10UU m 0 _ i - : f I~~~~~~- OC w 3 ~E II, BANK RESTRUCTURING Lessonsfrom the 1980s Edited by Andrew Sheng THE WORLD BANK WASHINGTON, D.C. C 1996 The International Bank for Reconstruction and Development/The World Bank 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing January 1996 The findings, interpretations, and conclusions expressed in this publication are those of the authors and do not necessarily represent the views and policies of the World Bank or its Board of Executive Directors or the countries they represent. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to copy por- tions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923, U.S.A. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors, and countries and regions. The latest edition is available free of charge from Distribution Unit, Office of the Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, The World Bank, 66, avenue d'1ena, 75116 Paris, France. Andrew Sheng is Deputy Chief Executive (Monetary) of the Hong Kong Monetary Authority. Library of Congress Cataloging-in-Publication Data Bank restructuring: lessons from the 1980s / edited by Andrew Sheng. p. cm. Includes bibliographical references. ISBN 0-8213-3519-7 1. Bank management. 2. Bank failures-Case studies. 3. Banks and banking-United States-Case studies. 4. Banks and banking-Spain- Case studies. 5. Banks and banking-Developing countries-case studies. I. Sheng, Andrew. HG1615.B3515 1996 332.1-dc2O 95-49776 CIP Contents Foreword v Acknowledgments vi Introduction 1 1. Banking Fragility in the 1980s: An Overview 5 Andrew Sheng 2. Bank Restructuring Techniques 25 Andrew Sheng 3. Resolution and Reforn: Supervisory Remedies for Problem Banks 49 Andrew Sheng 4. The United States: Resolving Systemic Crisis, 1981--91 71 Andrew Sheng 5. Bank Restructuring in Spain, 1977-85 87 Andrew Sheng 6. Structural Weaknesses and Colombia's Banking Crisis, 1982-88 99 Fernando Montes-Negret 7. Malaysia's Bank Restructuring, 1985-88 109 Andrew Sheng 8. Ghana's Financial Restructuring, 1983-91 123 Andrew Sheng and Archibald A. Tannor 9. Yugoslavia: Financial Restructuring in a Transition Economy, 1983-90 133 Andrew Sheng 10. Financial Liberalization and Reform, Crisis, and Recovery in the Chilean Economy, 1974-87 147 Andrew Sheng 11. Argentina's Financial Crises and Restructuring in the 1980s 161 Luis A. Giorgio and Silvia B. Sagari 12. Post-Liberalization Bank Restructuring 175 Andrew Sheng iii About the authors Andrew Sheng is Deputy Chief Executive Luis A. Giorgio is Deputy Director, Center for (Monetary) of the Hong Kong Monetary Authority. Latin American Monetary Studies. Fernando Montes-Negret is Principal Financial Silvia Sagari is Chief, Industry and Energy Economist, Financial Sector Development Division, Western Africa Department, at the World Department, at the World Bank. Bank. Archibald A. Tannor is Director, Banking Supervision, Bank of Ghana. iv Foreword The decade of the 1980s saw major stresses and Gerard Caprio, Jr., Izak Atiyas, and James Hanson, strains in the developing world. Fiscal adjustment, and Banking Institutions in Developing Markets, by trade liberalization, financial deregulation, and priva- Diana McNaughton, are companion publications to tization were major policy reforms that sought to deal this book. The first deals with the policy reform with the problems of high debt overhang, balance of agenda in financial sector policy, the second with payments problems, and slow growth in many devel- institutional reforms in individual banking institu- oping countries. The financial sector in many coun- tions. This book surveys experience with systemwide tries witnessed tumultuous change as bank balance bank restructuring, focusing on the policy and sheets reflected the devastating effects of external and process of the legal, institutional, financial, and man- internal shocks as well as major policy responses to agerial restructuring demanded by systemic bank such adjustments. Major initiatives were undertaken problems. in financial sector refotm: the building of new legal Bank restructuring is a process, not an event. The and regulatory frameworks, strengthening of banking causes of bank failure are often country specific, and and financial institutions, removal of interest rate and the solutions to bank failure may require special con- exchange rate restrictions, liberalization of market sideration of country conditions. However, lessons entry into the financial sector, commercialization and from international experience suggest that there are privatization of state-owned financial institutions, common techniques and approaches that can help and opening of financial markets to greater domestic clarify the issues, reduce their complexity, and identi- and foreign competition. fy possible solutions. During this period banking systems in industrial Given the magnitude of bank losses in many and developing countries alike underwent fundamen- industrial and developing countries today, we hope tal change. Many banks were devastated by the bur- that Bank Restructuring: Lessons from the 1980s will be den of nonperforming loans, brought on by a variety a useful guide for policymakers, bank supervisors, of causes. A centerpiece of financial policy reform was and bankers for dealing with these pressing problems bank restructuring, as policymakers sought to improve of the 1 990s and beyond. the process of financial intermediation that would fund and foster stronger growth. This book is part of a series of research studies Gary Perlin stemming from World Development Report 1989: Director Financial Systems and Development. Two books, Financial Sector Development Department Financial Reform: Theory and Experience, edited by The World Bank v Acknowledgments Having moved from the World Bank's development of the Bank, who helped identify, conceptualize, and advisory function to the core central bank function research the intricacies of financial sector market fail- of regulating and promoting financial markets in ure. Much of the research in this volume was under- Hong Kong, one of the freest of market economies, I taken by Susan Hart, who helped extensively in the have had the opportunity to reflect on what went study on the United States and sharpened many of right, what went wrong, and what should have been the arguments. Izabela Rutkowska continued where done during the banking crises in developing coun- Susan left off, and Bo Wang created many of the tries in the 1980s. Extrapolating lessons from the tables and figures. past runs the danger of boiling down essential truths Gerard Caprio, Yoon Je Cho, Samuel Talley, to such bland cliches that the advice becomes tauto- Melanie Johnson, and Ross Levine read chapters and logical. To say that macroeconomic stability was crit- offered valuable insights, comments, and suggestions. ical to recovery merely states the obvious-if there Helpful comments were also provided by several had been macroeconomic stability in the first place, anonymous referees. banking systems wouldn't have gotten into the mess Maria Raggambi, Wilai Pitayatonakarn, and they did. Karin Waelti, at the World Bank, and Helen Moy at This volume had its genesis in the pioneering the Hong Kong Monetary Authority, were able to work of Millard Long, who directed the World Bank's make out my unintelligible scrawls and assemble the World Development Report 1989: Financial Systems and drafts with great cheerfulness and accuracy. The Development. I was brought on board as the author for chapters were edited by Meta de Coquereaumont and the paper on the Malaysian case, and was delighted Paul Holtz and laid out by Christian Perez, of that a group of people, led by Manuel Hinds and American Writing Corporation. Aristobulo de Juan (then at the World Bank), was Finally, encouragement and support came from exploring banking problems in developing countries. Nancy Birdsall, Johannes Linn, and Jean-Fran,ois This volume forms part of three massive works that Rischard of the Bank, the more so for sparing me the emanated from these forays: Gerard Caprio's Financial valuable time to complete this task. Since moving to Reform: Theory and Fxperience, Diana McNaughton's Hong Kong, I have learned considerably about the Banking Institutions in Developing Markets, and this impact of the marketplace and free competition on attempt at understanding the techniques used in bank banking systems from my colleagues, David Carse restructuring in the 1980s. and Joseph Yam. To all the above, credit is due. All This volume therefore owes its intellectual debt to debits, errors, and omissions are mine. the above persons, but also to Alan Gelb, Vince Last but not least, my wife Suan Poh deserves Polizatto, Silvia Sagari, and Dimitri Vittas, colleagues thanks for putting up patiently with the trials and in the former Financial Policy and Systems Division tribulations of my writing. vi Introduction Not since the Great Depression had as many banks of such a complex subject can cover all these issues; failed as during the 1980s. To the casual observer thus readers of this volume should bear in mind two most bank failures are the result of abuses of power caveats. First, this volume addresses the resolution of and trust by bank owners and managers-and in systemic bank problems, not the restructuring of many cases this diagnosis is correct. But large-scale individual banks. As a guide to best practice on the bank failures are symptoms of a broader malaise. techniques and tools used by different countries to Accordingly, bank restructuring-defined as the resolve large-scale bank distress, its intended audience package of macroeconomic, microeconomic, institu- is policymakers-whether bank supervisors, central tional, and regulatory measures taken to restore prob- bankers, treasury officials, or informed bankers. lem banking systems to financial solvency and A strictly market-based approach to failed banks health-must address the causes and effects of wide- would call for their liquidation. Indeed, this is the best spread bank distress. solution for isolated, small bank failures: clean surgery Banks are important because they are the main is often less messy than slow medicine. But where channels of savings and the allocators of credit in an large banks suffer from a lack of public confidence economy. Banks offer instruments that are money and large segments of the banking system are insol- substitutes, and they operate the payments system. vent, liquidation only masks the problem. If banking Their efficiency affects the entire economy, and bank- fragility is a symptom of economywide problems, liq- ing system failure erodes public wealth and confi- uidation alone is neither practical nor useful. The res- dence in the economy. The failure of 10,000 banks in olution of systemic bank problems therefore must be the United States between 1930 and 1933 made the part of an overall strategy to restructure and reform Great Depression much deeper and long-lasting than fundamental inefficiencies in the economy. it might otherwise have been (LaWare 1994). That Second, this volume does not attempt to draw trauma led to the extensive U.S. deposit insurance quantitative empirical conclusions, mainly because of scheme and its associated regulatory framework, the lack of comparable cross-country data. Policymakers everywhere protect or regulate the Accounting standards in banking vary, particularly in banking system on efficiency, welfare, and public pol- loan classification and income accrual on nonper- icy grounds. As this volume shows, these regulations forming loans, making comparisons of bank losses sometimes build perverse incentives in the banks- extremely difficult. In addition, countries are reluc- such as moral hazard through deposit insurance- tant to publish data on bank losses because of the that themselves give rise to problems. potential impact on confidence in the banking sys- tem. Such empirical work will not be possible until Identifying Problems there is greater transparency in international account- ing and regulatory standards, as well as better data on Banking problems have many roots, ranging from the size of fiscal and quasi-fiscal deficits. distorted management incentives to institutional fail- Despite these limitations, important lessons can ure to misguided macroeconomic policies. No study be drawn from the eight case studies in this volume: 2 BANK RESTRUCTURING Spain and the United States in the industrial country bank fragility than low-inflation economies. Inflation group and Argentina, Chile, Colombia, Ghana, disguises the extent of damage to the banking system, Malaysia, and Yugoslavia in the developing country which a period of deflation quickly exposes. In the group. These studies indicate that banking problems cases studied, bank restructuring was part of price in the 1980s were essentially an outcome of bad stabilization or occurred during a period of deflation, policies, poor management, and weak institutional as bank assets became compressed and eroded avail- frameworks. able bank capital. Governments' ability to transfer Both external and internal factors contributed to real sector inefficiencies to savers through the infia- systemic bank failures. Dramatic changes in the inter- tion tax became severely weakened as financial mar- national economy created fragility and volatility in the kets globalized. Savers simply escaped the inflation macroeconomic environment. Globalization of trade tax through capital outflows, putting grave pressure and finance, liberalization and deregulation of real and on the exchange rate. Deregulation of financial mar- financial markets, and changes in technology all kets and liberalization of trade and capital flows open increased competition and risks for banking, while also up the possibility of large portfolio shifts from eroding the franchise value of protected bank markets. domestic financial assets (disintermediation) as The willingness of governments and enterprises to wealthholders perceive potential losses from bad poli- incur large debts in the inflationary 1970s and the cies, bad management, or bad institutional frame- ensuing debt crisis and adjustment in the 1980s creat- works. The restoration of financial discipline begins ed enormous strains on the banking community. with the restoration of fiscal and monetary discipline. Nevertheless, as a number of the case studies 2. Banks fail because of losses in the real sector, com- show, bank failures can occur because of perverse pounded by poor risk management and fraud. incentives even where growth is stable. Excessive con- Deregulation, technological advances, and globaliza- centration of bank resources, connected ownership of tion-in both the real and financial sectors-have banks and enterprises, inadequate supervision, and increased the volatility and risks to which banks are deposit insurance coverage (even if only implicit) can exposed. The move toward flexible exchange rates in result in extensive losses. the 1970s and subsequent interest rate deregulation opened bar;ks up to much higher credit and market Recognizing Causes-and Solutions risks relative to their capital base. Competition from nonbanks in the deposit and credit markets has large- Given the variations in initial country conditions- ly eroded the franchise value of banking, particularly legal framework, banking practices, industrial and for U.S. banks. At the same time, changes in relative ownership structure, resources, and policies-what is prices in real markets have brought about large losses successful in one restructuring could easily be disas- in the enterprise sector. The bursting of asset (real trous in another. There are many parallels between estate and stock market) bubbles-created as a result the resolution of domestic banking crises and the res- of low interest rates, excessive tax incentives, and olution of the international debt crisis of the 1980s. information asymmetry-further eroded bank capi- A case-by-case approach was applied for many coun- tal, as in Japan and Scandinavia. Bank losses in a tries affected by the debt crisis, designing custom number of countries were compounded by an over- solutions for each case. As recent experience with concentration of assets-geographically, sectorally, or bank problems in Eastern Europe and the former in terms of ownership (as in Latin America)-which Soviet Union has shown, there are common problems encouraged connected lending and credit abuses. At but no common solutions. both the microeconomic and macroeconomic levels, Still, a number of lessons emerge from the experi- bank managers and policymakers have not managed ence of the 1980s. These lessons can help guide the these risks very well. bank restructuring efforts that are under way in a 3. Liberalization programs often fail to take into number of countries: account the wealth effects of relative price changes, and 1. Financial stability rests on thegovernment's ability to inadequate supervision creates furtber losses. Rapid maintain a stable currency. Fiscal and financial disci- trade liberalization may create losses for previously pline are the anchors of a stable financial system; protected enterprises, leading to large bank losses. If without them reliable credit decisions cannot be these enterprises belong to groups that also own made. High-inflation economies are more prone to banks, in a situation without adequate bank super- Introduction 3 vision, these banks are likely to finance distress bor- require liquidating some institutions. At the micro- rowing at unrealistically high real interest rates. These economic level, changing management is vital. high rates quickly become a systemic problem, calling Managers who are part of the problem cannot be part for a public bailout. Strong bank supervision and of the solution. enforcement, together with laws that encourage debt 7. The method of loss allocation determines the success of discipline and avoid bank owner-borrower conflicts, the resiructuring program. Bank losses ultimately are are important components of bank restructuring pro- borne somewhere in the economy. Since no one is grams. This is perhaps the major lesson of financial willing to accept such losses voluntarily, loss allocation sector liberalization in the Southern Cone economies is a major political issue. The wealthy may attempt to (Argentina, Chile, and Uruguay). escape such losses through capital flight. The poor can- 4. Bank losses ultimately become quasi-fiscal deficits. not escape an inflation tax. Since there is no formula Large-scale bank failures are simply not acceptable in for the democratic distribution of losses, losses have to most economies. In none of the cases studied did gov- be allocated by accepted law or by arbitrary policy. The ernments dare to pass widespread bank losses on to ability to allocate losses depends on a country's politi- depositors. Most bank losses are absorbed by the bud- cal and institutional framework. The technique adopt- get or by the central bank through explicit or implicit ed, either a flow or a stock solution, depends on the deposit protection schemes. Such schemes place enor- degree of distress. As mentioned, when banks are in mous burdens on the budget. During financial crises severe distress governments tend to absorb bad debts governments are made to assume considerable debt: all and to exchange-"carve out"-government or central external debt (public or private), internal debt (includ- bank bonds for bad debts to recapitalize banks. ing debt of public enterprises), and losses in tk. bank- Insolvent institutions, however, cannot be rescued by ing system (public or private). In almost all the cases another insolvent institution. As shown in Argentina studied financial stability was restored only when gov- and Yugoslavia, overreliance on seigniorage to finance ernments were able to maintain a sustainable fiscal bal- bank debts ultimately explodes into hyperinflation. ance without monetary creation. Banking failures in Loss allocation that maintains macroeconomic stability the 1 980s were largely market failures, caused in large requires a budget that is able to generate a primary sur- part by moral hazard induced through implicit or plus to service its debts (including debts incurred by explicit deposit insurance. Deposit insurance demands the carve out) without excessive monetary creation. sound preventive bank supervision. Losses are allocated either to depositors through an 5. Failure recogzition is important because a banking inflation tax or to taxpayers. In the most extreme cases, crisis is a solvency problem, not a liquidity issue. The when the budget is unable to bear the huge internal resolution of banking problems is often delayed and external debt without creating hyperinflation, the because officials are unwilling or unable to determine government may have to undertake a deposits-to- the magnitude of the problems. In many cases banks bonds conversion, as occurred in Argentina and Brazil that may not appear to be insolvent in accounting in 1990. Such forced losses-borne by the deposi- terms are in fact insolvent when assets and liabilities tors-gave the government breathing space for other are priced at current market values. Any banking sys- reform measures to work, particularly in fiscal reform, tem with nonperforming loans (net of provisions) trade, and privatization of state-owned enterprises. exceeding 15 percent of total loans is probably reach- 8. Success depends on suffcient real sector resources to pay ing a crisis stage. Insolvent banks can hide such losses off losses, adequatefinancialsector reforms to intermediate with bad accounting, but failing to deal with hidden resources efficiently and safely and the budgets ability to losses can create perverse incentives in the banking tax "winners"and ivind down "losers" without disturbing system, leading to inefficient resource allocation and monetary stability. Because bank losses are rooted in adding to macroeconomic instability. real sector financial imbalances-enterprise losses or 6. Stopping the flow offuture losses is critical. large fiscal deficits-bank restructuring is inextricably Stemming future losses involves changing the incen- linked to fiscal and enterprise reforms. Recapitalizing tives within the real and financial sectors. Where loss- banks without addressing the underlying enterprise makers are public enterprises and banks, changing losses or inefficiencies runs the risk of repeating bank- the incentives structure requires changing ownership, ing problems in the future. The Yugoslav experiment particularly through privatization and foreign capital with worker ownership of enterprises and banks in the and expertise. Enforcing hard budget constraints may 1980s showed how the inability to change ownership 4 BANK RESTRUCTURING and to separate ownership between borrowers and Successful reforms call for, among other things: lenders can create large amounts of distressed borrow- * Political will and strong leadership, with a dedi- ing that ultimately culminates in inflation. cated, unified economic team. 9. Rebuilding a safe and profitable banking system * A carefully sequenced, coherent, and comprehen- requires good policies, reliable management, and a sive implemnentation plan. strong institutional framework. At the policy level, * The ability to sell the plan to every level of society governments must maintain credible macroeconomic (Rhodes 1992). policies that encourage stability, competition, and Bank restructuring techniques are simply a set of growth. At the management level, incentives have to tools. How these tools are used depends on how well be right and the ownership and governance of banks policymakers understand the nature of the problem must be addressed. Good managers should be and what combination of tools is best suited to deal rewarded for prudent risk management and punished with the problem. The most successful bank restruc- for speculative and fraudulent behavior. Bank laws turings have been those that are simple yet commit- and regulations should be enforced. The accounting ted. Political will and the ability to execute changes framework should encourage the measurement and simply and transparently worked, for example, in disclosure of economic performance using interna- Chile, Malaysia, and Spain. But as with any policy, tionally accepted accounting standards. The pay- the design of bank restructuring programrs involves ments system must work efficiently and robustly. tradeoffs between risk and return. And the choice New financial markets should be created to help ultimately depends on how each country values social mobilize risk and long-term capital and permit better welfare relative to efficiency. monetary management using indirect tools. Prevention is better than a cure. But can a fail- 10. Time and timing are of the essence. Policies can proof banking system be designed? U.S. Federal change overnight, but it takes much more time to Reserve Chairman Alan Greenspan has said that "the get management incentives right and to restructure optimal degree of bank failure is not zero, and in all the institutional framework-law, accounting, regu- likelihood, [is] not even close to zero." Even if banks lation, and infrastructure. Bank losses develop over held nothing but low-risk government obligations, time and should not be expected to disappear quick- there is no guarantee that the government itself ly. But the sooner the problem is recognized and would never default on its debt. dealt with, the lower the costs to the economy and The process of bank restructuring continues in the banking system. many countries. For the post-centrally planned From these lessons it is easy to assume that bank- economies in transition to market-based economies, ing crises arise from recession alone. But countries bank restructuring will be a challenge that continues with strong growth and good fundamentals also had well into the 1990s. For the post-liberalization banking problems, the obvious example being Japan economies, that is, economies that have opened their in the 1980s. Banking weaknesses can develop capital accounts, the challenge now is how to man- because of excessive risk-taking during periods of age the banking system risks in a volatile world of growth, particularly if insufficient supervisory atten- global capital flows (chapter 12). It is hoped that this tion is paid to such excesses. Without growth in real volume will help these countries better understand output or extensive reserves, it is difficult for a coun- the techniques and processes involved in the difficult try with banking losses to pay for them-but some- road ahead. one has to pay. Although each country must find its own solutions, international experience offers an array References of techniques to draw on. LaWare, John P 1994. "Bank Failures in a Sound Economy." In Implementing Change Frederick C. Schadrack and Leon Korobow, eds., The Basic Elements of Bank Superuision. New York: Federal Reserve Bank of New York. The pace and model of reform are determined by a Rhodes, William. 1992. "Third World Debr: The Disaster that country's political and institutional framework. Didn't Happen." The Economist, September 12. CHAPTER 1 Banking Fragility in the 1980s: An Overview Andrew Sheng The 1980s may well be remembered as the decade of In selected OECD countries provisions against debt, inflation, and adjustment. The decade began nonperforming loans rose toward the end of the with the deepest international recession since the decade (table 1.1). In the developing world the World 1930s, saw the eruption of the debt crisis, and ended Bank provided financial sector adjustment loans to with the fragmentation of the socialist economic twenty-two countries in the 1980s, and in almost all of bloc and its integration with the global economy. these countries financial distress-in which large parts Despite the increasing globalization of financial mar- of the financial system had negative capital-was kets and unprecedented financial innovation and lib- present to some degree. In the formerly socialist eralization, bank crises and restructuring were com- economies in transition to market economies, a new mon in both industrial and developing countries. By group of problem banks has emerged in the nascent the end of the decade the Scandinavian countries financial systems. These banks are struggling with (except Denmark), Spain, and the United States had inherited portfolios of dubious quality while trying to all experienced severe banking problems. Financial transform themselves into market-oriented institutions. fragility (defined as the deterioration of bank solven- Important lessons can be drawn from this decade cy due to poor asset quality and declining piofitabili- of bank distress and adjustment. In the United States ty) was evident in the banking systems of Taiwan more than 1,300 banks and 1,400 thrift institutions (China) and Japan and a number of other member failed or were merged and consolidated during countries of the OECD. 1980-91, compared with only 210 closures during Table 1.1 Provisions against nonperforming loans in OECD countries, 1981-90 (as a percentage of net income) Country 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Denmark Commercial and savings banks 57.0 56.8 21.7 89.2 20.7 - 66.5 53.1 75.9 129.0 Finland Savings banks 76.8 80.6 74.0 70.4 71.8 69.9 70.0 54.9 43.7 77.4 Japan Commercial banks 3.2 10.8 6.7 7.1 4.0 6.8 5.4 7.6 8.7 7.2 Norway Commercial banks 46.5 54.4 42.7 48.0 54.2 59.1 138.4 125.8 90.3 209.8 Spain Commercial banks 50.2 60.8 63.3 63.9 54.8 49.3 47.7 37.0 27.9 27.9 United Kingdom Commercial banks - - - 43.7 33.7 31.5 89.5 17.9 94.0 60.7 United States Large commercial banks 23.5 34.8 39.9 44.9 42.0 46.8 99.7 31.5 65.4 65.6 Mutual savings banks -1.8 -3.2 41.5 19.8 9.8 8.8 12.2 24.4 89.7 196.9 Savings and loan associations -4.8 -16.5 29.1 52.1 40.3 72.3 175.8 480.3 - - - NoE available. &Sur: Schuijer 1992. 5 6 BANK RESTRUCrURING 1945-79. At the end of 1991 the U.S. Federal Table 1.2 Growth and inflation, 1965-80 and Deposit Insurance Corporation (FDIC) estimated 1980-90 that there were 1,069 problem banks, or 8.6 percent (percent) of the banks in the United States, with assets totaling 1965-80 1980-90 $611 billion, or 18 percent of total bank assets. The Country group GDP Inflation GDP Infltion number of bank failures has declined since the late Lowandmiddleincome 5.9 16.7 3.2 61.8 1980s, but the assets of failed banks have been Developing Europe - 13.9 2.1 38.8 increasing. In 1991 the total assets of failed banks East Asia and the Pacific 7.3 9.3 7.8 6.0 reached $66.2 billion, the highest figure since the Latin America anid che Caribbean 6.0 31.4 1.6 192.1 1930s. Of the U.S. banks that failed in 1991, eleven Middle East and had assets of more than $1 billion. Resolution of the North Africa 6.7 13.6 0.5 7.5 South Asia 3.6 8.3 5.2 8.0 thrift crisis cost at least $180 billion, or 3 percent of Sub-Saharan Africa 4.2 11.4 2.1 20.0 gross domestic product (IMF 1993). Severely indebted 6.3 27.4 1.7 173.5 In Japan estimates of bank exposure to bad debt OECD 3.7 7.6 3.1 4.2 from real estate and other problem loans were as World average 4.0 9.2 3.2 14.7 high as V 40 trillion ($160 billion), or 10.4 percent Soure: World Bank 1992. of total loans outstanding, at the end of 1992 (Huh and Kim 1994). Banks' exposure to real estate debt Background to Financial Fragility was caused by their exposure to nonbank financial institutions, where 41 percent of total loans were The 1980s were a decade of slow growth, large made for property (IMF 1993). The Japanese gov- financial imbalances, and high inflation. Average ernment recently announced several packages of annual growth rates in the low- and middle- measures to stimulate the economy and aid the income developing countries dropped to almost banking industry. In Scandinavia loan losses half their late 1960s and 1970s levels, while infla- incurred by banks in Finland, Norway, and Sweden tion rates more than tripled (table 1.2). during 1991-92 amounted to 4.2 to 6.7 percent of After enjoying strong terms-of-trade gains averag- gross domestic product (GDP). The overnight fail- ing 8 percent a year in the 1970s, the developing ure of the U.K.'s Barings Bank in February 1995 countries' terms of trade deteriorated, on average, by due to speculative activities in East Asia demonstrat- 1.8 percent a year in the 1980s. Export growth in ed the vulnerability of banks to weaknesses in inter- developing countries slowed to 2.6 percent a year, nal controls. reducing their share of world exports from one-third In the developing countries problem loans of 15 in 1980 to one-quarter by 1990. As a result of these to 30 percent of total loans were not uncommon dur- downturns, the overall balance of payments of devel- ing the 1980s, while recent estimates of bank bad oping countries deteriorated considerably-from a debt in the transition economies run as high as 55 to surplus of $145 billion in the 1970s to a deficit of 60 percent of total loans (Sheng 1992). (See annex $245 billion in the 1980s. By any measure, the debt 1.1 for a summary of major banking problems burden of the developing countries roughly tripled around the world in the 1980s.) between the 1970s and 1980s (table 1.3). A number of studies have examined the causes and economic effects of financial crises (Hinds Table 1.3 External debt indicators for developing 1988, Davis 1989, Sundararajan and Balifio economies, 1970-75 and 1983-89 1991). This book focuses on the macroeconomic, (percent) microeconomic, institutional, and regulatory mea- Interestpaymentsl sures taken to restore problem banking systems to Externd debtdGNP exports financial solvency and health. It draws on the Countrygroup 1970>-75 198.3-89 1970-75 1983-89 lessons of bank restructuring in eight countries l , l r l , , , ~~~~~~Low income 10.2 28.5 2.9 9.8 based on a number of background papers prepared Low income for World Development Report 1989 (World Bank (exduding 1989). Specifically, it analyzes various bank China and India) 20.5 60.7 2.9 11.8 restructuring techniques and their applicability Middleincome 18.6 54.9 5.1 15.4 under different conditions of bank distress. Source:World Bank 1991. Banking Fragility in the 1980s: An Overview 7 Despite efforts to roll back state intervention had reached such a level that in a single day the value through privatization and tighter fiscal discipline, of payments through twenty-one countries studied by the state's role in the economy continued to grow the BIS in 1989 totaled $3 trillion. The annual trans- in both developing and industrial countries. In the actions turnover of the eleven BIS countries (Group OECD countries general government spending as a of Ten plus Switzerland) was more than fifty times share of gross national product (GNP) reached 42 annual GNP in 1989 (BIS 1990). BIS estimates in percent in 1989, compared with 33 percent in April 1992 indicate that foreign exchange market 1970. Central government deficits almost doubled turnover was about $880 billion a day. in the decade, while in Latin American countries Vokstile relative prices. The trend toward global fiscal deficits almost tripled. interdependence coincided with the emergence of higher volatility in relative prices, beginning with the Financial trends abandonment of fixed exchange rates in the early 1970s and the rise of global inflation. The world In contrast to the disappointing macroeconomic commodity price index rose by 269 percent between environment of the 1980s, world financial markets 1970 and its peak in 1980, before falling 27 percent today are characterized by globalization, liberaliza- to its trough in 1986. Inflation in the low- and mid- tion, innovation, and re-regulation. dle-income developing countries rose to an annual Globalization of financial markets. The roots of average of 62 percent in the 1980s, compared with bank distress in the 1980s can be traced to the emer- 17 percent in the 1970s. As inflation rose, real inter- gence of global financial markets and flexible est rates reached record levels. In the United States exchange rates in the 1970s. During that decade the real lending rates averaged 6.2 percent in the 1980s, Eurocurrency markets expanded rapidly as banks compared with 0.5 percent between 1974 and 1979. began to internationalize their operations, seeking Real lending rates for countries that underwent higher profits offshore in order to escape the interest financial crisis varied from more than 40 percent a rate ceilings imposed on domestic markets. Gross year in Chile (1981-83) to over 200 percent in new issues of Eurobonds rose from $27 billion in Argentina (1984-85). Nominal exchange rates 1981 to $319 billion in 1991. During 1972-82 plunged as inflation rose. The U.S. dollar, in which international banking assets grew two-and-a-half more than half the foreign currency assets of the BIS times faster than the GDP growth of OECD coun- reporting banks are denominated, depreciated by 30 tries and an average of 10 percent a year faster than percent in real terms during the 1980s, compared world trade (Bryant 1984). By the early 1980s devel- with a variation of less than 10 percent during oping countries had become highly leveraged and 1976-80. were not prepared for the sharp increase in interest The world banking system, which had experi- rates when the U.S. Federal Reserve tightened in enced relatively stable interest rates during the 1950s, 1981. By the time the Mexican debt crisis erupted in 1960s, and most of the 1970s, suddenly had to cope 1982, the developing countries had accumulated an with rapidly changing interest and exchange rates and external debt of $720 billion, or roughly one-third of large capital flows as funds moved rapidly both the assets of the Eurocurrency market. Global inter- domestically and internationally in search of higher dependence had reached the point where the solven- yields at lower risks. cy not only of debtor countries but also of the lender Innovation and competition. At the same time, banks was at stake. rapid improvements in international transport, By 1990 total foreign liabilities of the global telecommunications, and computer technology trans- banking system had grown to $7 trillion, more than formed the business of finance. The emergence of twice the level of annual world exports. Much of the credit cards and electronic funds transfer technology growth in the international banking system was in made severe inroads into the traditional payments the industrial world, whose share of international system business of the banking sector. New nonbank bank liabilities grew from 71.1 percent of the total in competitors-especially money market funds, travel 1976 to 76.6 percent in 1990. At the end of 1991 companies, retailers, insurance companies, mortgage crossborder interbank claims within the area of the specialists, and pension funds-began to offer high- Bank for International Settlement (BIS) alone were er-yielding deposit substitutes that eroded the low- $4 trillion. Global trading in financial instruments cost "captive" deposit base of the banking system. 8 BANK RESTRUCTURING On the asset side, innovations in financial engi- was followed by major liberalization efforts in neering pioneered by investment banks created finan- Australia (1983) and New Zealand (1984). The liber- cial instruments that offered lower costs and higher alization of Japan's financial market after the 1985 liquidity to borrowers-eating into the previously Plaza Accord had a profound impact on capital flows bank-dominated loan market. Highly leveraged around the world, particularly in the Asia-Pacific transactions (junk bonds) evolved to assist the merg- region. ers and acquisitions and management buyouts in the By the late 1980s deregulation of the financial United States. These high-risk, high-yield bonds sector had swept most of the OECD countries and attracted many investors, including banks, thrifts, many developing economies. The trend was toward and pension and insurance funds. complete freeing of the capital account, liberalization Moreover, the rise of international banking of interest rate controls, freeing of competition brought competition from foreign banks, which between banks and nonbanks, and allowing banks to began to rapidly penetrate domestic markets, espe- enter new fields, such as securities trading and even cially if they had lower costs of capital and were not (under certain conditions) insurance. Change was subject to domestic deposit constraints. Technology coming not only to banks, but also to bank regula- made the cost of intermediation cheaper and the tors, who had to adjust to more complex supervision transmission of credit information faster. In short, in a more volatile market environment. the traditional "franchise" of banking, which was pro- The banking sector also responded to new cor- tected by legal and market barriers, came under petition by lending more aggressively in traditional severe attack from changing market conditions, com- but higher-risk markets such as real estate. In the petition, and technology. United States the threat of large losses from banks' Banks in the industrial countries responded to the exposure to developing countries' sovereign debt changing environment in three ways: innovation, cred- swung banks toward domestic lending, particularly it expansion, and deregulation. First, they engaged in consumer lending and commercial real estate. The financial innovation-such as asset securitization and real estate boom in the United States was partly stim- use of derivatives-that "saved" on capital require- ulated by tax incentives and by the ready availability ments by placing assets and risks off the balance sheet. of credit, which arose from competition to lend By the end of 1990 more than one-third of U.S. mort- among the distressed thrifts, the commercial banks, gages were securitized and sold in the secondary mar- and insurance and pension funds-all in search of ket. Credit card and consumer debt were also being higher yields. packaged and sold. One of the most spectacular Real estate exposure was particularly evident in a growth areas in financial innovation is trading in number of countries that exhibited the "Dutch dis- derivative instruments (swaps and options). Trading in ease," with overvaluation of the exchange rate stimu- financial futures and options has grown phenomenally lating domestic spending on nontradables. This was because of lower transaction costs, high liquidity, and most noticeable in Malaysia, Norway, Sweden, and lower capital costs. Total outstanding derivatives the United States. Banks in Japan and the United (mainly interest rate futures and options and currency Kingdom also faced high exposure in the real estate and interest rate swaps) amounted to $7.5 trillion at and housing sectors. the end of 1991, a fivefold increase since the end of Domestic policy distortions. Most of the banking 1986. systems and financial markets in developing countries Second, the threats of new competition and inno- were unprepared for these structural changes. Many vation induced banks to lobby policymakers to dereg- of these financial systems were underdeveloped ulate their range of activities to allow them to engage because of severe financial repression-governments in businesses previously barred to them, such as controlled interest rates, directed the allocation of stock-market trading and funds management. By the credit, and used highly negative real interest rates to early 1980s most OECD countries had embarked on finance fiscal deficits and inefficient state-owned both interest rate liberalization and removal of enterprises. The shortfall between domestic savings exchange controls. The United Kingdom began its and investment was financed mainly by external bor- "big bang" liberalization by removing exchange con- rowing. Financial markets were highly segmented, trols in 1979 and thereafter allowing commercial with little or no competition among markets. Banks banks to enter into securities market trading. This were highly protected, and nonfinancial instirutions Banking Fragility in the 1980s: An Overview 9 were prohibited from innovation and competition. currencies depreciated in the wake of structural Legal and accounting structures became obsolete in adjustment programs (table 1.4). The largest expo- an era of changing market prices and new financial sures were in Eastern Europe (Hungary, Romania, instruments and market practices. Turkey, and Yugoslavia) and Latin America, with net Thus, when external shocks came in the form of foreign liabilities exceeding $40 billion. By 1990 the severe international recession of 1980-81 and the African and Latin American banks were still facing cutoff of external resources as a result of the interna- large net foreign liabilities, while Asian banks had tional debt crisis, many developing countries were moved to a large net foreign asset position. totally unprepared for the severity of the impact of relative price changes on their financial sector balance Deposit insurance sheets. By the mid-1980s the combined effect of high There was, however, a fundamental asymmetry in the levels of bank lending and deterioration of asset qual- financial sectors' risk profile. Although formal deposit ity (in particular, the developing countries' debt bur- insurance exists in a limited number of countries, the den on OECD banks) had eroded banks' capital base unwillingness of governments to allow banks to fail to a historic low of 5 to 6 percent, compared with 7 for fear of systemic failure (bank capital is negative to 8 percent in the 1960s and 1970s. The interria- systemwide) has meant that implicit and explicit gov- tional regulatory response to the marker changes was ernment guarantees exist for almost all banking liabil- to stem the capital erosion by harmonizing risk-based ities. Despite the limited stated coverage of most for- capital adequacy standards and improving overall mal deposit insurance schemes, actual coverage has supervision. In 1988 the Basle Committee on Bank been almost 100 percent because bank failure would Supervision agreed to enforce a minimum risk- create social and economic upheaval. Consequently, a weighted capital-asset ratio of 8 percent by the end of central problem in the global banking system is that, 1992. In a number of countries, however, the damage irrespective of public or private ownership of banks, had already been done. commercial bank losses in excess of capital have become de facto quasi-fiscal deficits (for a survey of Structural weaknesses quasi-fiscal deficits, see Blejer and Cheasty 1991). The blanket state guarantee on bank deposits, In a world of more volatile relative prices, changing coupled with weak bank supervision and enforce- technology, and intense competition, the global ment, created massive problems of moral hazard in banking system was caught in a double bind. On the almost every country. Bank management could and liabilities side of the balance sheet, which comprised did take risks far beyond prudential levels because mostly deposits, disintermediation from banks losses were ultimately borne by the state. Under per- occurred wherever banks faced deposit rate ceilings or verse incentives and poor supervision, even good high inflation or where nonbanks with lower reserve bank managers became bad managers, engaging in or intermediation costs offered more competitive speculation, excessive spending, and ultimately fraud rates. On the assets side, increasing competition and (de Juan 1987). Recent studies in the United States improved capital and money markets lured away low- blame moral hazard for major losses in the savings risk customers, forcing banks to take higher credit and loan crisis. risks. Interest rate and exchange rate risks rose with higher inflation as policymakers sought to engage Table 1.4 Net foreign liabilities of domestic banks more actively in macroeconomic stabilization mea- (billions of U.S. dollars) sures, particularly in countries caught in the debt cri- sis. As real interest rates rose and exchange rates depreciated, the asset side of bank balance sheets World 42.9 64.9 330.2 Industrial countries 4.6 -16.0 315.1 de:eriorated, while banks in countries with net for- Developing countries 38.3 81.0 15.2 eign exchange liabilities were hit with large revalua- Africa 2.0 4.0 8.9 tion losses. Asia 0.7 8.4 -29.9 Europe - 42.1 19.4 By 1980 the developing countries banks had a Middle East -4.3 -15.9 -52.5 net foreign liability exposure of $81 billion, which Western hemisphere 21.3 42.5 69.3 subjected them to large revaluation losses when their Source: IMF, various years. 10 BANK RESTRucrURING Inadequate capital base Figure 1.1 Capital deterioration as a finction of nonperforming loans The capital base of many banks around the world is Capital/asset ratio (percent) probably inadequate relative to the risks in a world of 8 volatile relative prices and their impact on asset val- 7 - ues. Despite efforts to raise the level of capital in 6 industrial countries, total capital-asset ratios have fall- 5 - en from around 50 percent in the nineteenth century, 4 - to around 15 to 20 percent during the 1930s, to less 3 - than 10 percent today. 2 - The Basle Committee on Bank Supervision I - agreed to impose minimum risk-weighted capital- 0- asset ratios of 8 percent for the end of 1992. Capital- -1 - asset ratios of banks in the Euromoney Top 500 banks -2 - display an average ranging from 4.6 percent in -3 OECD countries to 20.1 percent in Latin America - 4 l l l l l l l l l (annex 1.2). International comparisons of capital ade- 0 2 4 6 8 10 12 14 16 18 20 quacy are deceptive, however, because of different Percentage of nonperforming loans standards of loan provisioning. Note: Assumes the loan portfolio has a maturity of seven years and standards yields 10 percent a year. Nonperforming loans would reduce yield The impact of declining market yields (price proportionally. volatility) on the market valuation of bank assets is also deceptive (figure 1.1). Because banks have percent decline in loan yield due to a corresponding increasingly extended loans of longer maturity, par- increase in nonperforming loans would totally decap- ticularly to the real estate and industrial sectors, and italize a bank with an 8 percent capital-asset ratio. because of the practice of short-term loan rollovers, As a rule of thumb, therefore, financial distress is the average maturity of a bank loan today is much likely to become systemic when nonperforming longer than reported in the books. According to de loans, net of provisions, reach roughly 15 percent of Juan (1987), who draws on his experience in the total loans. An alternative way of confirming this is to Spanish banking crisis, the worst loans are those that assume that the average ratio of loan loss provisions are reported "current," because banks roll over bad against nonperforming loans is 50 percent. When debts and continue to make loans to borrowers who nonperforming loans exceed 15 percent of total are "too big to fail" through the process of "ever- assets, the capital base of 8 percent would be totally greening" (the extension or rollover of bad loans to eroded by loan loss provisions. loss-making borrowers to cover up the extent of dam- age from nonperforming loans). Inadequate loan loss provisioning The economic maturity of a bank loan is not its contracted maturity, but depends on the borrower's The 15 percent threshold is confirmed when we ability to service his or her debts. Theoretically, the examine time-series data on nonperforming loans in maturity of a debt in which the borrower is able to a number of countries (table 1.5). International service only interest but not principal is infinity. comparisons of nonperforming loans are not totally During a recession many borrowers fall into this valid because of varying definitions of nonperfor- category, while at the same time the maturity profile mance. Until recently, for example, loans in one of the bank's deposit base shortens as depositors try South Asian country that were not serviced for more to reduce their exposure to the risks of bank default. than three years were still treated as performing. The At the height of the Argentine bank crisis in 1989, U.S. norm for a nonperforming loan is one that has the bulk of the deposit base had only seven days' not been serviced for more than ninety days, maturity. although general international practice varies Assuming that the maturity of the loans of a between 90 and 180 days. Some countries treat all banking system is roughly seven years, figure 1.1 loans to state-owned enterprises as performing, since illustrates the discount in asset value if all the assets these are state guaranteed, even though many of are marked to market like a seven-year bond. A 15 these loans have not been serviced. Banking Fragility in the 1980s: An Overview II Table 1.5 Ratio of nonperforming loans to total loans, selected countries, 1980-90 (percent) Country 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Argentina - - - 16.9 29.1 30.3 24.6 25.1 27.1 - - Colombia 7.2 8.7 14.6 16.2 25.3 25.1 22.6 18.5 11.7 - - Ecuador 9.9 13.5 16.2 17.4 13.9 11.9 10.8 - 13.4 - - Malaysia - - - - - 10.3 20.8 26.8 32.6 24.5 20.8 Philippines 11.5 13.2 13.0 8.9 12.7 16.7 19.3 - - - - Uruguay 8.9 14.6 30.4 24.7 22.3 36.2 45.9 25.2 - - - Venezuela 7.6 8.3 9.3 15.6 15.3 13.3 9.8 7.0 10.8 - - United States - - 3.3 3.5 3.3 3.2 3.3 4.1 3.6 3.7 - - Not available Source: Morris 1990: Monires-Negret 1990: Yusofand others 1994; U.S. Bureau of the Census 1991. Whatever the variation in definitions, banking sys- blame bad bankers, and depositors are quick to biame tems with problem loans of more than 15 percent inadequate supervision. The evidence from the eight inevitably have encountered crises and failure of some case studies suggests that no single cause, but rather a institutions. Bad debts do not occur overnight, but combination of causes, may be nearer the truth. they can quickly build up over two to three years, and In all eight cases banking crises or problems were once they exceed 10 percent of total loans, the likeli- associated with declines in real economic growth either hood of bank failure escalates rapidly unless the bank- before or during the crisis period (table 1.6). The ing system is completely state owned. In contrast, even sharper the decline in growth, the more severe the eco- though the ratio of nonperforming loans to total loans nomic "losses" to the private sector. The real wealth has never exceeded 4 to 5 percent in the United States losses to the private sector in Chile in 1982 were esti- (because of early detection through good accounting mated at 28 percent of GDP (based on a proxy that standards and fairly strict supervision and enforce- included the decline in dollar terms of the fall in the ment), many banks still failed because of their expo- capitalization of the stock market and the real value in sure to geographical or sectoral concentration of risks. money and quasi-money; table 1.7). In 1989 holders For example, four of the five largest banks in Texas of Argentine broad money lost the equivalent of 27 failed and were merged or sold because of large expo- percent of GDP in real terms, but borrowers gained sure to property, commercial, and energy loans during from the inflation tax (erosion in the value of real the downturn of energy prices in 1985-87. debt) to the extent of 16 percent of GDP In Malaysia, where nonperforming loans were Thouglh the twin macroeconomic imbalances of defined as those that were not serviced for more than fiscal and current account deficits existed in all eight 365 days (reduced to 180 days in 1990), the sharp countries, there was no clear relationship between the increase in provisions from 51 percent of total non- size of fiscal or balance of payments deficits and non- performing loans in 1986 to 66 percent by 1990 performing loans, partly because of major problems brought the net exposure to nonperforniing loans to in standardizing the measurement of fiscal deficits only 7.1 percent of total loans by 1990 (Yusof and and nonperforming loans. In the United States data others 1994). The banking system was also required on nonperforming loans are stringently applied, and to increase capital and provisions through rigorous industry averages for nonperforming loans, at I to 5 supervision and, where necessary, injectioni of capital percent of total loans, are very low by international by the central bank (chapter 7). Once the economy standards. Nevertheless, more U.S. banks failed, part- began to recover in 1987-89, the proportion of non- ly because of the large number of banks in the United performing loans began to decline rapidly. States (29,000) and partly because the system is designed to merge or liquidate problem banks much Who is to blamie? faster than in other countries, as a deliberate exit poli- cy. Thus, despite the well-known problems of the Controversy still persists over whether bank failures thrift industry and concern in certain pockets of the have been due to bad bankers or the bad economic industry about the fragility of the banking system, environment. Bankers are likely to blame bad govern- the safety and soundness of the U.S. banking indus- ment policies, banking system supervisors are apt to try as a whole is not in question. 12 BANK RPSTRUCTURING Table 1.6 Banking problems in selected countries (percent unless otherwise specified) Indicator Argentina Chile Colombia Ghana Malaysia Spain United States Yugoslavia Picriod of hankinig probLems 1980-90 1974-87 1982-88 1983-89 1985-88 1977-85 1981-91 1983-90 I:ak declinc in DI)P -9.8 -14.1 (0.9 -4.5 -1.0 -0.2 -2 5 -7.6 1'cak curreiit accouiit defilct/(',DI' -8.5 -14.5 -7.5' -4.3 -6.0 -2.6 -3.6 -5.4 P'eak fiscal dficbit/GDP -15.9 -3.0 -4.8 -2.7 -11.2 -6.9 -6.0 -0.1 Peak inflation rare (annlual) 4,923.6 30.7 24.1 122.9 5.8b 24.5 13.5 1.239.9 Peak dccline in ctrises of trade -7.9 -29.3 -29.6 -18.5 -18.2 -14.9 -34.4 -7.2 l'cak real itsreresr atc (lenIdinig) 44.8 56.9' 16.1 9.8 12.0 5.0 18.9' 1.539.9 I'ak non nperfornming Iloans as a Ipercentrage of soial loan s 30.3 18.7' 25.3 39.5' 32.6 n.a. 3.7. 27.69 -40.0 Real estate losses Yes Yes Yes Yes Yes Yes Yes No Forcign cxchanige loswse Yes Yes Yes Yes No No No Yes Cossiiected lending Yes Yes Yes Yes Yes Yes Yes Yes LxLessive govcrnnment borrowing Yes No No Yes No No No No Stare-owned enterprises Yes No N-lo Yes No No No No Numjiber of banks 168 financial 13 banks and 7 9 4 banks. 52 2,203 69 iissolvset inistitutions 6 financieras 4 finance companies, and 32 deposit-taking cooperatives Restructuring Forced Swap for Guarantee Swap for Central bank Swap for Deposit Across-the- nclibod deposit cenitral fund central capital unaranree itisurance board and coiiversion bank bonds banik bonds injection fund bonds liquidationl case-by-case andi merger A'oi,: Reported peaks are during period of banking problems. a. (Cuirrenit accousts deficit/GNP. 1P. 1982. before sIte crisis. Itflatioti was belowv I percent during nmost of the crisis period. -. Yearly average. d. Prine lendinig rate, sot adjusted for inflation. c. Comtmercial bank loan defaults as a percenctage oftotal loan portfolio. t provisions only. I iDl(-inSlArd comiimercial banks in 1990 and FSLIC-insured thrfits it 1988 (no data available for suibsequenit years, which were worse). %ounr: Morris 1')90; Velasco 1991; chapters 4-11. The size of the published fiscal deficit -an also External shocks probably triggered banking be misinterpreted. Although Yugoslavia's central problems in Chile, Colombia, Malaysia, and go%ernment fiscal deficit in the 1980s was almost Yugoslavia. Chile suffered significantly from a negligible, the quasi-fiscal deficit in the books of the decline in its terms of trade, mainly because of its Nationial Bank reached 11.8 percent of gross social heavy reliance on copper exports. Colombia suf- product in 1986 (Gaspari 1989). The fiscal deficits fered a decline in terms of trade that exacerbated its of Argentina and Chile would similarly have banking problems during 1982-88, which were increased had they included the quasi-fiscal deficits triggered in part by the cutoff of external resources of the central bank in absorbing the foreign caused by the Mexican debt crisis of 1982. exchange losses of the private and public sectors, as Malaysia also suffered an across-the-board decline well as losses incurred in bank interventions. By the in its terms of trade in 1983-86, when the prices of end of 1986 the Central Bank of Chile had "carved its major commodities-tin, rubber, oil, and palm out" bad loans from banks equivalent to 20 percent oil-declined simultaneously. Yugoslavia was hit of GDP (chapter 10). hard by oil price shocks. Banking Fragility in the 1980s: An Overviev 13 Table 1.7 Banking crises and private sector "wealth" capping of bank spreads and tough action by the cen- (change in U.S. dollars as a percentage of GDP) tral bank in preventing further lending to nonper- Decline forming borrowers curbed interest rate growth. in bank Another common feature was losses in lending to Crisis Private Decline credit to the real estate. Caprio, Atiyas, and Hanson (1994) have Countiy peak year 'Yoss" in M3 prvate sector suggested that banks tended to expand their real Argentina 1981 6.2 13.8 9.1 estate lending immediately after financial sector liber- 1989 7.3 26.7 15.8 alization or relaxation of lending guidelines. Colombia 1985 4.0 5.1 2.3 Information asymmetry or bank myopia may be Chile 1982 28.3 14.3 31.2 involved in lending to real estate. First, banks assume Ghana 1983 20.3 22.6 2.3 that collateral value alone, particularly real property, Malaysia 1985 13.9 5.0 6.0 is sufficient to demonstrate good credit, instead of Spain 1982 6.4 9.0 7.0 assessing the underlying cash flow capacity of real United States 1981 5.7 4.0 4.0 estate developers to service their debt. Because devel- Yugoslavia 1988 0.6 7.2 6.6 opcrs assume that debt can always be serviced out of Note: Because changes in inflation and exchange rates make measuring th rentals or property sales, they forget that the "lumpi- decline in "real" wealth difficult, moncy stock and private sector credit have ness" of real property and the thinness of markets can been converted into U.S. dollars at the prevadiing rate. Private losses are create situations where property cannot be sold estimated as the decline in the value of stock market capitalization, plus the decline in real value of the money stock, less the decline in real value of bank except at massive losses. credit to the private sector. Second, there is a fallacy of composition problem in real estate lending. Each developer assumes that In the United States terms-of-trade changes his or her project is good at the margin, but forgets affected different regions. The savings and loan crisis (as does the banker) that if all developers were to was sparked by high interest rates in 1980-81, make the same assumption, there would be such an which, because of the institutions' fundamental inter- oversupply of property that prices would fall sharply. est rate and maturity mismatch, caused their eco- Information on cumulative property construction nomic insolvency. In the early 1980s the farm belt and its oversupply generally is not available to the was hurt by declining exports brought on by the high market until the economy turns downward, by which value of the U.S. dollar. In the mid-1980s the energy- time it is too late. producing states, particularly Texas, were hurt by the Third, both bankers and real estate borrowers have decline in energy prices. By the second half of the frequently been deluded by the initial high returns on 1980s excessive investment in real estate, stimulated property. Bankers forget that property "booms" are by tax incentives and by consumer hedging against sometimes fueled by the increase in bank credit pro- inflation, caused large bank losses in the Northeast. vided and that the oversupply can result in large loan One phenomenon common to all eight countries losses. However, because property price cycles are long, was the close association between high real interest the next generation of bank credit managers may rates and banking problems. In the United States repeat the same mistake. annual real interest rates in the 1980s were 3 to 6 per- Fourth, speculative "bubbles" in real assets have cent, significantly higher than in the 1960s and been closely associated with financial liberalization, 1970s. In Argentina and Chile high annual real inter- particularly deregulation in bank lending, interest rate est rates of 30 to 50 percent were the norm rather decontrol, and the opening of the capital account than the exception during the debt crisis. At such high (Park and Park 1992). In the absence of high returns real interest rates, borrowers quickly became decapital- in tradables, new funds seek shelter in nontradables, ized. The result was "to transfer the ownership of real such as land. Before deregulation, cartelized banks enterprise wealth from debtors to creditors, a mecha- earned protected profits. After liberalization, banks lost nism doomed to failure when no more shareholders' their market share to capital markets and, faced with wealth was left" (Diaz-Alejandro 1985, p. 16). When such downsizing, increased their risk appetite, "a strate- the central bank itself had to pay such high real inter- gy that was also encouraged by the existence of a net of est rates in attempts to maintain monetary stability, explicit and implicit government guarantees that both the result was unsustainable growth of internal debt, protected depositors and made 'failure' a less credible as occurred in Argentina and Chile. In Malaysia the deterrent to excessive risk taking" (IMF 1993). 14 BANK RESTRUCTURING In almost all countries connected lending In almost all cases there was clear agreement on between banks and their shareholder-managers was the need for better supervision and more transparent found to be a problem. In Argentina, Chile, and legal and accounting frameworks. Yet the evidence Spain the presence of economic grupos that owned from this survey of bank failures is that, so far, even banks was a major factor. Yugoslavia's banking prob- the strongest supervision available in the OECD lems were inextricably linked to the nation's experi- countries has failed to prevent bank failures. Clearly, mentation with socialized ownership of banks and improved supervision and banking laws are necessary enterprises, in which worker councils owned groups but are not sufficient to prevent bank failure. of enterprises, including the banks that served as trea- Evidence suggests that while fraud and bank mis- suries for such groups. Consequently, socialized management were responsible for many individual banks were unable to exercise financial discipline (the bank failures and losses, macroeconomic factors such hard budget constraint) on their borrower-owners, as external shocks, policy mistakes, and inadequate risk leading to considerable distress borrowing. In management at all levels-institutional, sectoral, and Malaysia and the United States, where regulations national-created the conditions of financial imbal- limited the amount of connected lending, the dam- ance that led to widespread bank distress. No unique age was felt only in areas where supervision was weak, set of factors, macroeconomic or microeconomic, cre- as in the case of Malaysian deposit-taking coopera- ated the distress, nor did bank failures happen tives and in U.S. savings and loan institutions. overnight. Many policymakers failed to correct key Institutional and structural issues also led to structural defects, particularly legal impediments to bank problems. In the United States legal and struc- geographical or sectoral distribution of risks or inher- tural issues of geographical and sectoral segmenta- ent interest rate risks, before liberalizing the banking tion prevented banks and thrifts from diversifying system. In Argentina, Chile, and Uruguay, with de their risks adequately. Competition among more facto state guarantee of the deposit base, it was wide- than 29,000 banks and thrifts and such newcomers spread moral hazard behavior by banks, worsened by as money market funds, corporate bond markets, the presence of connected lending to economic grupos and foreign banks led depository institutions to take that controlled banks, that generated large losses in the risks in real estate, highly leveraged securities, and banking system. Moreover, the severe imbalances in developing country debt without adequately pricing these economies, such as the large fiscal deficits such risks. In Japan financial liberalization allowed financed by excessive external debt, were entwined banks and nonbank financial institutions to finance with the political economy and so were beyond the the asset bubble. Fortunately, the economy was powers of bank supervisors to control. strong enough to withstand the large asset losses that Trade liberalization, price reforms, and industrial came with deflation of the bubble. policy changes also generated large losses to banks, In Chile and Spain the entry of new financial which had lent heavily to support inefficient state- institutions without adequate separation of ties owned enterprises or domestic private sector firms between owners of banks and enterprises created that borrowed heavily to finance investments behind overgearing of the private sector, which required a high tariff barriers. In Africa and Eastern Europe "shakeout" when the economies went into a reces- rapid trade liberalization exposed the large inefficien- sion. Large fiscal deficits were clearly the source of cies of state-owned enterprises, and their losses were problems in Argentina, Chile, and Malaysia. In quickly passed on to their creditor banks. In Ghana, Argentina and Chile the central banks' absorption of for example, trade and price reforms made state- private external debt complicated the conduct of owned enterprises uncompetitive, and they were monetary policy. Chile's and Malaysia's ability to unable to repay their debts to the bank. reverse their fiscal deficits generated enough In sum, banking problems in the 1980s resulted resources to stabilize their economies and turn from a combination of bad policies, bad manage- around the banking problems. Ghana's banking ment, and a bad institutional framework. problems were the legacy of years of financial repres- sion and economic decline. As in other socialist Banks as intermediators ofeconomic loss economies, banks were state owned and controlled, lending mainly to finance the budget or state-owned Although banks are seen primarily as intermediators enterprises (chapter 8). of savings and investments, it is useful in the context Banking Fragilty in the 1980s: An Overview 15 of financial crises to treat banks as the allocators of The inflationary impact of bank lending to loss- loss in an economy. The coincidence of widespread making enterprises was first grasped by Schumpeter bank losses throughout the world cannot be attrib- (1934): uted solely to incidental speculative or fraudulent behavior by bank management. Fraud and misman- This loss always occurs if the entrepreneur agement occur when conditions of incentives are dis- does not succeed in producing commodities torted; that is, when speculative risks are privately at least equal in value to the credit plus inter- rewarded and socially absorbed, and enforcement est. Only when he succeeds in so doing has against fraud and criminal misappropriation is lax. the banlc done good business-then and Sound supervision and tough enforcement can pre- only then, however, is there also no inflation. vent cases of individual fraud and mismanagement at the margin but may not be able to prevent systemic The work of Sundararajan and Baliino (1991) or economywide losses from being absorbed by the empirically confirmed the effect of banks increasing banking system. credit before and after a financial crisis. They found Wealth losses for the private sector in peak crises that total domestic credit increased in real terms, years can be very large. For the eight countries stud- despite falls in real output and foreign exchange ied here they varied from 0.6 percent of GDP for reserves. Yugoslavia in 1988 to 28.3 percent of GDP for Chile The deterioration of bank portfolios and their res- in 1982 (see table 1.7). Even these broad numbers cue by central banks or the state have large monetary can be deceptive, because the household sector in and fiscal implications because of feedback effects Yugoslavia lost the equivalent of 7.2 percent of GDP that lead to a vicious circle of macroeconomic insta- in the real value of money holdings through hyperin- bility. As demonstrated in the cases of Argentina and flation and devaluation, while the socialized enter- Chile, large current account and fiscal deficits prise sector gained from the depreciation of the real required devaluation as part of the package of macro- value of its debt. The Chilean case was the most economic stabilization. Because enterprises and the severe because of the massive decline in GDP (14.1 public sector had large external debt, however, the percent) and the sharp jump in the real value of pri- borrowers suffered large foreign exchange losses from vate sector debt (31.2 percent of GDP) due to devaluation, with an immediate impact on the banks' unusually high real loan rates averaging 42.4 percent portfolios. Central bank lending to rescue insolvent a year in 1981-82 (Velasco 1991). banks injected excessive reserve money into the sys- Such wealth "compression" or shocks were clearly tem. A central bank can become insolvent if it too large for the private sector net equity to absorb acquires liabilities of greater market value than the and were therefore transmitted to the banking system capacity of its seigniorage to service. At that point the in the form of bad debts. Banks transmit their own central bank can service its liabilities only by acceler- losses in the economy through additional credit. ating inflation (Fry 1991). Alternatively, the govern- They attempt to recover losses in two ways: by gener- ment can run higher fiscal deficits to finance the ating further credit or by widening spreads. The first recapitalization of banks by borrowing from the cen- approach involves additional lending or investments, tral bank. Either method increases money supply, and in the hope that new credits will earn sufficient prof- unless there is accompanying tight monetary mea- its to cover existing losses. Credit expansion often sures, domestic inflation exceeds international infla- includes the extension or rollover of bad loans to loss- tion, leading to further devaluation and enterprise making borrowers to cover up the extent of the dam- losses (figure 1.2). age of nonperforming loans (evergreening). Just as The second transmission mechanism of bank loss- distressed borrowers are willing to pay higher interest es is the widening of spreads. Banks must raise their rates to maintain liquidity in the face of deteriorating spreads to cover the nonaccrual of income from non- solvency, insolvent banks are apt to raise deposit rates performing loans. If the writeoff of nonperforming to attract funds to maintain their own liquidity and a loans (provisioning) is also included, spreads have to facade of solvency. Unless stopped, these banks widen further. Thus if nonperforming loan levels increasingly will channel scarce resources to loss-mak- reach 20 percent of total loans, spreads might have to ing enterprises in an economy, thus distorting widen from an average of, say, 4 percentage points to resource allocation further. as much as 12 percentage points (figure 1.3). Spreads 16 BANK RESTRUcruRlNG Figure 1.2 The vicious circle of financial distress experience foreign exchangeand Devaluation occurs J\ Th fscal impact Commerca creates a need for bank ) - - banks are recapitaliztion dccapitalzd Inlto ises higher than theworld rate \ (dCentral bank b rl are increased either by reducing deposit rates, which distressed borrowing, bank runs, capital flight, and leads to disintermediation from the banking system, monetary and price instability. Consequently, bank or by raising lending rates to such punitive real levels restructuring is a key element in achieving sustainable that even good borrowers have incentives to become and stable growth. But as Hinds points out, bank bad borrowers. restructuring makes sense only when coupled with a restructuring of the real sector in which inviable firms A Framework for the Resolution of Bank Distress are closed and viable but troubled firms are restruc- tured. The problems faced by banks in Eastern The pioneering work of Long (1987) and Hinds Europe have brought fresh insights to these views and (1988) has cogently argued that macroeconomic new challenges to the process of bank restructuring. imbalances, such as large fiscal and balance of pay- ments deficits, sharp changes in relative prices, exter- A stock-flow-consistent approach to bank restructuring nal shocks, and policy errors can lead to weak finan- cial structures, large portfolio losses, and weak bank The relationship between bank problems and the management and supervision. These factors in turn problems of the real sector is probably best depicted reduce the efficiency of resource allocation, causing in a stock-flow-consistent matrix of sectoral and national accounts. A stock-flow-consistent analytical framework is used to examine the causes and effects Figure 1.3 Spreads and nonperforming loans because bank losses are not just flows but also evolve Spreads (percent) from changes in relative prices and their impact on the portfolio (stock) wealth of asset or liability hold- ers. In this framework financial imbalances in one lo / sector of the economy show up as potential losses in other sectors, causing major portfolio shifts that can 8 - disrupt and destabilize the economy. 6 Consider this example, using the flow of funds accounts for the Republic of Korea in 1990 (table 4 1.8). The economy is divided into four domestic sec- Spread_with_wri__off_of nonperforming loanstors and one external sector. Deficits of expenditure Spreadwithtwriteoffofnonperformingloans over income in a sector are financed either by the 2 - Spread without writeoff of nonperforming loans dadw fast rb orwn rmohrsc drawdown of assets or by borrowing from other sec- tors. The deficits of the enterprise sector (17.0 per- 0 5 10 15 20 cent of GDP) are financed by the surpluses of the Nonperforming loans as a percentage of total loans household sector (11.4 percent), government (3.7 Banking Fragility in the 1980s: An Overview 17 Table 1.8 Sectoral and national financial balances, based on the Republic of Korea's financial balances, 1990 (percentage of GDP) Account Households Enterprises Banks Government Domestic total Externalfinds National total Revenues 22.4 12.4 39.7 4.3 78.7 2.1 80.8 Expenditures 11.0 29.4 39.5 0.5 80.4 0.4 80.8 Surplusordeficit 11.4 -17.0 0.1 3.7 -1.7 1.7 0.0 Financed by: Buildup or drawdown of assets -11.4 -11.5 -3.7 -26.7 -1.7 -28.4 Borrowing 17.0 11.4 28.4 28.4 Source: Bank of Korea 1992. percent), banking (0.1 percent), and the external sec- system that does not address the financial imbal- tor (1.7 percent). ance of the corporate sector would only invite Sectoral imbalances arising from a variety of caus- another financial crisis. es thus need to be financed, and financing restrictions * Domestic losses cannot be passed on to the external force adjustment in the economy due to the stock- sector unless there is external debt forgiveness or uni- flow consistency of the national accourts. For exam- lateral debt repudiation. In fact, the cutoff of ple, imbalances in the economy can be created either external funds resulting from the 1982 Mexican by large fiscal deficits, requiring financing by the debt crisis stopped the growth of unsustainable banking system or externally, or by excessive spend- debt accumulation by both the enterprise and the ing by the enterprise or household sectors. These government sector in many developing countries imbalances can be caused by changes in policy or rel- in the 1980s. The net external resource transfer ative prices, by external shocks, or even by bad enter- from servicing external debt created a large inter- prise and bank management (box 1.1). nal debt problem in a number of countries when The stock-flow consistency of this framework governments and enterprises were initially unwill- helps to conceptualize the diagnosis process by iden- ing to reduce consumption to bring the imbal- tifying the sources of financial imbalance that caused ances to sustainable levels. the losses in the banking system. The framework also * Wealthholders disintermediatedfirom domestic bank- helps policymakers think through the issues of loss ing systems or engaged in capitalflight to avoid losses allocation, because bank losses would have to be from potential bank failure as well as from a poten- borne sooner or later somewhere in the economy. tial wealth tax to pay for the national debt over- Unfortunately, not enough data exist to construct hang. This not only narrowed the domestic infla- this matrix for any actual country. Sectoral balance tion tax base, which worsened the fiscal deficit, it sheets for the enterprise sector and the government also worsened the balance of payments. As is now sector are not normally compiled in national income well known, any central bank liquidity support accounts. Even if government and enterprise balance during a banking crisis in an economy with an sheets could be compiled, serious inconsistencies open account only results in loss of foreign would occur because governments operate on a cash exchange reserves. accounting basis while banks and enterprises usually * An insolvent sector cannot rescue another insolvent prepare accounts on an accrual basis. Moreover, sector. Stability in the system requires that surplus value-added flows are not consistent with the size of resources be found to pay for the deficits in other changes in the sector balance sheet items. sectors if the external sector is unwilling to lend The stock-flow consistency of this model, however, new resources. In Chile and Malaysia, for exam- helps to clarify several issues in bank distress that were ple, the supply response arising from dzvaluation, sometimes confused in partial equilibrium analyses: liberalization in the real sector, and fiscal Bank restructuring cannot be executed independently retrenchment generated sufficient growth from of enterprise restructuring and budget reform. renewed foreign direct investment and a current Because of implicit or explicit state guarantees on account surplus to alleviate problems in the the deposit base, bank losses ultimately become banking system. In Argentina and Yugoslavia, quasi-fiscal responsibilities of the state. At the where normal seigniorage of the central banks was same time, any recapitalization of the banking insufficient to service the large burden of net 18 BANK RESTRUCTURING Box 1. I Rewriting the national balance sheet The effects of bank losses can be demonstrated from Consequently, the government may have to absorb the the sectoral and national balance sheet presented in bank losses by either guaranteeing the banking system or table 1.8. Suppose that in year one, growth slows and recapitalizing it through an issue of bonds in exchange the enterprise sector begins to default on loans, equiva- for the losses. Note that the losses have de facto become lent to 10 percent of its loans. In the table loan losses a quasi-fiscal deficit. Macroeconomic stability then would be equivalent to 1.7 percent of GDP (10 per- depends on whether the government has the debt-servic- cent of enterprise borrowing of 17 percent of GDP). ing capacity to service its increased debt burden, equiva- Assuming that these are all domestic loans, the losses lent to 5.1 percent of GDP. Assuming that the nominal would fall on either the banking system, as the major interest rate is 10 percent, the government must bear an lender, or the government. In either case the surplus of additional debt-servicing burden equivalent to 0.5 per- the banking or the public sector would have to adjust cent of GDP. If the government is unable to borrow by 1.7 percent of GDP domestically or internationally to finance the higher If the losses are not recognized because of defective internal debt, then it may have to resort to either higher accounting, then all the losses would be reflected in the taxation or monetary creation. books of the banking system as impaired assets. These Recent experience with loan losses in three are financed by bank deposits. If the rate of losses carries Scandinavian countries indicated that loan losses aver- on for three years, the banks would have carried aged 5.5 percent of GDP between 1991-92 and that the impaired assets equivalent to 5.1 percent of GDP. If at government provided support to the banking system that point the household sector loses confidence in the equivalent to 3.3 percent of GDP banking system and engages in capital flight, domestic The national balance sheet and financial flows savings would fall and the balance of payments would framework suggest that both stock and flow losses have worsen, causing a further decline in domestic growth to be taken into consideration. Stock-flow consistency and worsening the enterprise losses and bank loan losses. suggests that there is no free lunch. If the external sector At that point the government can only stabilize the will not bear losses, then domestic bank losses must be situation by raising domestic real interest rates (to stem borne elsewhere in the economy, initially in the public capital outflows) or by recapitalizing the banking system. sector because of either government action or deposit A market-based solution is possible if the failed banks insurance, but ultimately through higher taxation or are allowed to collapse, passing the losses on to the monetary creation. depositors. But this would worsen domestic savings and encourage disintermediation, including capital flight. Source: IMF 1993. foreign exchange liabilities, the central banks Nowhere was the interlocking relationship among themselves became major sources of monetary banks, the budget, and enterprises so evident than in instability, which eventually led to hyperinflation. the formerly socialist economies of Central and Eastern Europe. Former President Gorbachev of the Bank restructuring, enterprise restructuring, andfiscal Soviet Union used to say that the workers pretended reform in transition economies to work, and the state pretended to pay them. The institutional structure of the centrally planned econo- The stock-flow-consistent framework suggests that my is best portrayed in the "troika" model of budget, losses in one sector must appear in or be financed by enterprise, and banking system, where every element other sectors. Thus the unwillingness of one sector to is owned by the state (figure 1.4). The state based its continue financing inefficient behavior in another sec- budget revenue almost totally on resources extracted tor could trigger massive balance sheet and flow adjust- from the enterprises and the banking system. In 1989 ments. The rapid growth of domestic debt in many tax collected from state enterprises (including sales highly indebted countries after recourse to external tax) accounted for 89 percent of Soviet federal rev- debt was shut off in the 1982 debt crisis demonstrated enues. Prior to reform, enterprises extracted monop- how enterprises, governments, and banks all tried to oly profits from consumers and banks extracted mobilize internal resources to meet excessive expendi- monopoly interest revenue from enterprises because ture commitments and to service their external debt. enterprises were locked in to designated state banks.' Private wealthholders, sensing that heavy taxation was Moreover, because enterprises acted as tax agents, inevitable in such an environment, stashed their capital there was no parallel revenue-collecting machinery to abroad to avoid either a wealth or an inflation tax. extract excise or consumer taxes (McKinnon 1991). BankingFragility in the 1980s: An Overview 19 Figure 1.4 The troika model increases. In China there was a clear positive correla- tion between decentralized enterprise "ownership' BUDGE T ' ut and productivity. The average annual growth rate of Tax and borrow \\nor prots) labor productivity during 1980-87 was 5.2 percent fro banks \\ Subsidize for state-owned enterprises, 12.3 percent for urban Allocate credit loss-makers collectives, and 18.0 percent for rural town and vil- according to plan lage enterprises. "BANKS" { . > ENTERPRISES Moreover, as prices remained tightly controlled, Captive deposits Ilnterenterpnse losses began to increase. In Yugoslavia current losses n1 debt of enterprises escalaved from 0.2 percent of gross Enierprises social product in 1984 to 14.3 percent by 1989 (National Bank of Yugoslavia 1990). In Romania The old Soviet accounting system, vlhich was losses of state enterprises were as much as 19 percent designed to check compliance with plan rather than of GDP in 1989 (IMF 1990). In China as many as measure the profitability or solvency of enterprises, one-third of state enterprises were reportedly incur- tended to overstate profits (measured according to ring losses in 1991. These losses were largely reflect- internationally accepted accounting standards) by 12 ed in the books of the banks that lent to these enter- to 15 percent, due mainly to underdepreciation and prises. Moreover, because there was no inflation underallocation of costs (Enthoven, Sokolov, and accounting during this period of high inflation, the Petrachkov 1992). government was overtaxing the inflation profits of Liberalizing prices, deregulating the real sector, the enterprises, especially through the sale of old and moving the banking system into a two-tier struc- inventory at historic costs. Even sound firms with ture had the effect of eroding the tax base in most surplus cash kept in banks were subject to severe transition socialist economies, particularly in terms of inflation tax, because such demand deposits paid no the enterprise tax. For example, the Chinese tax sys- interest. tem allows the repayment of loans as a deduction on Bank credit is not the only source of credit for enterprise income tax. Between 1978 and 1988 total enterprises. Because of suppliers' connections among revenue from enterprises in China fell from 60 per- enterprises, interenterprise credit is often more cent of government revenue to 25 percent, while gov- important than bank credit. Loss-making enterprises ernment revenue fell from 34 percent of GNP to 21 have often relied on credit from other enterprises to percent (World Bank 1990). Subsidies to loss-making survive. Central bank attempts to tighten bank credit enterprises also became a heavier burden. The net often have had the effect of temporarily increasing revenue extracted from enterprises fell from 20 per- interenterprise credit. In Yugoslavia interenterprise cent of GNP in 1978 to only 2 percent in 1988. credit rose from about 14 percent of total domestic Giving enterprises greater autonomy in the liber- credit in 1980 to more than 33 percent by 1988. In alization phase also eroded monopoly profits and Poland interenterprise credit in June 1990 (at a time encouraged enterprises to evade taxation through of tight liquidity) was one-and-a-half times larger spending on inventory, higher wages, and higher than bank credit. Consequently, even sound enter- investment. The budget compensated for the loss of prises can be dragged down because of losses on revenue by shifting the burden of investment financ- uncollectible interenterprise credit. ing to banks and to the retained earnings of enter- Other than the inheritance of enterprise loans at prises. During 1986-89 budget financing of invest- overstated book values, banks in transition economies ments of Polish enterprises fell from 4.5 percent of also suffer from the following structural and institu- GDP to 2.9 percent, while borrowing from banks tional flaws: rose from 6 percent to 20 percent of GDP * Geographic and sector concentration. Almost all the The troika system was stable as long as enterpris- banks in Central and Eastern Europe suffer from es were relatively efficient and output growth was geographic and sector concentration in their loan expanding. But state-owned enterprises under cen- portfolio. This has exposed them to large shocks tral planning were inefficient-they invested in out- arising from the collapse of the Council for moded equipment, geared products toward the Mutual Economic Assistance market or from domestic market, and granted excessive wage downturns such as a decline in agriculture. 20 BANK RESTRUCTURING *Foreign exchange mismatches. Yugoslav and Polish reliable payments and fund management com- banks had large foreign exchange mismatches in pound the problems of the enterprises. their books because all foreign exchange proceeds In market-based economies the relationships were surrendered to the national bank, while the among banks, enterprises, and the budget have not banks collected foreign currency deposits from been as sharply defined as in the troika, but variations residents. Unless the national bank is able to guar- of the negative features listed above have been evident. antee the liability exposures, such banks are high- The prevalence of economic grupos in Chile, ly vulnerable to exchange rate depreciations. Colombia, and Spain, for example, created connected * Shareholder-borrowers. In Hungary, Yugoslavia, iending and a concentration of power that con- and the states of the former Soviet Union enter- tributed to banking problems in these countries. prise borrowers are also significant shareholders in As the stock-flow-consistent framework demon- banks. This was a major problem in Yugoslavia, strates, the objective of bank restructuring is thus to and led to uncontrolled lending (chapter 9). rewrite the national balance sheet and profit and loss There are now more than 1,800 banks in Russia, accounts such that economic efficiency, sound credit many of which are owned by borrower-enterprises allocation, and macroeconomic stability are restored and operated under an inadequate supervisory as an integral part of a national program. framework. The political challenge of bank restructuring lies in * Segmentation of credit and deposit markets. Most the difficult area of loss allocation. Bank losses are ulti- banks are still segmented into the enterprise mately reflections of real sector losses in the enterprise financing market and the retail deposits and and budgetary sectors, compounded by inefficiencies housing finance markets. Household deposits are or fraud in the process of financial intermediation. The largely concentrated in national savings banks, success of bank restructuring depends on the ability to which lend substantially to finance housing at allocate such losses to the rest of economy without suf- subsidized interest rates. With interest rate liberal- fering macroeconomic instability. ization, most of these specialized banks with large portfolios of below-market yields became techni- The Process of Bank Restructuring cally insolvent. These banks are only gradually diversifying their liabilities into the retail market. Like financial sector liberalization, bank restructuring * Inadequate loan loss provisions. Because of the is a process, not an event (Caprio, Atiyas, and strain on their budgets, most authorities in the Hanson 1994). The techniques of this process are transition economies are unwilling to concede to well defined, but their application depends on indi- banks pretax provisioning for bad debts. vidual country conditions. The process of bank Consequently, these banks generally have negligi- restructuring may be distilled into four main phases, ble loan loss provisions. some of which may overlap: * Inadequate legal and supervisory framework. Most * Diagnosis. transition economy banks do not have a strong * Damage control. and clear legal framework to enforce financial dis- * Loss allocation. cipline through debt recovery ancl liquidation * Rebuilding profitability and creating the right procedures. Bank supervisory authorities are incentives. unable to enforce discipline on banks due to the The process of bank restructuring operates in a lack of strong banking laws and the supervisory dynamic context, where techniques of restructuring capacity to carry them out. can play major roles in changing the incentives with- * Lack of banking skills and outdated systems and pro- in the economy. Chapter 2 discusses the problems of cedures. The inadequacies of accounting, legal, and diagnosis and loss allocation, which lie more in the supervisory frameworks all underscore the dire realm of political economy. Chapter 3 focuses on shortage of human capital. Bank skills in credit and supervisory remedies for bank restructuring, the issue project evaluation are grossly inadequate. The inef- of damage control, and the building of safe and ficiencies of the banks in providing speedy and sound banking systems. Banking Fragility in the 1980s: An Overview 21 Annex 1.1 Bank problems in the 1980s Africa Ghana In 1989 the government recapitalized ten state-owned and -controlled commercial banks rhat had suffered large foreign exchange losses and nonperforming loans. A nonperforming asset recovery trust was established to dispose of bad debt. Guinea In 1985 the government closed six state-owned specialized banks, with losses amounting to almost 97 percent of total assets. Kenya In 1989 eight failing institutions were merged into a "turnaround" bank, Consolidated Bank Ltd. The Kenya Deposit Protection Fund also was established. In 1992 the central bank had to intervene in two local banks. Madagascar In 1988, 25 percent of all loans were deemed irrecoverable and 21 percent more deemed "difficult to collect," compared with capital and reserves of less than 5 percent of assets. Mauritania In 1984 the five major banks had nonperforming assets ranging from 45 to 70 percent of their portfolios. The cost of rehabilitation was estimated at about 15 percent of GDP in 1988. Senegal Between 1988 and 1991 six commercial banks and a development bank were closed. A single asser recovery body took over the nonperforming assets and corresponding liabilities of the liquidated banks. Tanzania In 1987 the main financial institutions had arrears amounting to half their portfolio and implied losses were near- ly 10 percent of GNP. Asia Bangladesh In 1987 four banks accounting for 70 percent of total credit had an estimated 20 percent nonperforming loans. Hong Kong Between 1983 and 1986 the Commissioner of Banking had to take over or liquidate seven banks and deposit- taking companies. Malaysia Between 1986 and 1988 the central bank intervened in thirty-two deposit-taking cooperatives, four commercial banks, and four finance companies. Nepal In 1988 reported arrears of three banks accounting for 95 percent of the financial system averaged 29 percent of all assets. Philippines Between 1980 and 1987 the Central Bank closed 173 banks. The government took over $5.1 billion of foreign debt of the two largest banks, which were subsequently partially privatized. Sri Lanka State-owned banks, which make up 70 percent of the banking system, have estimated nonperforming loans of about 35 percent of the total portfolio. Thailand During 1983-84 fifteen finance companies and securities companies with assets of B 9.8 billion went under, and the rescue scheme cost B 8 billion. Between 1984 and 1987 three commercial banks had their capital reduced and recapitalized with assistance in low-interest loans from the central bank. The Thai Financial Institutions Development Fund was established to assist rehabilitation of problem institutions. Latin America Argentina Between 1980 and 1989 the Central Bank intervened in ninety-three institutions, of which only seven were reha- bilitared or sold. In 1990 more than 200 banks were in the process of liquidation. The restructuring process is proceeding. Bolivia During 1986-87 five banks were liquidated, and nonperforming loans reached an estimated 30 percent of total loans in 1987. Chile In 1981 the government liquidated eight insolvent institutions that held 35 percent of total financial system assets. In 1983 another eight institutions were taken over, accounting for 45 percent of system assets. Colombia Between 1982 and 1987 the central bank intervened in six banks, accounting for 24 percent of system assets. Costa Rica In 1987 public banks accounting for 90 percent of loans considered 32 percent of loans "uncollectible." Uruguay Between 1984 and 1987 the central bank intervened in five domestic banks and de facto "nationalized" 75 per- cent of total deposits. 22 BANK RESTRUCTURING Annex 1.1 Bank problems in the 1980s (continued) Industrial countries Australia During 1989-90 two large state banks received capital injections from the government to cover loan losses. Canada Two small provincial banks failed in 1985, following the failure of a number of trust and loan companies. Finland The central bank took over control of the Skopbank, the apex bank for the Finnish savings banks, in August 1991. Several banks also suffered losses due to bad loans and share investments. Japan Banks have suffered from the sharp decline in the stock market (down 56 percent from its peak in 1987) and dedines in real estate prices, with official estimates of nonperforming loans at Y 12.3 trillion ($120 billion) as of September 1992. Of this, Y 4 trillion was not covered by collateral. Several small trust banks had to be taken over by stronger banks. In August 1992 rhe government announced Y 10.7 trillion ($86 billion) of fiscal stimulation. The program included the endorsement of the establishment of a private entity (owned jointly by the financial institutions) to buy nonperforming loans of banks with real estate collateral. Norway The Central Bank provided special loans to assist six banks suffering from the recession that followed the oil bust of 1985-86, particularly from problem real estate loans. The state took control of three of the largest banks, part- ly through a government bank investment fund (NKr 5 billion),and the state-backed bank insurance fund had to increase capital to NKr 11 billion. Spain Between 1978 and 1983 fifty-one institutions holding one-fifth of all deposits were rescued. In 1983 the govern- ment nationalized Rumasa, a holding company controlling more than 100 enterprises and 20 small and medium- size banks. Sweden In 1991 the government had to inject SKr 5 billion ($800 million) into the state-controlled Nordbanken, and to guarantee a $609 million loan to save the largest savings bank. United Kingdom The Bank of England acquired Johnson Mathey, a key player in the gold market, recapitalized it, and subsequent- ly sold it. The Bank of Credit and Commerce International, onc of the largest bank frauds in history, was closed in 1991. United States Between 1981 and 1991 more than 1,400 savings and loans and 1,300 banks failed. The thrift crisis may have cost between $315 billion and $500 billion. The Resolution Trust Corporation acquired $357 billion of bad assets of thrifts by the end of 1991 and had disposed of $228 billion. The government granted $70 billion to strengthen the Federal Deposit Insurance Corporation. Transition economies Bulgaria Nonperforming loans are reported to have exceeded 60 percent of assets. Hungary The government has included guarantees in the state budget to cover a portion of nonperforming loans. Poland A recent audit of seven state-owned commercial banks indicated that substandard loans accounted for 25 to 60 percent of assets. Yugoslavia According to a statistical survey of the National Bank of Yugoslavia, problem loans in 1988 amounted to 35 to 40 percent of banks' portfolios. Source: World Bank 1989 (updated); Financial Times, various issues; The Economist, various issues; and other reports. Banking Fragility in the 1980s: An Overview 23 Annex 1.2 Euromoney top 500 banks: selected indicators, 1992 (percent) Number of Equity/asset Profit/asset Return Countrygroup banks ratio ratio on equity Africa South Africa 6 4.91 0.78 15.32 Asia 46 5.19 0.63 11.96 China 5 2.07 0.50 17.29 Hong Kong 1 4.47 0.45 6.89 India 2 2.84 0.39 12.64 Korea, Rep. of 15 6.41 0.39 6.15 Malaysia 2 6.12 0.47 6.82 Singapore 5 8.78 0.92 10.93 Taiwan (China) I1 4.53 0.91 20.56 Thailand 5 6.32 1.04 14.37 Latin America 24 20.15 0.84 2.39 Argentina 3 14.37 -0.47 -12.91 Brazil 8 11.75 0.67 6.83 Mexico 7 7.68 0.55 4.22 Puerto Rico 4 12.30 1.23 10.25 Uruguay 1 24.29 0.00 0.00 Venezuela 1 50.49 3.06 5.97 Middle East and North Africa 23 7.95 1.00 9.08 Algeria 3 6.76 0.38 5.42 Bahrain 1 7.88 0.22 1.42 Iran 1 0.67 0.00 0.58 Israel 3 4.65 0.07 1.03 Jordan 1 6.53 0.68 10.44 Kuwait 1 9.74 1.42 n.a. Qatar 1 12.13 1.99 n.a. Saudi Arabia 5 7.61 0.94 19.34 Turkey 4 11.77 3.31 25.98 United Arab Emirates 3 11.70 1.00 8.43 OECD 392 4.64 0.33 5.69 Australia 8 5.89 0.62 9.58 Austria 7 4.10 0.29 5.56 Belgium 8 3.14 0.08 5.60 Canada 10 5.45 0.64 8.47 Denmark 6 6.25 0.19 2.80 Finland 5 5.15 -1.07 -22.30 France 23 3.53 0.23 5.41 Germany 41 2.82 0.15 4.40 Greece 3 3.12 0.44 14.70 Ireland 2 4.87 0.44 4.28 Italy 42 4.79 0.40 8.54 Japan 81 3.44 0.19 5.51 Luxembourg 1 0.95 0.23 8.10 Netherlands 6 4.25 0.39 6.80 New Zealand 1 4.65 -0.36 -5.77 Portugal 3 9.48 1.05 9.91 Spain 19 5.83 0.89 12.06 Sweden 5 3.12 1.00 15.44 Swimzerand 11 5.94 0.45 5.94 United Kingdom 27 4.73 0.39 8.82 United States 83 6.00 0.39 5.72 Transition economies 9 6.82 0.87 12.39 Bosnia-Herzegovina 1 8.55 0.50 5.80 Bulgaria 1 4.41 0.83 18.74 Cuba 1 2.70 0.00 0.00 Czechoslovakia 1 4.70 1.38 29.28 Poland 2 10.04 4.05 39.12 Slovenia 1 9.11 -1.23 -13.31 Yugoslavia 2 8.24 0.58 7.07 Source: Euromoney 1992. 24 BANK RESTRUcrURING Note Huh, Chan, and Sun Bae Kim. 1994. "Financial Regulation and Banking Sector Performance: A Comparison of Bad Loan 1. Under Soviet-style funds management all current receipts of Problems in Japan and Korea." Economic Review 2: 18-29. state enterprises had to be placed in the current account with a Federal Reserve Bank of San Francisco. designated state-owned bank, on which little interest was earned. IMF (International Monetary Fund). 1990. Recent Economic These enterprises were allowed to service interest, pay for inven- Developments, Romania. Washington, D.C. tory purchases, and pay wages from this account. Any surplus - . 1993. International Capital Markets, Part 11. Systemic was transferred to the state in the form of taxes or dividends. Issues in International Finance. Washington, D.C. Funds required for investment were allocated from the budget - . Various years. International Financial Statistics. under the central plan or approved for bank borrowing under Washington, D.C. the credit plan, on which (after interest liberalization) market Long, Millard F. 1987. "Crisis in the Financial Sector." World interest was paid. This explains the unusually high enterprise Bank, Washington, D.C. deposits in these banking systems and the high spreads earned McKinnon, R 1991. "Taxation, Money, and Credit in Liberalizing by their banks. For example, 88 percent of Chinese enterprise Socialist Economies: Asian and European Experiences." deposits with banks were sight or current deposits (September Stanford Universiry, Department of Economics, Stanford, Calif. 1989) and accounted for 36 percent of total deposits in banks. Montes-Negret, Fernando. 1990. "An Overview of Colombia's Even at the height of hyperinflation, 56 percent of Yugoslavian Banking Crisis, 1982-1987." Background paper prepared deposits were demand or sight deposits. for World Development Report 1989. World Bank, Washington, D.C. References Morris, Felipe. 1990. Latin America's Banking System in the 1980s: A Cross-Country Comparison. World Bank Bank of Korea. 1992. Economic Statistics Yearbook. Seoul. Discussion Paper 81. Washington, D.C. BIS (Bank for International Settlement). 1990. "Large-Value National Bank of Yugoslavia. 1990. Quarterly Bulltin. Sarajevo. Funds Transfer Systems in rhe Group of Ten Countries." OECD (Organization for Economic Cooperation and Basle, Switzerland. Development). 1992. "Bank Restructuring in Central and Blejer, Mario, and Adrienne Cheasty. 1991. "Measurement of Eastern Europe: Issues and Strategies." OECD Financial Fiscal Deficits: Analytical and Methodological Issues." Market Trends (February). Paris. Journal of Economic Literature 29: 1644-78. Park, Yung Chul, and Won Aun Park. 1992. "Capital Market, Bryant, Ralph. 1984. "The Progressive Internationalization of Real Asset Speculation, and Macroeconomic Adjustment in Banking." Brookings Institution Discussion Paper 14. Korea." Organization for Economic Cooperation and Washington, D.C. Development, Paris. Caprio, Gerard, Izak Atiyas, and James Hanson. 1994. Financial Schuijer, Jan. 1992. Banks Under Stress. Paris: Organization for Reform: Theory and Experience. New York: Cambridge Economic Cooperation and Development. University Press. Schumpeter, Joseph. 1934. The Theory of Economic Development. Davis, E. P. 1989. "Instability in the Euromarkets and the Cambridge, Mass.: Harvard University Press. Economic Theory of Financial Crisis." Bank of England Sheng, Andrew. 1992. "Bad Debts in Transitional Socialist Discussion Paper 43. London. Economies." World Bank, Financial Policy and Systems de Juan, Aristobulo. 1987. "From Good Bankers to Bad Bankers: Division, Washington, D.C. Ineffective Supervision as Major Elements in Banking Sundararajan, Vasudevan, and Tomas Balifio, eds. 1991. Banking Crises." Economic Development Institute Working Paper. Crises: Cases and Issues. Washington, D.C.: International World Bank, Washington, D.C. Monetary Fund. Diaz-Alejandro, Carlos. 1985. "Goodbye Financial Repression, U.S. Bureau of the Census. 1991. Statistical Abstract of the United Hello Financial Crash." Journal of Development Economicr. States. Washington, D.C.: Government Printing Office. Enthoven, Adolf, J. V. Sokolov, and A. M. Petrachkov. 1992. Velasco, Andres. 1991. "The Chilean Financial System, Doing Business in the Russian Federal Republic and the Other 1975-85." In Vasudevan Sundararajan and Tomas Balifio, Republics of the CIS: Managerial and Financial Issues. eds., Banking Crises: Cases and Issues. Washington, D.C.: Montvale, N.J.: Institute of Management Accountants. International Monetary Fund. Fry, Maxwell J. 1991. "Can a Central Bank Go Bust?" World Bank. 1989. World Development Report 1989: Financial University of Birmingham International Finance Group Systems and Development. Washington, D.C. Discussion Paper 91-09. Manchester School of Economic . 1990. China: Between Plan and Market. A World Bank and Social Studies, Manchester, England. Country Study. Washington, D.C. Gaspari, Mitja. 1989. "Balance of Payments Adjustment and . 1991. World Development Report 1991: The Challenge of Financial Crisis in Yugoslavia." In C. Kessides, T. King. M. Development. Washington, D.C. Nuti, and C. Sokil, eds., Financial Reform in Socialist . 1992. World Development Report 1992: Development Economies. Economic Development Institute Seminar and the Environment. Washington, D.C. Series. Washington, D.C.: World Bank. Yusof, Zaimal Aznamn, and others. 1994. "Finanacial Reform in Hinds, Manuel. 1988. "Economic Effects of Financial Crises." Malaysia." In Gerald Caprio, Izak Atiyas, and James Policy Research Working Paper 104. World Bank, Hanson, eds., Financial Reform: Theory and Experience. Washington, D.C. New York: Cambridge University Press. CHAPTER 2 Bank Restructuring Techniques Andrew Sheng Bankers and policymakers often fail to respond effec- interests, resolving problem banks becomes nor a tively to evidence of an impending banking crisis. technical issue but a political one. Bankers, for example, tend to engage in "cosmetic" In other cases supervisors are relLctanit to con- behavior, evergreening bad credits, assetizing losses, front complex structural issucs, such as fundamental and hiding material risks and losses from the public changes in the law, that could involve issues of and from bank supervisors (de Juan 1987). Proper national history and political economiiy. An examiiple understanding of the scope and causes of bank losses is the political difficulties encountered in clinliliatinig necessarily precedes effective treatment andi accurate the geographical and functional segmentation of ril prognosis. Although individual bank crises are trig- U.S. banking system, where many belicved in Uilit- gered by specific events-such as bank runs, capital or community-based banking and werc deeVply ljs- flight, or the disappearance of prominent bankers- trustful of nationwide banks. such events only mark the point of public or official In addition, there are few incentives lor governi- recognition of problems that have been building up ment agencies to deal with problems that involve over a long period. political economy issues of bank ownerslip and control and of borrower and depositor propetty Recognizing Failure rights. For example, many central banks art mole concerned with issues of bank complian'e with Proper diagnosis does not necessarily imply correct credit allocation guidelines or monetaly targets prognosis. In many cases bank supervisors are unable than with issues of solvency. Many developing or unwilling to measure the true extent of bank loss- countries interested in developing rural bankilng es, especially where inadequate accounting standards systems have required state-owned banks to openl allow income to accrue on nonperforming loans or extensive rural branches even wheln thtese were allow such loans to be rolled over. Many countries operating at a loss. In such cases state-owned binzks also fail to recognize the size of the problem, whether become pools of political patronage and employ- because of a lack of political will or because of a per- ment generation. ceived inability to deal with such losses. Moreover, ownership and sL'pervision of banks There are further incentives to hide losses or delay and other deposit-taking agencies may be diffused action. Private banks are reluctant to reveal losses across different ministries or agencies with varying because they fear government intervention and a run skills and powers of supervision. Central banks cani- on banks. State-owned banks have no incentive to not always take action against banks owned by the operate profitably since they lend largely to state- ministry of finance. A weak bank superintenden:cy owned enterprises or in accordance with policy-based may not be able to enforce bank supervision regula- credit directives. Governments resist dealing with tions against state banks whose chief executives are bank problems where they rely on the banking sys- political appointees. Large rural banks may be olrt- tem as a source of taxation or finance for fiscal side the supervision of the ministry of finance or deficits. Where banks are controlled by political the central bank. Politicized credit unions have 2,) 26 BANK RESTRUCTURING been a source of problems in many economies, There is therefore a significant time lapse between notably Japan and Taiwan (China). loss recognition and loss measurement and allocation. Where bureaucrats believe that problems are The lag occurs because the parties involved- cyclical and not structural, some authorities engage in bankers, supervisors, and policymakers-often refuse "regulatory forbearance"-allowing more and more to acknowledge that losses have occurred. The diffi- ailing institutions to breach the law, hoping that time cult question of political economy is how to allocate and economic recovery will resolve the problems. In the losses across different sectors and over time. Who the United States such behavior worsened the prob- pays for bank losses? lem of moral hazard, since undercapitalized problem A recent example is the reluctance of authorities institutions have a competitive advantage over sound- in the post-centrally planned economies to recognize ly capitalized institutions and are more likely to the economic insolvency of state banks that had engage in high-risk activities (Brewer and Monschean inherited loans to loss-making state enterprises. 1992; chapter 4). These banks have not suffered bank runs because Although ignoring failure can go on for decades, they carry a state guarantee on deposits and are fully the cumulative effects of weak intermediation, exces- state owned. With massive changes in relative prices sive overhead, loan losses, mismanagement, and fraud in these economies, the banks may continue to often surface during recessions, leading initially to finance loss-making enterprises for years to come- illiquidity in the weaker institutions and ultimately to further distorting resource allocation. In many coun- default and failure. The government is forced to deal tries policymakers have been reluctant to recognize with the problems when bank runs threaten systemic these losses because of the large resources required to bank failure. recapitalize these banks (Sheng 1992). The major arguments against marking bank assets Defining insolvency to market value are that there may be no markets in such assets and that thin markets may yield inaccu- Because not all financial institutions have adopted rate valuations. Asset valuation is relatively straight- market value accounting, which requires that all forward in industrial economies, where established assets and liabilities be marked to market, two types markets, valuation expertise, and set accounting and of bank insolvency should be distinguished. A finan- asset valuation standards are in place. Asset valuation cial institution is in economic insolvency when the is much more difficult for less-developed econonmies, market value of its assets (capital and reserves, exclud- where none of these conditions prevail. ing the value of deposit guarantees) is lower than the Asset valuation-particularly the valuation of col- market value of its (nonequity) liabilities (Kane lateral-is crucial for loan provisioning and hence for 1985). A financial institution is in accounting insol- basic bank solvency. In most countries real estate vency when the accounting report of its net capital forms the bulk of bank collateral (box 2.1). The share and reserves, according to generally accepted of bank lending in real estate peaked at 37 percent in accounting principles, is negative. A financial institu- Malaysia and reached 42 percent of total loans in the tion failt when it is unable to meet its obligations and United States. At the 35 percent level, the exposure authorities intervene to restructure or liquidate the of banks to real estate would be 350 percent of bank failed institution. capital, assuming a loan-assets ratio of 80 percent and Many financial institutions are economically an 8 percent capital base. By the end of 1991, 15 per- insolvent but appear to be accounting solvent because cent of U.S. construction and development loans different laws and accounting standards allow such were noncurrent, three times higher than the average institutions to operate, often under a national deposit (U.S. FDIC 1991). guarantee scheme. For example, if market-value accounting had been applied to OECD banks during Measuring losses the height of the international debt crisis in 1985-86, a number of them would have been judged Loss diagnosis involves four steps: analyzing the causes both economically and accounting insolvent. of bank losses, applying uniform accounting standards Moreover, many accounting insolvent banks do not to measure these losses, assessing the condition of fail because the government intervenes through banks, and calculating the costs of restoring banks to bailouts or other assistance. capital adequacy or, alternatively, of liquidating them. Bank Restructuring Techniques 27 Box 2.1 Bank exposure to real estate Since real esrate (land and buildings) form the bulk of yields are low and property developers typically recover loan collateral, banks are particularly vulnerable to real cash flow through property sales. Cyclical overbuilding, estate price changes. In the United States about half of however, means that commercial property prices can fall all loans from banks with holdings of less than $1 bil- by as much as 50 percent from peak to trough, as has lion are collateralized by real estate; banks with assets of occurred in Japan (1991-92), Malaysia (1984-86), and more than $1 billion have 35 percent of total loans the United States (1989-92). secured by real estate. Declines in collateral value reduce A bank with 35 percent exposure to real estate is bank profitability in two ways: an income effect, since rapidly decapitalized by a fall of 50 percent in collateral the borrower may not be able to service debt, and a value over a three-year period. Assume that a bank lends wealth effect, since the bank will have to provide for the 80 percent against estimared market value. If collateral shortfall between collateral value and estimated market value falls below the original loan principal by 10 per- value. Commercial property is particularly vulnerable to cent in each consecutive year, a bank with an initial 8 price fluctuations. All property loans have effective percent capital-asset ratio would fall below the 2 percent maturities of seven to ten years because average rental level by the third year. An important step in the diagnostic process is to Depending on the financial reputation of the bor- improve bank accounting and auditing standards- rower, some banks provide such loans on an unse- especially loan classification and interest accrual stan- cured basis. In Colombia, when the Superintendency dards-so that the degree of bank distress can be of Banks limited the exposure of banks to their relat- measured uniformly in all areas. Although there is no ed group of companies, some banks channelled funds set of international standards for loan classification, to specially created subsidiaries in Panama that lent two sets of criteria determine loan performance: the the money back to the related companies in lag in servicing of loans and the profitability or per- Colombia (chapter 6). In other cases banks simply formance of borrowers. rescheduled or rolled over debt, providing fresh loans In most cases a loan is nonperforming if it has not to service interest. In some developing countries been serviced in 90 to 180 days. Income on nonper- banks have been known to exchange their bad debts forming loans should not be tallied in the profit and with one another and to treat these as new loans, thus loss account, and realistic provisioning should be delaying the requirement of reporting nonperforming made against the likelihood of losses in the recovery loans to the central bank. of debt. In many countries markets for valuing loan The World Bank has published a set of guidelines collateral such as plants and equipment are absent. for bank supervision that provide minimum stan- Prior to the establishment of special banking tri- dards of best practice for developing countries in loan bunals, courts in Pakistan took as long as ten years to classification and income accrual, accounting for debt deliver a judgment on debt recovery. Banks are also rescheduling, credit concentration, and accounting reluctant to make provisions where these are not tax- disclosure along generally accepted accounting prin- deductible against profits. Accordingly, the extent of ciples (World Bank 1993). nonperforming loans in many developing countries is grossly understated because of the lack of uniform, Causes of Losses enforced accounting standards, and hence bank prof- its are overstated. The causes of bank losses are as important as their Even where reasonable accounting standards value. A 1988 study by the U.S. Comptroller of the apply, banks generally have found ways to disguise Currency found that U.S. bank failures could be the reporting of nonperforming loans. De Juan attributed mainly to poor asset qua'ity (found in 98 (1991) found that the most problematic loans in clas- percent of cases) and poor manageinent (90 percent). sification schemes are those that are reported as "cur- A weak economic environment was a factor in 35 rent." Many banks in developing countries operate percent of U.S. failures, and fraud was an issue in overdraft loans, which are recallable on demand. only II percent. In developing countries the eco- These cash loans, or even short-term loans, are in nomic environment may play a larger role-but effect long-term loans, since they are almost always probably not more than, say, 50 percent. In many rolled over, often with the interest capitalized. financial crises-including severe decade-long cases 28 BANK RESTRUCTURING (as in Argentina)-private banks with good manage- responsible for a significant volume of loan losses. ment survived, indicating that bank losses are not Banks made loans according to policy directives and inevitable and can be avoided through good risk were lax in their credit evaluation because they management and adequate capital. believed that the state would underwrite the loans. In Banks incur losses as a result of bad debt, opera- Kenya several small banks and finance companies tional losses, speculation, inefficiencies, excessive failed because of excessively liberal licensing policies taxes or regulation, and fraud. Thus bank losses in the early 1980s-policies that were intended to reflect significantly the impact of changes in relative improve indigenous ownership in the finance sector. prices (asset prices, tax rates, interest and exchange Weak management skills and fraud, complicated by rates) on a bank's operations and balance sheet. the ownership of banks by political factions, led to When such flow and stock losses exceed the bank's the failure of several smaller institutions. These insti- total capital and reserves, bank insolvency occurs. tutions were subsequently consolidated under the While losses can often be controlled or minimized supervision of the central bank. A deposit insurance with quality management, some changes in relative fund protected depositors from losses. In the after- prices, such as taxes and exchange rate or interest rate math of political change in Kenya in the early 1990s, controls, are outside a bank's control. Studies of the several loss-making banks were closed. causes of bank insolvencies therefore must consider the behavior of bank management, borrowers, depos- Commodity shocks itors, and policymakers in response to external shocks, policy changes, and sectoral imbalances. Market-based banking systems with reasonable Assessing a bank's standing requires setting objec- supervision standards in economies that depend on tive standards of measurement, obtaining indepen- commodities-such as Malaysia (chapter 7), dent verification of bank operations (through audi- Norway, and Texas in the United States (chapter tors or supervisors), and providing full disclosure of 4)-also suffered severely in the wake of the col- results to all parties involved in bank restructuring: lapse of oil and gas prices in 1984-85. In the decade depositors, shareholders, bank employees, manage- before the collapse, banks in these areas enjoyed ment, and supervisory authorities. The full extent of high profits as business boomed. The boom created stock losses (historic losses at a point of time) and a bubble in real estate and other asset markets as potential flow losses (estimated future performance if speculation in property and shares surged, financed restructuring fails), as well as the implications of loss by banks in pursuit of higher profits. The asset bub- distribution, should be transparent to all parties, ble burst as commodity prices turned downward since perceptions of inequitable distribution of gains and inflation fell. The real value of debt increased or losses, based on imperfect information, can sabo- relative to collateral assets, thus creating a spate of tage any restructuring effort. enterprise defaults. Governments in all three areas During the 1980s severe bank losses appeared in intervened through the central bank or a state both industrial (Norway, Spain, the United States) deposit insurance fund. and developing countries for a variety of reasons. Inherited portfolios Credit losses A number of banks created in the reform of banking Bank losses cannot be attributed solely to external systems in formerly socialist economies-such as the shocks. India, Kenya, and Pakistan all enjoyed several Czech and Slovak Republics, Hungary, and Poland- decades of uninterrupted growth, but each country's inherited large portfolios of nonperforming loans to banking system has suffered problem loans for a vari- state-owned enterprises. Since state-owned banks had ety of microeconomic and macroeconomic reasons. no credit risks under a centrally planned economy, In India and Pakistan the credit allocation policies of many of these enterprises were loss-making ventures. state-owned banks played an important role. Banks The quality of most of these loans remains highly were expected to maintain a large number of loss- suspect, but it has not been possible to estimate the making branches, particularly in rural areas, and to losses accurately in the absence of stable prices. In lend at below-market rates for economic and social 1990-91 many of these enterprises suffered trade welfare purposes. Such policy-based lending has been shocks from the collapse of the trade arrangement Bank Reoructuring Techniques 29 through the Council of Mutual Economic was caused primarily by the maturity mismatch in Assistance, creating large loan losses for the banks. the balance sheet of savings and loans, which invested in twenty-year or more fixed-rate mortgages but Connected lending relied on short-term deposits at flexible rates. After short-term rates rose in 1980-81 the thrift industry Banks in Argentina, Chile, Colombia, and Spain suf- had an estimated net insolvency of about $100 bil- fered from a system in which banks were owned by lion in 1982 (U.S. Treasury 1991). large economic groups that also owned enterprises, with substantial bank lending to finance activities Foreign exchange mismatch within the same group. The largest failure of such a group was the Rumasa group in Spain, which was The 1974 failure of Herstadtt Bank in Germany nationalized in 1983 (chapter 5). Seventeen of the highlighted the dangers of mismatches in foreign cur- twenty Rumasa-owned banks had made loans of rency positions. Inadequate management attention to more than 73 percent of their total credit to related such dangers could lead to large speculative losses, companies, which numbered more than 300 sub- such as those reported by Bank Duta in Indonesia in sidiaries. Connected lending also brought about the 1990. Until 1989 in Poland and Yugoslavia, for failure of a small German bank, Schroder example, banks took substantial amounts of foreign Munchmeyer, which was caught in excessive lending currency deposits from residents but surrendered to its parent construction company in 1983. these funds to their national banks. Since foreign exchange reserves were considerably smaller than Excessive regulatory taxes external debt during most of the 1980s, these bank- ing systems suffered large losses when the domestic Efforts by central banks to control inflation through currency was devalued. noninterest-bearing statutory reserves or low-interest liquidity reserves reduce bank profitability and may Fraud eventually be passed on to depositors or borrowers through high spreads. In Argentina reserve require- The more infamous cases of bank failures have been ments averaged 15.8 percent in 1987, and forced associated with fraud. Early cases of fraud such as investments reached 50 percent of total deposits. In Banco Ambrosiano (Italy, 1982), Johnson Matthey Yugoslavia regulatory costs in the form of statutory (United Kingdom, 1984), and BMF Hongkong and liquidity reserves added nearly 20 percent to (Malaysia, 1982) seem small compared with the spec- bank spreads in 1989. It has been estimated that tacular closure of the U.K. Bank of Credit and financial sector taxation in selected Sub-Saharan Commerce International in 1991, where fraud and African countries has averaged up to 10 percent of mismanagement may have caused losses of $5 billion GNP (Chatnely and Honohan 1990). to $10 billion out of total assets of $20 billion. In Guinea five state-owned banks-accounting for the Excessive overhead bulk of the banking system's assets-were closed in 1985 when 78 percent of their assets were found to Developing country banks incur high administrative be fictitious (Tenconi 1989). Even in traditionally costs through overstaffing, excessive branching, and orderly markets such as Japan, cases of forged deposit wasteful expenditures. Prior to restructuring, banks certificates have caused large losses to the smaller in Ghana incurred noninterest operating costs equiv- trust banks. alent to 6 percent of total assets, compared with an average of I to 2 percent in OECD banks. These Flawed liberalization policies costs accounted for more than 75 percent of total interest income. Many of the banking problems of the 1980s can be traced to poorly designed financial liberalization poli- Interest rate mismatch cies (Zahler 1993). Liberal entry rules and the expan- sion of new banks and deposit-taking institutions in All banks suffer structurally from maturity mismatch- the absence of adequate capital and managerial skills es in their portfolios. The U.S. savings and loan crisis and sufficient supervision were responsible for 30 BANK RESTRUCTURING failures in Argentina, Chile, Kenya, Spain, and led to oversupply and resulted in heavy real estate Uruguay. Caprio, Atiyas, and Hanson (1993) found loan losses in U.S. banks. Delays in obtaining court that the lifting of credit ceilings and deregulation of orders to force debt repayment encouraged firms to interest rates allowed banks to take excessive risks in default in times of tight liquidity. In some developing areas where they had no prior experience, such as real countries the total lack of enforcement of banking estate. Sundararajan and Balifio (1991) conclude that laws and regulations emboldened bank management the connection between financial reform and crises to ignore laws and undertake speculative risks. derives from an unstable macroeconomic environ- While there is consensus that adequate supervi- ment, unsound liability structures of nonfinancial sion and enforcement are required to maintain stabil- firms, and weaknesses in the institutional structure ity in a financial system, even the most sophisticated for banking. bank supervisors in advanced OECD countries have not been able to prevent bank failures completely. Understanding the bigpicture Bank supervision is thus a necessary but insufficient condition for bank stability. Supervisory authorities Thus banking problems have no single cause. Neither must pay greater attention to issues of sector and the monetary causes of banking crises according to national imbalances that may destabilize the banking the approach by Schwartz (1985) nor the business sector. Similarly, there is a need for overall national cycle models of Minsky (1977) and Kindleberger risk management (Sheng and Cho 1993). (1978) are fully explanatory. To blame solely manage- ment failure, policy mistakes, or a business cycle ten- Assessing Problem Banks dency to develop Ponzi-type bubbles would not suf- fice. A proper diagnosis calls for a thorough under- Once uniform accounting standards and the causes standing of the policy and economic environment, of losses are agreed on, the actual assessment of bank the institutional framework, banking practices, the solvency takes place. Bank examiners in the United quality of bank supervision (if any), and the structure States use a uniform CAMEL rating system, which of incentives (box 2.2; Glaessner and Mas 1991). assesses capital adequacy, asset quality, management, Policymakers are becoming increasingly aware earnings, and liquidity (see chapter 3). In the that perverse incentives can lead to bank failure if left absence of bank examiners, some bank supervisors unchecked. For example, recent evidence suggests use external auditors for an independent review of that it was moral hazard-not simply bad luck or the financial condition of banks. In World Bank delayed closure-that led to the U.S. savings and operations in the financial sector, special portfolio loan crisis and increased its cost (Brewer and audits have been commissioned using international Monschean 1992). Excessive tax incentives to real auditors under carefully designed terms of reference estate, for example, generated a property bubble that (box 2.3). Such audits should provide the bank Box 2.2 Assessing the risk of bank failure A review of the policy and institutional environment in Institutional environment which banks operate is a good start for the diagnostic * Are the legal framework and judicial processes con- process. ducive to enforcement of debt recovery? Policy environment * Do domestic accounting and auditing standards meet * Is there significant financial repression? internationally accepted accounting standards? * Does the state own a large stake in the financial sector? * Is information on credit and borrower performance * Is there liberal entry into the financial sector? available and transparent? * Is there a nonbank financial sector that is growing * Is there good bank supervisory capacity? rapidly without supervision? * Do supervisory authorities fully understand the prob- * Are credit allocation and forced lending policies hurt- lems facing the banking system? ing banks' autonomy in credit decisions? * Is sufficiently trained bank management staff in * Are banks being taxed considerably higher than non- place, for example, in foreign branches? banks? * Do bankers have a good understandi 'g of the costs of * Are banks substantially owned by large enterprise intermediation and the sources of their profits and groups? losses? Bank Restructuing Techniqus 31 Box 2.3 Special audits Terms of reference for special audirs of banks should rate, exchange rate, and maturity mismatches. include: * Compliance with regulations, especially for capital Accounting diagnosis adequacy. * Assessment of the bank's portfolio under agreed uni- * Review of internal controls and potential exposure to form accounting standards and loan classification fraud. methodologies. Institutional diagnosis * Assessment of the bank's credit process, including the * Review of bank's strategic plan to operate safely and quality of information used in loan classification and efficiently in a competitive environment. estimation of provisioning. * An audit of the adequacy of management systems * Evaluation of adequacy of provisioning. and staffing and training policies. * Analysis of profit and loss accounts, including cost of * Review of organizational responsiveness to opera- intermediation. tional and strategic needs. * Analysis of reliability of reported earnings and operat- * Assessment of credit policies and procedures. ing expenditures. * Assessment of treasury operations (asset and liability * Evaluation of credit concentration to largest borrow- management capacity, including management of liq- ers and the extent of connected lending. uidity, interest rate, and exchange rate risks). * Evaluation of contingent liabilities such as lawsuits, * Review of the management information system (abil- unfunded pension liabilities, guarantees, and expo- ity to obtain reliable and timely accounting and sure to futures and options. strategic information for sound decisionmaking). * Restatement of the accounts (balance sheet as well as * Review of the efficiency and reliability of operational profit and loss accounts), including adequate provi- procedures, including computer systems. sioning for all known losses or adjustments on assets * Review of the internal audit function. and liabilities, as well as off-balance items. * Role of board of directors in monitoring and super- * Solvency assessment, including exposure to interest vising management performance. supervisor with sufficient information to assess the Loss Recognition and Allocation basic causes of bank losses, a realistic (and indepen- dent) assessment of the size of losses, and suggested An important watershed in the bank restructuring remedies. process is loss recognition, the point at which policy- Recent experience with portfolio audits in Central makers become convinced of the need to act despite and Eastern Europe, where there has been great the enormous challenges involved. This usually volatility in prices and policies, suggests that mea- occurs when it becomes apparent that failing to act surement of the extent of losses is still subject to sig- would not only jeopardize confidence in the banking nificant variation. This is because the accounts of system, it would also destabilize the macroeconomic enterprises (the banks' main clients) are not yet main- environment. This stage cannot be reached until a tained according to generally accepted accounting sound assessment of past losses in the financial sys- principles and auditing standards that conform to tem has been completed, as well as a projection of European Union requirements. Even where these future losses if the problems are not addressed. standards have been attempted, international auditors As discussed in chapter 1, bank losses should have been unable to impose inflation accounting. As never be underestimated. Underestimating the size of a result accounting information is quickly outdated losses creates credibility problems later with the pub- under conditions of high inflation. A comprehensive lic and policymakers, and insufficient resources- audit, followed by regular and quick portfolio financial or human-may be allocated to deal with reviews, is required to establish the trend of losses in the problem. Time and again, delays in recognizing banks operating in such environments. failure and taking action have escalated the cost of Since using external auditors can be costly, some failure resolution. For example, bank losses in countries have built up their own capacity for off-site Norway first surfaced in 1987, when eight commer- surveillance and on-site examination. Such teams cial banks and seven savings banks began reporting provide another assessment of banks' solvency, which losses. The banking system as a whole reported nega- may be grossly underestimated by bank management tive profits in 1988 (Solheim 1990). The banks ini- or even external auditors. tially relied on bank-funded commercial banks' and 32 BANK RESTRUCTURING savings banks' guarantee funds, but by 1991 the is likely-"when most everyone (who counts) is funds were insufficient to deal with the widespread bankrupt, nobody is!" (Diaz-Alejandro 1985). problems. The Norwegian government ultimately Indiscriminate bailout of bad borrowers also may had to create a Government Bank Insurance Fund to induce good borrowers to delay debt repayment, deal with the problem (box 2.4). thus imposing higher bailout costs. And borrowers Once bank losses have been recognized, they must may have no incentive to repay if repayments can be be allocated. But who should bear the losses? Not nec- delayed due to inefficient cour t enforcement of essarily the holders of bank obligations. Conventional bankruptcy or debt recovery. thinking assumes that banking losses should be borne, in broad descending order, by borrowers, sharehold- Shareholders ers, fellow bankers, employees (in situations of liqui- dation), government, and depositors. The failure of If losses are met by shareholder deposits of new capi- banks is like market failure. By not allowing banks to tal or bank liquidation, shareholders would lose their collapse, and hence effectively passing the losses on to wealth. But shareholders with political clout may not all holders of the failed banks' liabilities, there may be lose their shares if they induce a government bailout. no equitable or fair method of allocating such losses. In Colombia the central bank's injection of addition- Loss allocation decisions generally have been made on al equity into the rescue of a bank preserved the the basis of the existing legal framework and the cur- shareholders' shares, although the shareholders lost rent conditions of political economy. Because most controlling interest. In Sri Lanka shareholders were governments do not feel that depositors should lose, effectively protected from loss when the central bank the least painful political option is for the government injected liquidity into loss-making finance companies to bear the losses. But is this solution fiscally sustain- without requiring a reduction in capital. Before laws able? Moreover, will a government bailout create were changed in Malaysia and Thailand, initial bank problems of moral hazard? Each of the potential bear- restructuring efforts imposed only partial losses on ers of bank losses plays a role in determining the opti- shareholders. In Thailand finance company share- mal form of bank restructuring. holders surrendered only 25 percent of theii shares to the Ministry of Finance in return for lifeboat sup- Borrowers port, while in Malaysia central bank injections of capital into three commercial banks absorbed equiva- Defaulting borrowers pay in terms of the liquidation lent losses, protecting existing shareholders. Both of their collateral. If banks have difficulty collecting countries subsequently required full capital reduction from bad creditors, they may pass the loss on to of losses before government assistance was made other borrowers by raising lending rates. Losses allo- available. cated too heavily to borrowers through high lending Because of the principle that share capital should rates and rapid debt collection may cause even good be at risk, it is now standard practice that bank losses borrowers to default, leading to a credit crunch that should be written off against existing share capital may retard growth and recovery. Even good borrow- and reserves before governments intervene. Where ers have incentives to default if a government bailout reduction of capital is governed by corporate law, Box 2.4 Norway: An institutional solution Norwegian commercial and savings banks' cumulative directly acquire shares in banks. Between 1991 and 1992 losses on loans and guarantees during 1987-90 amount- the fund provided support loans and preference capital ed to NOK 35 billion. This followed six years of operat- amounting to more than NOK 16.7 billion to six finan- ing losses, attributable to falling property values, exten- cial institutions, including three of the largest banks in sive business failures, and economic recession. These the country. The fund generally injected capital into the losses eroded the banks' two lines of defense, their own banks to meet their statutory capital requirements, sub- capital and the bank industry-financed guarantee funds. ject to agreement by the banks to improve their operat- To strengthen bank safety the government created the ing efficiency and return to sound banking. Government Bank Insurance Fund (GBIF) with a capi- tal of NOK 5 billion to provide support loans or to Source:GBIF 1992. Bank Restructuring Techniques 33 however, that procedure may be extremely cumber- the case of state-owned banks, employee action has some and impractical-a bank supervisor's applica- delayed the process of privatization. The Argentine tion to a court for reduction of capital would trigger banking system has a labor force of 145,000, three an irreparable bank run. Accordingly, banking laws times that of Malaysia with only twice that country's such as Spain's allow deposit guarantee funds to write GDP. Restructuring state-owned Argentine banks off losses against existing capital before any invest- became difficult because bank employees became a ment in capital by the fund. powerful lobby against the closure of banks, inflicting further losses on the budget. Other bankers Depositors Industrywide cooperation in rescuing a single fail- ing institution is now common practice. This may A common technique to keep weaker institutions take the form of a larger, stronger bank taking over solvent is to lower deposit rates while maintaining a weaker institution, either vo!untarily or with high lending rates. Where no explicit deposit insur- some incentives provided by the central bank (such ance scheme exists, depositors may be willing to bear as a soft loan). Alternatively, the costs of liquidating some losses, such as lower deposit rates, so long as a small bank can be distributed among the other, the return of their principal is assured. Some coun- healthy banks. The issue of persuading other tries have used the inflation tax to pass on losses to bankers to absorb losses of failed banks is an open depositors, particularly through negative real interest one. In Europe and the United States policymakers rates on deposits. Deposit insurance schemes protect have moved toward greater concentration in bank- only the nominal value of deposits, not their real ing by providing incentives to profitable and well- value. managed banks to absorb smaller, failed banks. In Some depositors in low-inflation economies have the United States concentration of problem banks borne losses where there has been no explicit deposit was encouraged through "phoenixes," "bridge insurance scheme. When Thai finance companies banks," or management consignments. This failed in the early 1980s, authorities agreed to repay process may work where banks still have a "fran- depositors only the deposit principal, without inter- chise" value due to limited licensing, protection est, over ten years, in installments of 10 percent a from competition, or valuable branches or net- year. With average inflation of 5 percent a year, works, but liberalized entry policies in a number of depositors bore real losses of about 50 percent. When countries are making the process less appealing. In deposit-taking cooperatives failed in Malaysia, the the past large Japanese banks helped absorb weaker, authorities agreed to repay the principal, 50 percent smaller banks. There is now resistance to such in cash over two to three years at a positive (but low) moral suasion. interest rate and 50 percent in equity or convertible The technique of strong banks absorbing weak bond in a licensed finance company that absorbed banks works in the exceptional failure or two, and the assets of the failed cooperatives and whose shares where the rest of the banking system is reasonably were to be sold publicly (chapter 7). When the com- strong and profitable. Where losses are larger and pany was sold to a successful, publicly listed company more widespread, the stronger banks often refuse to in exchange for shares in 1991, some depositors help because of the danger of contagion. There is ended up gaining from the exercise. potential for disaster when weak, state-owned banks take over failing private banks, as demonstrated by External creditors the Consolidated Bank of Kenya, which later ran into substantial difficulties. Some losses can be passed on to the external sector if foreign creditors are willing to absorb some of Employees the losses on their foreign currency loans to the country in which the failed banks reside. An inno- Bank unions are generally among the strongest vative technique for recapitalizing rural. banks unions in a country and have a significant say in the using the discount value of the Philippine govern- resolution of failed banks. Making failed institutions ment's debt in the secondary market is described in more efficient often involves large staff cutbacks. In box 2.5. 34 BANK RESTRUCTURING Box 2.5 Debt-equity swaps in bank restructuring Some countries with external debt at discounts to their dollar promissory notes of the Philippine government at par value can use the discounts to recapitalize banks. a discount from foreign noteholders. These were sold to Commercial banks in Ecuador, for example, bought the the central bank at 100 percent par value, in pesos plus nation's foreign debt from the secondary market at a interest, subject to the Land Bank's participation in the discount in U.S. dollars and sold these bonds to the Rural Bank Capital Enhancement Program. Under the central bank, which bought them in domestic currency program a rural bank places new capital funds with the at full face value. The commercial banks used the prof- Land Bank, and the Land Bank issues the rural bank a its between the discount and the face value to recapital- seven-year promissory note for double the funds ize themselves. received as capital injection. The rural bank then uses A variation of this approach was used in the the Land Bank promissory note to liquidate debt to the Philippines to recapitalize the rural banks, which owed central bank. The system essentially uses the discount substantial sums to the central bank. The Land Bank of "profit" to recapitalize rural banks and settle their debts the Philippines bought restructured seventeen-year U.S. to the central bank. Government On the other hand, central banks that have been weakened by excessive lending to the public sector, Governments intervene in bank failures directly and with large interest-bearing liabilities and net foreign indirectly through central bank assistance, guarantees exchange liabilities, do not have the capacity to of bank deposits, and tax writeoffs for bad loans. absorb bank failure losses. These central banks run What share of bank failure losses should be borne by the risk of monetizing the losses through the growth the public sector? Should the losses be absorbed by of their monetary liabilities, generating an inflation- the treasury or by the central bank? In theory, the ary pressure that distributes the losses to the holders central bank is a part of the public sector, and there of currency and central bank liabilities. Large quasi- should be no difference between treasury and central fiscal deficits incurred by monetary authorities were bank financing of the losses. In practice, the source of major causes of hyperinflation in Yugoslavia and financing can be critical in bank restructuring Argentina (chapters 9 and 11). schemes, with the choice depending on the solvency Experience suggests that, where central banks do of the central bank and the size of its foreign not have the capital and reserves to absorb bank fail- exchange reserves. ure losses, the treasury should absorb the losses. The A traditional central bank-one with only nonin- treasury can finance bank failure losses by raising terest-bearing monetary liabilities, high capital, and taxes, cutting spending, using the inflation tax, sell- reserves represented mainly by foreign exchange ing its real or financial assets, or borrowing (long- assets-should have few problems absorbing bank term) domestically or abroad. In all cases bank failure failure losses. The stock of public sector savings (in losses are passed on either to taxpayers or to the hold- the form of the net worth of the central bank) is ers of public debt, including currencyholders. reduced by the losses absorbed. Since the central bank Despite the range of possibilities in loss alloca- can always earn seigniorage on its currency liabilities, tion, most governments choose the most politically it can absorb large bank losses. The central bank may, palatable course, that of bearing the losses in the pub- however, lose foreign exchange reserves-in the form lic sector through the central bank, a deposit insur- of capital flight-if it issues reserve money to rescue ance fund, or directly in the budget. One of the least banks. The losses from the failure of some small banks successful methods, with a total lack of transparency, in Hong Kong in the early 1980s were absorbed is to use the deposits of state enterprises to bail out against the reserves of the Exchange Fund. weak banks or finance companies, as happened in For solvent central banks or currency boards with Kenya during the 1980s. This approach hid the scale large reserves, bank losses can be absorbed with few of losses in the banks and financially weakened the monetary effects. Indeed, a central bank with large state enterprises. foreign exchange reserves and no external debt gains As long as the state provides implicit or explicit from a devaluation when large revaluation surpluses deposit insurance for the banking system, bank loss- are created. These surpluses can be used to absorb the es greater than available bank capital are equivalent losses from banking failures. to an internal debt of the government that is being Bank Restructuring Techniques 35 financed by short-term bank deposits. Accordingly, Rewriting the National Balance Sheet and Incentives the capacity of the state to deal with bank losses Structure depends on its debt capacity, the domestic growth rate, and the real interest rate paid on state debt. As mentioned earlier, any effort by the state to Governments in strong fiscal positions, with low address systemic bank losses must include a rewriting debt or strong borrowing capacity and favorable of the national balance sheet and incentives structure. access to foreign or domestic capital markets, may Authorities have to decide where losses should be be able to finance bank losses by swapping the bad allocated and whether these losses should be allocated debts in the banks with government bonds. As long to a one-time stock adjustment or to a gradual adjust- as they are able to borrow without significantly rais- ment over time through flow changes. Both tech- ing domestic real interest rates, crowding out the niques affect the net solvency of different sectors, private sector, or disturbing monetary stability, the with significant implications for the incentives struc- state will be able to transfer most of the burden of ture in the economy. bank losses to the future. The losses are written off This effect is best illustrated by reviewing sector over time, with the burden borne largely by taxpay- and national financial balances (see chapter 1). For ers. The dynamics of government debt and bank example, if total bank credit amounts to 20 percent losses are examined in box 2.6. of GDP and losses amount to 15 percent of total On the other hand, governments that are credit, then recapitalization of the banks would already saddled with high fiscal deficits and that require resources equivalent to 3 percent of GDP. If are paying high real interest rates may not have the the government decides to absorb bank losses capacity to deal with bank losses except through through a bond-bad debt swap (sometimes called a some form of inflation tax. These losses are spread carve-out), then the government would have to issue through the economy on financial assetholders bonds equivalent to 3 percent of GDP to shift the and often create distortionary resource allocation losses from the banking sector to the government sec- effects-which worsen the macroeconomic tor. If domestic interest rates are 10 percent a year, environment. the additional cost to the budget is 0.3 percent of Box 2.6 The fiscal impact of bank losses Since bank losses are quasi-fiscal deficits, the cumulative tively low inflation, high growth, and low fiscal deficits losses in the banking system are equivalent to an internal or small surpluses, d may be declining over tine. This debt of the state. The long-term consequences of running category includes Malaysia, Spain, and Chile after the such internal debt can be projected using the following macroeconomic stabilization program. equation: the change in the government debt to GNP On the other hand, countries with large primary ratio (d) is equal to the primary (or noninterest) deficit deficits and excessive real interest rates allow their debt of the public sector, less what is financed by seigniorage, ratios to become unsustainably large, so much so that plus the current debt ratio (d) times the average real the debt can only be reduced through higher and interest rate on the debt minus the growth rate of GNP: higher inflation. This category includes Argentina, Chile in the early 1980s, and Yugoslavia. The growth Change in d= (primary deficir/GNP) - of the domestic debt also rose sharply because of exces- (seigniorage/GNP) + d(real interest rate - growth rate) sive real interest rates, ranging as high as 40 percent a year. The quasi-fiscal deficit due to bank losses tends to How dwill perform depends on whether bank losses increase the primary deficit. To the extent that the cen- continues to flow or converge toward zero, provided the tral bank is already financing bank losses through liq- bank restructuring exercise has stemmed all future loss- uidity creation, the revenue that can be obtained from es. Carve-outs have not worked well in many Eastern seigniorage is correspondingly reduced. The change in European economies in transition because the problem the overall debt ratio d will increase with the primary stems from the continuing flow of enterprise losses. As deficit or if the real interest rate exceeds the domestic long as the real sector is still suffering from economic growth rate. On the other hand, d will decline with losses, lending to such enterprises will ultimately show inflation or high GNP growth. up as bank or quasi-fiscal losses. The equation above helps explain the varying success of bank restructuring programs. In countries with rela- Source: Fischer and Easterly 1990. 36 BANK RESTRUcrURING GDP Assuming that the government has a small fis- of and willing to absorb the risk of bank failure. In cal deficit of, say, not more than 2 percent of GDP, many countries deposit protection is afforded such additional costs may be fiscally sustainable with through a small deposit insurance scheme that helps no major monetary implications. There is no net small depositors but presumes that large depositors impact on the domestic economy's solvency since the can take care of themselves. transaction has no foreign exchange implications. Liquidation was the approach used in the case of If, however, the public sector has a large fiscal the United Kingdom's Bank of Credit and Commerce deficit of, say, 8 percent of GDP, and inflation is run- International (BCCI) in 1991, when extensive fraud ning at 30 percent a year, then the total fiscal deficit was uncovered. The bank was liquidated after negotia- would be worsened by the additional debt-servicing tions for additional capital injection to recapitalize the costs of 0.9 percent of GDP (30 percent times 3 per- bank failed. It was left to the liquidator of the bank to cent new bonds for the carve-out). At that point the recover and distribute any remaining assets according government may be faced with the combined prob- to the law. If the process is applied according to the lems of a declining tax base (due tc the losses of state- law, the depositors of the failed bank will suffer losses owned enterprises and consequent losses of banks, plus beyond any deposit coverage they had under the rising unemployment) as well as capital flight (by British deposit guarantee scheme. wealthholders concerned with higher taxation). Another option, particularly successful where the Thus different solutions involve some form of number of failed banks is small and there is 'fran- rewriting of the national and sectoral balance sheets chise" value, is to sell or merge the ailing banks, and, by implication, the incentives structure. Carve- without government assistance, to other banks, outs may recapitalize the banking system by shifting including foreign banks or new entrants to the bank- the burden to the state, but they also may create ing sector. This method was adopted in Spain and moral hazard incentives that worsen future banking Thailand, where the capital was reduced to the losses. Moreover, without parallel reforms in the real extent of the losses and the bank was recapitalized by sector, banks may continue to incur losses so long as offering the shares to new shareholders (Sundaravej they continue lending to the enterprise sector. Thus and Trairatvorakul 1989). In the Spanish case some bank restructuring often must be accompanied by of the larger banks that gained market share by changes in the real sector and in the incentives struc- absorbing smaller problem banks later got into trou- ture. It is futile to engage in paper transactions that ble (chapter 5). appear to address bank solvency but that do not If private buyers cannot be found for a failed address the fundamental causes of the losses. bank, experience indicates that liquidations are gener- Sustainable bank restructuring requires that the ally cheaper than keeping insolvent banks open. But national balance sheet and incentives structure be in some African banking systems with dominant rewritten so that the economy moves toward an state-owned banks, such as in Ethiopia, Tanzania, open, competitive, and stable financial structure. and Uganda, loss-making banks have been kept open Different countries have rewritten the national bal- because they are the only payments mechanism and ance sheet and incentives structure in different ways. major savings mobilization systems in the country. Bank restructuring works where concomitant mea- sures have been taken to reform the real sector, Box 2.7 Restructuring mechanisms accompanied by major changes in the fiscal and poli- Market-basedsolutions cy environment. * Shareholder capital injection * Sale or merger Liquidate or restructure? a Privatization * Liquidation without deposit compensation A first key policy decision is whether a market-based Government intervention solution or government intervention should be * Nationalization adopted (box 2.7). Market-based solutions are gener- * Liquidation with deposit insurance ally the most efficient, with least cost to taxpayers. * Asset recovery trust For example, a natural market solution to a failed * Bank "hospitals" bank is liquidation without any protection for the Supply-side solutions * Forced conversion into bonds d epositors. This presumes that depositors were aware . Bank Restructuring Techniques 37 Bank supervisors in a number of countries hesi- credibility. In almost all the cases examined, the gov- tate to liquidate banks because of either the "too big ernment (including the central bank) absorbed the to fail" principle or because of the possible contagion bulk of bank losses in order to protect the depositors. implications. But the real fear is perhaps "too big to One of the two most drastic forms of state inter- liquidate," where the costs of liquidation-in terms vention is nationalization. Nationalization of banks of economic disruption, political backlash, and fail- was undertaken in the 1980s by Costa Rica, El ure of the payments system-would be too large. Salvador, and Mexico. In Mexico the banks were Such hesitation creates perverse incentives-liberal nationalized in 1982 for populist reasons rather than entry into banking without exit results in moral haz- for reasons of national exigency. Nationalization ard behavior. The experience of the 1980s suggests internalizes the losses of the banking sector into pub- that as the costs of bank rescues rose, more bank lic sector debt, thus removing any doubt about the supervisors were willing to take the bold step of liqui- safety of private sector deposits in the banking sys- dation in order to break the moral hazard trap. tem. Even though Mexico shared the same debt crisis Market-based solutions cannot succeed, however, as other heavily indebted countries, it avoided a crisis where the problems stem from market failure itself. of bank runs (although not capital flight) because of As indicated earlier, such failures could have many the nationalization. The cost of avoidance of bank causes, including restrictive practices, bad macroeco- runs was inefficient credit allocation by state-owned nomic policies, and inadequate supervision. Once banks. It was mainly to promote the more efficient liquidation has been ruled out, government interven- allocation of resources that Mexican banks were pri- tion becomes the next logical step. vatized in 1990-91 (Barnes 1992). Another drastic solution is hyperinflation fol- Government intervention in the financial sector lowed by currency reform. Conceptually, hyperinfla- tion erodes the real value of the bank debt of the Although market-based solutions were tried in most enterprise and public sectors by passing all losses on of the countries studied, government intervention in to the depositors. Proponents of this strategy often bank restructuring occurred in all eight. There are cite post-World War II German hyperinflation, after good reasons for government intervention in bank which a successful currency reform placed the econo- restructuring. First, unlike enterprise failure, bank my on a strong path of growth and recovery. failure involves thousands of deposit and loan con- The hyperinflation approach was tried in the tracts, which involve extremely high transaction costs 1980s in Argentina, Poland, and Yugoslavia- through the normal process of liquidation or court although in each case it was the consequence of a resolution. number of factors, not a deliberate act of policy- Second, since banks are both the custodians of and in each case it failed because the structural prob- private savings and operators of the payments system, lems within each economy were not corrected (chap- bank failures carry the systemic risk of causing eco- ters 9 and 11). In all three cases hyperinflation was nomic disruption and loss of confidence in the sys- unable to erode external debt and only increased dis- tem itself. The externalities of bank failures are very intermediation from the banking system. In Poland large. and Yugoslavia, where a significant share of deposits Third, the fact that banks are closely regulated with domestic banks were denominated in foreign compared with enterprises gives depositors an currency, hyperinflarion and the accompanying implicit (and sometimes explicit) public guarantee devaluation worsened the internal debt obligations of that the government will safeguard the value of the governments because banks carried an implicit or deposits. Consequently, despite awareness of the dan- explicit state guarantee on these deposits. The gov- gers of moral hazard in the "too big to fail" doctrine, ernments ended up owing large sums denominated governments have intervened in almost all known in foreign currency to their own residents. cases of problem banks. Even after Chile's period of In the absence of inflation accounting, inappropri- doctrinaire free-market policies (1974-80), for exam- ate decisions can be made based on distorted informa- ple, when the banking crisis emerged in 1981 the tion. For example, in the immediate post-inflation government not only gave an explicit guarantee on period (1990) Polish banks and enterprises reported all bank deposits, it also assumed or guaranteed all exceptionally high profits, derived from the sale of private sector external debt to preserve international inventory-without adjusting for replacement costs- 38 BANK RESTRUCTURING and a sharp cut in real wages. The government also enterprise-borrower problem (the supply-side solu- temporarily enjoyed a budgetary surplus because of a tion) or with the banking problem. higher tax on the inflation profits of enterprises. But If the state assists the enterprise-borrower sector enterprise losses reemerged quickly after 1990 because first by directly injecting funds to restore the capital state-owned enterprises did not restructure to interna- base of borrowers, then the bad debts of the banking tionally competitive levels, real wages increased, and system are correspondingly reduced by the amount of lost inventory had to be replaced. state aid. But state assistance to the banking system Thus a stock-flow-consistent framework suggests does not automatically mean that there will be a "pass- that even temporary erosion of the debt of the deficit through" effect, in which the banks pass-through their sectors (the enterprise and public sectors) at the state aid in the form of debt rescheduling, reduction, expense of the surplus (household) sector will not or writeoffs. Recent cases of bank restructuring, partic- work. Structural adjustments to the asset side of the ularly in Eastern Europe, suffer from the complaint by national balance sheet, that is, changes in the real sec- distressed borrowers that the state bailed out banks and tor, must accompany bank restructuring. Bank fail- depositors at the expense of the enterprise sector. ures arise from structural imbalances in the real Sundararajan (1992) suggests the following sector-such as large fiscal deficits or enterprise sequencing for financial sector reforms: losses-and until these are addressed, financial * A minimal system of prudential regulation, restructuring alone will bring only temporary relief, including enforcement, and recapitalization of the The lessons of the Argentine case are important banking system. (chapter 11). Argentina started and ended the 1980s * Key monetary control reforms and support for with banking crises. The primary cause of the early changes in money markets. 1980s crisis was financial liberalization in an era of * The speed of liberalization of interest rates and lax fiscal discipline, overborrowing, and inadequate credit controls depends on the pace of fiscal and bank supervision. These failings were also common external policies in achieving overall macroeco- to neighboring economies, such as Chile and nomic balance. Uruguay. With the cutoff of external credit-due * A wide range of central bank operational reforms, mainly to the debt crisis but also to the government's including policy research, foreign exchange mar- inability to bring its fiscal deficit under control-the ket operations, prudential regulation, clearing and economy reeled into hyperinflation in 1989. Since settlement systems, public debt management, and the state had reached the limits of its debt capacity, accounting and bank reporting systems. the only way to reduce its interest and debt burden While this sequence is appropriate for the finan- was to force the conversion of commercial bank time cial sector, real sector reforms-particularly in the deposits into cash (up to the equivalent of $500) and enterprise, trade, and fiscal areas-are much more dollar-denominated ten-year bonds with two years' difficult to achieve because of entrenched political grace called Bonex. This unilateral writedown of and institutional interests. In the transition socialist public debt to the private sector was a major blow to economies with newly created parliamentary process- the financial system, but it gave breathing space to es, needed changes in legislation and the political the public sector to carry out other real sector debate over privatization and equitable distribution reforms that gradually restored economic growth in of former state property often get entangled in a 1991 and 1992. Brazil also was forced to adopt this morass of debate that delays resolution. drastic solution in 1990. The national balance sheet approach suggests the following: Sequencing ofaction * Rewriting the losses of the banking sector will involve distributing losses across almost all other Bank restructuring cannot be undertaken indepen- sectors-enterprise, banking, government, and dent of real sector (or enterprise) restructuring. even external. Financial reforms made well in advance of real sector * Rewriting the national balance sheet has to be liberalization could lead to subsequent reversals as done in such a way that sector balance sheets are recapitalized banks continue to make loans to loss- stable, sustainable, and credible-shifting large making enterprises. The debate remains open, how- imbalances to a fiscally unsustainable public sector ever, on whether the state should first deal with the would only inflate macroeconomic imbalances. Bank Resnucturing Techniqus 39 * The national balance sheet can only be stable if Second, the authorities "cleaned up" the assets of changes to the balance sheet (that is, flows) follow the former state banks of eastern Germany by issuing the right incentive structure-toward sound risk "equalization claims" to these banks and selling the management, competition, and efficiency. cleaned banks to the larger banks in western * Resolving sector or national losses requires fresh Germany. The debt of eastern state-owned enterpris- resources, but not necessarily additional debt. es were "cleaned" with "equalization claims" or guar- Thus borrowing externally to settle domestic lia- antees issued by the Restructuring Agency, all of bilities could worsen the national balance sheet if which are backed by the federal government. As a external resources leak out in the form of imports result of these initiatives the eastern banking system or capital flight. can function without the burden of past bad debts, Governments faced with difficult choices can opt and the risks against future bad debts are controlled for a Schumpeterian supply side-oriented approach by private banks operating under market rules, as to generate additional resources to deal with bank defined by German law (Meinecke 1991). losses, or use Keynesian demand management to In its August 1992 package the Japanese govern- extract the resources to do so. ment went for measures aimed at fiscal stimulation (box 2.8). Other than some advanced purchase of Supply-side solutions land from the banks by local authorities and a pro- posal to create a quasi-official agency to buy land that In two cases governments with a historical inclination would have been held by banks as collateral against toward industrial policy and strong fiscal capacity nonperforming loans, the package focused on the chose supply-side solutions to deal with potential supply side. In other words, if the burden of nonper- bank problems. In the case of the former German forming loans was due to slow economic growth, Democratic Republic, the German government then a stimulation package aimed at improving pub- avoided problems, first, by restricting entry into the lic infrastructure would generate sufficient growth banking sector with high capital and licensing stan- and profits for the banks to clean out their portfolios. dards (Meinecke 1991; Siebert, Schmieding, and This approach hinges on the belief that there is little Nunnenkamp 1992). No new banks were created in wrong with the structure of the banking system that the five Eastern Landes, and thus Germany avoided correct incentives, improved supervision, and fiscal the problem of failure of new entrants, which has stimulation would not cure. The incentive structure been a common problem in the rest of Eastern and is also supply side, since the fiscal measures reward Central Europe. efficient producers so that profits are generated in the Box 2.8 Japan: A supply-side solution The Japanese rescue package of August 1992 was a sup- to buy developing country debt from Japanese banks. ply-side solution aimed at stimulating the economy to The Cooperative Credit Purchasing Company was relieve banks of their nonperforming loan burden. The established by 162 financial institutions with a capital of package involved a total of Y10.7 trillion ($86 billion), V 7.9 billion to buy land from banks' collateral portfolio comprising V 8.6 trillion of public works spending at book or below-market prices, thus allowing banks to (including V I trillion of advance purchases of land by obtain tax relief on the writeoff of nonperforming loans. local governments) and V 2.1 trillion of lending for By March 1993 the company had bought 229 nonper- small firms and private capital investments. The lending forming loans worth V 681.7 billion at an average dis- and investment portion was designed to assist small count of 34 percent. firms suffering from the credit crunch as banks restrict- In December 1994 the Bank of Japan announced ed credit to prime customers. Funding for the program the creation of a special "rescue bank," using for the first came from the Ministry of Finance's Trust Fund time resources of the Bank of Japan to bail out two Bureau, which is funded largely from postal savings. insolvent credit associations in Tokyo. Of the V 40 bil- The government also asked the public utilities to accel- lion capital of the rescue bank, half was provided by the erate investments of V 700 million and to add Y 1. I tril- Bank of Japan and half by private banks. The Deposit lion from public funds for investment in shares. Insurance Agency will provide Y 40 billion to cover the The Japanese banks have established an asset holding losses of the failed institutions. company along the line of the Japan Bankers Association Investment Inc., an offshore vehicle established in 1987 Source: Numachi 1993; Sasaki-Smith 1994. 40 BANK RESTRUCTUIUNG system. The banks will still try to enforce financial loss burden on to either the borrowers (through high- discipline by weeding out the loss-makers and using er lending rates) or the depositors (through lower their profits to write off the bad loans. deposit rates). Some countries try to protect borrow- ers by placing lending rate ceilings on loans, which Flow and stock solutions forces the losses on the depositors. In 1980-81, when the Republic of Korea's heavy industrialization pro- Government intervention in ailing banking systems gram stalled in the face of the international recession involves one of two key approaches-flow or stock and large debt-servicing problems surfaced in the solutions (box 2.9). Depending on the degree of banking system, the authorities chose not to raise bank distress, intervention can involve a variety of interest rates, which would have been the orthodox these approaches, including a combination of flow stabilization solution. Instead they lowered deposit and stock solutions. and lending rates to negative real levels for two con- Flow techniques. Flow solutions have worked quite secutive years in order to reduce the real burden of well in most mild cases of banking distress. Where debt on Korean enterprises-and passed the losses on real sector problems are the result of a temporary to the depositors. Korea was also able to impose a shock and governance of banks, enterprises, and the financial repression tax on depositors because of tight civil administration are fairly effective, flow solutions exchange controls. The negative effect was the retar- can be useful. The most common flow technique is dation of financial deepening in Korea for a number central bank liquidity support at subsidized rates. For of years, as well as the emergence of a large curb mar- example, the low interest rate regime in the United ket, which had to be addressed separately. States from 1991-93 was largely responsible for In contrast, Chile was unable to impose a repres- recapitalizing the domestic banking system. Between sion tax in the 1981 crisis because of an open capital 1991 and 1993 net interest income and writeback of account. In order to keep savings within the country, provision for loan losses each accounted for about abnormally high real interest rates prevailed-averag- half of the factors contributing to the recovery of ing 46 percent a year in real terms during 1981-85- U.S. banks' earnings (Bank of Japan 1994). thus exacerbating the distress of borrowers. Similar Such techniques have high monetary costs conditions applied in Argentina and Yugoslavia. A because liquidity support for loss-making banks are repression tax would not have worked in Yugoslavia expansionary in nature and counter the need for tight because of the foreign currency deposits in Yugoslav monetary policy, particularly when the economy is banking systems. Attempts to impose heavy financial also undergoing a period of structural adjustment. repression taxes to pay for bank losses would have led On such occasions bank supervision responsibilities to either disintermediation or capital flight. conflict with the central bank's own monetary poli- Another flow solution is to deregulate banking to cies. Some Latin American economies, such as Chile, permit banks to engage in businesses outside their have separated bank superintendency from the cen- traditional avenues of income, in areas such as securi- tral bank to ensure that each objective is carried out ties trading, investment banking, credit card and independently and objectively. travel services, and even insurance. Deregulation is The second most popular flow technique is to still proceeding in a number of countries, but in most permit higher spreads, allowing the banks to pass the cases deregulation should not be attempted without adequate supervision. There are high learning costs associated with new entrants into the banking busi- Box 2.9 Flow and stock solutions ness or diversification into new businesses. In Spain a Flow solutions large number of the new entrants into the banking * Central bank liquidity support at subsidized rates industry in the late 1970s failed in the early 1980s. * Allow high spreads-use of inflation tax In the United States thrifts were allowed to diversify * Regulatory and accounting forbearance out of fixed-rate home mortgages into commercial * Deregulation to allow new income sources and industrial loans, including commercial real Stock solutions estate, and many failed because of insufficient under- * Carve out bad assets to central agency * Capital injection by private or public sector standing of the risks associated with these loans. * Liquidation Community-based thrifts were taken over by new entrepreneurial groups that took more speculative Bank Restructuring Techniques 41 risks under the umbrella of the nationwide deposit bonds or central bank liabilities. Alternatively, the insurance scheme. In addition, supervision of these bad assets can be retained in the books (with a gov- thrifts was relaxed in the early ] 980s in the first flush ernment guarantee) and debt recovery becomes the of deregulation, resulting in greater exposure of the responsibility of the banks. thrift industry to fraud, connected lending, and There are two basic models for the carve-out excessive risk. approach: the U.S. Resolution Trust Corporation Flow solutions work well in a market-based envi- (RTC) and the Spanish Guarantee Fund. The RTC is ronment where the degree of economic insolvency is essentially a liquidation arm of the federal deposit fairly small. The most promising environment for insurance system (chapter 4). It centralizes the dis- flow techniques is one where the capital adequacy of posal of assets of failed thrifts in a corporation run by the problem bank is below statutory levels but is still the U.S. Federal Deposit Insurance Corporation positive, at between 0 and 2 percent of risk assets. (FDIC). Diagnosis, damage control, and discussions Under such conditions, and assuming that bank on the ailing banks are still the responsibility of the spreads can be increased to the extent that banks have FDIC. Funding for the RTC is wholly an obligation a good chance of regaining positive economic solven- of the U.S. government. cy within two or three years, flow solutions can work. T he Spanish Guarantee Fund acts as a bank But allowing spreads to widen runs the risk of "hospital," with the deposit insurance scheme serv- increasing real lending rates to a level where even ing as the vehicle for rehabilitation and liquidation good borrowers begin to fail, or if the burden is (chapter 5). Before the fund steps in, all bank losses passed on primarily to depositors in the form of are applied against the existing paid-up capital lower negative deposit rates, there is a danger of sys- according to the "accordion" principle, so that exist- temic disintermediation. Worse, overall bank efficien- ing shareholders bear the first tranche of bank loss- cy can suffer where high spreads are allowed, since es. Thus the balance sheet of the problem bank is the rent from high spreads can easily be consumed collapsed (like an accordion) depending on the size through higher operating costs instead of being used of the losses. Funding was initially provided joindy to charge off bad debts and build up capital. by the Bank of Spain and the commercial banks, Stock techniques. The stock solution involves three but later the main financial burden was absorbed by techniques that address the balance sheet and capital the central bank. adequacy of the banks. If the problem bank is consid- The carve-out technique has two structural mod- ered viable over the long run, then restructuring els: debt workout by the banks and a total carve-out could involve a capital injection by either the public to a restructuring (debt collection) agency (box or private sector. If the bank is not considered viable, 2. 10). Both techniques try to separate problem banks it should be liquidated, and depositors paid through into "good bank-bad bank" components. Bad assets the deposit insurance fund. The third technique, are concentrated into a "bad bank," with specialized adopted in Chile, Ghana, and Spain, is to carve out staff to deal with the problems. The remaining assets the bad debts of banks in exchange for government are hived into a new institution-supported with Box 2.10 Should carve-outs be handled by a centralized restructuring agency or left to the problem banks? There is ongoing debate over the best model for asset restructuring and workout within one agency. and debt recovery: a decentralized debt workout by Funding can be concentrated in one agency banks or transfer of bad debts to a centralized work- with state guarantee, such as the German out agency. Debt workout by banks assumes that the Treuhandanstalt or the U.S. Resolution Trust banks have sufficient skills and resources to deal with Corporation. Laws may have to be changed, their problems. Although banks have the best infor- however, to allow transfer of debt rights to a state mation on the problems of their borrowers, they may agency. This approach works well where large debt not have the skills required to address these problems. threatens normal operations of the banking system Bank workouts work well with case-by-case resolu- and centralization allows banks to concentrate on tions that have little economywide impact. No their normal banking business. A private sector vari- changes in the law are required. This was the solution ant of this is the Japanese Cooperative Credit adopted in the Polish bank restructuring exercise. Purchasing Company, established by Japanese banks The centralized asset recovery agency approach to buy nonperforming loans from individual banks permits a consolidation of skills and resources in debt at market price. 42 BANK RESTRUCTURING adequate capital-so that the "good bank" can be The fiscal issue became very clear from the eight revitalized and eventually sold to the private sector. cases studied. In the two industrial country cases, The appropriate technique depends on each Spain and the United States, fiscal issues were not a country's institutional, legal, and market conditions. fundamental problem. Both economies had modest Both schemes work well where the legal framework is fiscal deficits that could be increased through addi- strong, where there is a large pool of bank profession- tional borrowing or fiscal adjustments that were not als to staff the complex restructuring and liquidation disruptive to the stability of the financial system. But work, and where the size of the problem is not sys- the six developing countries studied-Argentina, temic in nature. One of the most common com- Chile, Colombia, Ghana, Malaysia, and Yugoslavia- plaints about centralized institutions is that they all suffered from reduced access to foreign financing become administrative bottlenecks in the resolution when the international debt crisis broke in 1982. of banking problems, particularly where asset trans- Almost all had weak fiscal positions (including large fers are embroiled in legal suits. quasi-fiscal deficits) at the beginning of the banking Stock solutions, particularly the carve-out, only problems. In several, large fiscal deficits were a major buy time for problems in the real sector to be cause of the problems in the banking sector. Fiscal addressed, and hopefully correct the resource alloca- and external account adjustments in each country tion and financial stability conditions that jeopardize had severe effects on both domestic growth and the other efforts at macroeconomic stabilization.' Stock stability of the banking system. In Argentina and solutions appear to be very straightforward, with a Yugoslavia central bank losses from foreign exchange financial engineering exercise that cleans the banks at losses were so high that such losses were eventually one stroke and transfers all the bad debts to the state. monetized, resulting in hyperinflation. The vicious Successful carve-outs, however, have required extreme- circle of high debt, capital flight, recession, unem- ly large budgetary resources. Both the U.S. RTC and ployment, and widening fiscal deficits eventually led the German Treuhandanstalt operate in well-estab- to loss of monetary control. lished market economies, with strong liquidation and In Chile, even though central bank losses were debt recovery laws and skilled bank turnaround spe- high from the absorption of bank losses, the correc- cialists. Such budgetary and human resources are not tion of the central government deficit through tax available to many developing countries. reform and radical reform (including privatization) of An insolvent institution cannot be rescued by the social security system generated sufficient surplus another insolvent institution. Merging two weak in the budget not to disturb monetary balance. institutions such that their combined capital is still Together with devaluation, trade liberalization, and negative is not a long-term solution. Similarly, a bad the opening of the economy to foreign direct invest- debt carve-out at the banking system level will not ment, the real economy turned around sufficiently to solve the problem if the central budget cannot mobi- restore stability in the banking system without a lize sufficient resources to service the additional debt recurrence of inflation (chapter 10). burden. Many developing country banking crises Some form of stock solution involving direct gov- ultimately are a fiscal crisis. If the state can generate ernment intervention is probably inevitable when the sufficient resources to pay for the losses of the bank- economic insolvency of the banking system is greater ing system through additional taxes, spending cuts, than 2 percent of risk assets.2 The calculation of the or borrowing, then there is sufficient time for the fiscal cost to the government depends on the size of state to tackle the root causes of bank losses- the domestic and foreign debt, the total fiscal and whether at the institutional level, within the banking quasi-fiscal deficits, and the costs of servicing such system, or in the real sector. But if the fundamental debt. Experience suggests that countries with moder- problems stem from the fiscal imbalance itself, or if ate debt levels and deep financial markets-which the state cannot mobilize sufficient tax resources to allow the government to obtain financing at low costs service the bank losses without resorting to inflation- and long maturity-are likely to weather financial ary financing, then the carve-out may itself become crises better than governments with shallow financial destabilizing. The public will not believe that the markets. For example, in 1987 Malaysia had external public sector has tackled the roots of the problem debt and fiscal deficit ratios comparable to those of and may engage in capital flight, further exacerbating Argentina (75 percent and 8 percent of GDP, respec- the problem of bank instability. tively), but Malaysia had already begun to deal Bank Restructuring T7chniques 43 decisively with its bank restructuring through direct would require a minimum 5 percent increase in rev- central bank assistance to ailing banks in a noninfla- enue or an equivalent reduction in expenditure, tionary manner (chapter 7). The solution lay in high considering that total government revenue is only domestic saving rates (28.0 percent of GDP in 1989, 14 percent of GDP. This approach provokes the compared with 5.2 percent for Argentina) and a deep usual arguments against taking any action in bank domestic financial market that allowed domestic bor- restructuring: rowing with up to thirty years' maturity. In contrast, * The government cannot afford the losses.3 as domestic disintermediation occurred, debt-servic- * The banks should be able to make profits in the ing burdens became unsustainable in high-inflation future to regenerate their capital. economies with shorter maturities and higher real * Since the banks are responsible for the losses, they interest rates. should do the cleaning up themselves. The fiscal costs of bank restructuring can be cal- What the authorities did not realize was that culated from the following example. Country A is a real interest rates began to rise as bank losses accu- post-centrally planned economy in transition to a mulated because depositors demanded a risk premi- market economy. Its domestic debt to GDP ratio is um from what they perceived as a potentially 47 percent and its foreign debt to GDP ratio is 28 unsafe banking system. The combination of rising percent, with an annual fiscal deficit of 7 percent. losses (adding to the quasi-fiscal deficits) and rising Domestic bank credit to the private sector amounts real interest rates, as well as declining growth due to 30 percent of GDP. Recent studies indicate that to misallocation of resources, would increase the the banking system has nonperforming loans total- ratio of total debt to GNP (d) until it reached ing at least 20 percent of its portfolio, equivalent to unsustainable levels (Fischer and Easterly 1990; see 6 percent of GDP. If the government undertook a box 2.6). carve-out, the fiscal costs would depend on the level There is no alternative in macroeconomic stabi- of interest paid on the government bonds. Current lization efforts to privatization, cutting back gov- lending rates are 17 percent a year, while bank ernment spending to sustainable levels, allowing deposit rates are 12 percent, slightly below the real sector liberalization in trade, opening to for- inflation rate. Assuming that the government eign direct investment, and a synchronized package wished to issue bonds at a rate equal to the deposit of monetary and financial reforms to generate the rate, the annual fiscal cost would be about 0.7 per- growth that brings economic growth (and hence cent of GDP (12 percent times 6 percent). This the level of d) back to a sustainable level. Box 2.11 Bank restructuring mechanisms UK model(Lifeboat Fund, 1974) * Liquidation or sale of banks to private sector. * Funded by large clearing banks and the Bank of * Losses borne by RTC (funded by federal guarantee). England. * Initial liquidity support for viable secondary banks. Spanish model (bank "hospital" and carve-out mech- * Failed banks liquidated. anism) * Bank of England took over a failed bank that was -Accordion principle. subsequently privatized; losses were borne by the cen- - Joint funding by commercial banks and the Bank of tral bank. Spain. * Deposit Guarantee Fund buys bad assets. US. model(deposit insurance) * Provides banks with guarantees and long-term soft Federal Savings and Loans Insurance Corporation loans. (until 1989) * Sale of banks to the private sector. * Acquisition or merger. * Nationalization of the Rumasa group. * Income maintenance program. * Accounting forbearance. Chil variation * Phoenix and bridge banks. * Central bank issues bonds to buy bad assets, with * Management consignment. buyback schedule. Resolution Trust Corporation (after 1989) * Central bank loans to banks converted into equity. * Concentration of failed assets in RTC. * Sale of banks to the private sector. 44 BANK RESTRUCTURING Models of Bank Restructuring until the 1980s, when excessive competition, changes in market structure, speculation in real estate, mis- Beginning in the 1970s, several models of bank management and fraud, and inadequate supervision restructuring appeared (box 2.11). The first industrial led to unprecedented failures, first in thrifts and then country model was the market-based flow solution in banks. adopted in the U.K. secondary banking crisis of The Federal Savings and Loans Insurance 1973-76. A Lifeboat Fund was created with joint liq- Corporation (FSLIC), which became defunct in uidity funding by the large clearing banks and the 1989 when it became insolvent, applied a number of Bank of England. Twenty-three "fringe" banks, of market-based but government-assisted acquisition which eighteen were deposit-taking institutions and and merger solutions to deal with failed thrifts. five were authorized banks, were badly affected by the Liquidation and direct deposit payout was often the 1973-74 recession, with a number overspeculating in last-resort solution. The preferred method was acqui- the property market. Most were reconstructed or sition of the failed institution, followed by merger or merged into other companies, while a few eventually sale of the "cleaned" institution to a private buyer. were liquidated or placed into receivership. None of When FSLIC resources were inadequate to do stock the major clearing banks were affected. Total support cleanups it resorted to income maintenance programs was £ 1.2 billion, equivalent to 40 percent of the in which it made up the income shortfalls between total capital reserves of all English and Scottish clear- the yield on the assets of the failed bank and the mar- ing banks. The primary burden of losses was borne ket yield agreed with the buyer. These later turned by shareholders of the failed institutions, with some out to be extremely expensive solutions, since the liquidity support from the Lifeboat Fund. Some FSLIC bore most of the risks of subsequent failure. residue losses arising from the takeover of one bank When the number and size of thrift failures con- and subsequent resale were absorbed by the Bank of tinued to increase in the second half of the 1980s, the England. FSLIC began to create "phoenix" or "bridge" banks The Thailand Financial Institutions Development by consolidating failed thrifts for subsequent resale. Fund, created in 1981, was one of the first develop- The corporation also engaged in management con- ing country bank restructuring institutions signments, where the management of problem banks (Sundaravej and Trairatvorakul 1989). It was mod- was give to bidders or established banks with proven eled partly on the U.K. model, with funding from management records. By'1989, however, thrift losses mandatory contributions from financial institutions had became so large that the FSLIC became insol- and borrowing from the Bank of Thailand. The fund vent. It was dissolved and the funds were reconstitut- is managed by the Bank of Thailand. It is not a ed in the Saving and Loans Insurance Fund under deposit insurance scheme, but it has the flexibility of the charge of the Federal Deposit Insurance using its funds to recapitalize (buy equity in) problem Corporation (FDIC) (chapter 4). financial institutions or of giving financial assistance The FDIC used broadly similar techniques of pur- to depositors of problem banks. chase and assumption in dealing with problem banks, The United States has the most complex and but because it had greater resources than the FSLIC- comprehensive deposit insurance scheme. The both financially and managerially-it was able to scheme, which is market funded but government supervise and manage the rising number of failed and guaranteed, was established in the 1930s after the problem banks in an orderly manner. The 1989 massive bank failures of the Great Depression. The Financial Institutions Reform, Recovery, and scheme is unique in that it fostered the freedom of Enforcement Act delegated to the FDIC the responsi- entry and operations of more than 29,000 deposit- bility of running the thrift insurance fund, which was taking institutions, mostly small and with state or also in charge of the newly created Resolution Trust national charters. Since these operate in highly geo- Corporation (RTC). The RTC was created because graphically and functionally segmented markets with- the number of failed thrifts and their assets had grown in the United States, the scheme institutionalized the so large that it was administratively and legally over- orderly exit of failed banks without disrupting depos- whelming for the FSLIC to handle. The RTC was itor confidence. Supervision of deposit-taking institu- given the task of managing and resolving failed thrifts tions is divided among a complex of state and federal in a market-based and orderly manner that would agencies (chapter 4). The system functioned well minimize government losses. In mid-1990 the RTC Bank Restructuring Techniques 45 had more than 200 institutions under "conservator- The carve-out technique was widely adopted in a ship," that is, as ongoing businesses to be sold in part number of countries, notably Chile, Colombia, and or whole; and another 200 under receivership, that is, Ghana and more recently in Hungary, Slovenia, and for liquidation in full or in part. By the end of 1991 the Czech and Slovak Republics. The Chilean bank- the RTC had assumed more than $357 billion in ing crisis of 1975-85 was one of the most severe in assets of failed thrifts and disposed of $229 billion. the 1980s. The government intervened in thirteen of The RTC asset disposal model was adopted by twenty-five domestic banks (including two of the Ghanaian authorities, who created the Nonperforming largest banks) and six of eighteen financial compa- Asset Recovery Trust (NPART) in 1989 to deal with nies, covering 45 percent of total bank loans. the recovery and disposal of bad debts and their coUlat- Nonperforming loans reached as high as 113 percent eral. NPART is still engaged in asset disposals. of total bank capital and reserves by May 1983 and Asset recovery trusts are useful where the number bad loans sold to the central bank amounted to 24 and size of banks' bad assets are so large that they are percent of total loans. administratively cumbersome to handle in a decen- The Chileans began addressing their banking cri- tralized manner. Asset recovery trusts also can be a sis by converting emergency loans to banks into equi- useful transitory tool where enterprise restructuring ty, including heavy central bank subsidies to banks and debt recovery specialists are scarce and court pro- and finance companies to allow rescheduling of cedures do not enforce quick debt recovery and dis- loans. A major step was the central bank purchase of posal of collateral. bad loans, initially up to total bank capital and Problems with asset recovery trusts can occur reserves, by swapping them for noninterest-bearing because the information capital of banks on their bor- central bank bonds, with a repurchase schedule of 5 rowers often is destroyed or lost in the transfer. The percent buyback every six months for ten years. In transfer of rights of banks over their borrowers to the 1984, with a deepening crisis, this method was asset recovery trust may not be legally perfect, and the revised to cash purchase of bad loans up to 150 per- detailed knowledge of hidden assets of borrowers by cent of capital and reserves. Loans sold for cash had individual banks may not be passed on to the trust, to be rebought over ten years at a 5 percent real rate. thus inhibiting the debt recovery process. Thus it is Assisted banks could not declare dividends until the not yet entirely clear whether centralized asset trusts central bank's debt was fully repaid. or decentralized debt recovery is preferable. The Chileans forced the problem banks to The Spanish Guarantee Fund-created in increase their capital, first from existing shareholders, 1980-was a major institutional improvement over then from new subscribers. When private sector the U.S. and U.K. models since it embodied explicit interest failed to develop, the state development intervention by the government with a recognized agency Corfo bought equity (Corfo was limited by "carve-out" technique. Instead of intervening at the law to 49 percent equity ownership). Corfo's shares margin through liquidity support (as in the U.K. were to be resold to the private sector within five model) or purchase, assumption, and resale through years, with any unsold balance to be transferred back the deposit insurance fund (as in the U.S. model), to shareholders that had contributed to the equity the fund operated as a bank hospital. Funding was increase. shared by the larger commercial banks and the Bank The Chilean technique was quite successful-the of Spain, but the supervision and referral of problem banking system stabilized and the economy recovered banks to the hospital was left to the Bank of Spain. after massive structural adjustments in the fiscal and The technique used was, first, the application of the trade sectors. But the burden of bank losses fell accordion principle: a full capital reduction to the almost exclusively on the central bank, which in extent of losses so that the shareholders bore the August 1985 held bad debt purchases of $2.36 bil- brunt of the losses; and second, the fund would pur- lion and internal debt equivalent to 44 percent of chase the bad assets in exchange for bonds. GDP Sometimes the fund provided guarantees or long- The Philippines also successfully applied the bad term soft loans to the problem institution, but once debt carve-out in relieving the Philippines National the institution was "cleaned up" it was resold to the Bank and the Development Bank of the Philippines. private sector, sometimes to foreign banks interested About 104 billion pesos-including $5.1 billion of in entering the Spanish market (chapter 5). external debt-equivalent to 16.6 percent of GDP 46 BANK RESTRUCruRING were transferred to the national government in 1986 mentation of measures for damage control. Bank (Lamberte 1989). Both banks were then restructured restructuring is therefore often part of a larger pack- and partially privatized. age of macroeconomic stabilization and reform. While the carve-out remains a popular technique, As a general rule, if banking problems are detect- it is not always the best solution. In Malaysia, ed early, with a reasonably stable macroeconomic Thailand, and the United Kingdom the bad assets environment and effective administrative and super- were retained in the banks' books and central bank visory machinery, flow solutions will be adequate to assistance was provided in the form of loans deal with banking system that still have 0 to 2 per- (Thailand and the United Kingdom) or equity cent capital-asset ratios. If problems are allowed to (Malaysia). deteriorate beyond this level, a combination of stock In the 1985-88 Malaysian banking crisis the cen- and flow solutions will have to be implemented, tral bank intervened in four of thirty-nine banks, depending on the size of the distress. A summary of four of forty-seven finance companies, and thirty-two the conditions for successful bank restructuring are of thirty-five deposit-taking cooperatives involving listed in box 2.12. M$ 9.4 billion, or 10.4 percent of total deposits. Middle- and high-income countries with good Total financial institution losses in 1985-86 totaled accounting standards, ample resources, and effective 4.7 percent of 1986 GNP To handle the crisis, emer- bank supervision agencies (whether in the central gency legislation empowered the central bank to bank, bank superintendency, or deposit insurance freeze the deposits of failed deposit-taking coopera- agency) have been able to restructure their banks rea- tives and merge twelve failed cooperatives into one sonably successfully. In dealing with problem banks, licensed finance company. The central bank provided these countries adopted bank restructuring in tandem soft loans to assist the rehabilitation and refund of with enterprise and fiscal reforms, and it was the suc- depositors, who received 50 percent in cash and 50 cess of real sector reforms that generated the resources percent in shares in the "phoenix" finance company. to clean up the banking system. The central bank also provided M$ 672 million in In Malaysia, Spain, and the United States the bor- equity to the four banks and four finance companies, rowing capacity of government (including the credi- as well as soft loans of M$ 1.3 billion. More recent bility of the central bank) was such that absorbing cases of bank restructuring, particularly those in bank losses as part of the budget or as a quasi-fiscal Eastern Europe, are discussed in chapter 3. deficit within the central bank did not disturb mone- tary stability and hence did not lead to inflation. In Conditions for Successful Bank Restructuring all three countries bank losses were partly due to dis- inflation or asset compression, as competition, cor- The choice of the appropriate restructuring mecha- rections for overlending, and falling real estate prices nism depends on several key factors: the size of the created large losses in parts of the banking system, losses and the political economy of loss allocation, particularly in the more highly geared, managerially the legal and institutional framework, problems of weaker new entrants. the real sector, and the effectiveness of the imple- The other five countries (Argentina, Chile, menting agency. Timing is also important. The Colombia, Ghana, and Yugoslavia) all suffered urgency of the pending problems, the pressure brought on by specific events or macroeconomic pol- Box 2.12 Conditions for successful bank restructuring icy considerations, and the proper sequencing and . . *~~~~~~~~~~~~ Clear links with enterprise restructuring. policy mix also affect the choice and outcome of * Stable macroeconomic environment. bank restructuring schemes. * Strong political will. The objective of bank restructuring-to restore * Effective restructuring agency. sustainable bank solvency and profitability-is relat- * Transparent accounting standards. ed to the restoration of solvency and viability of the * Legal framework favoring financial discipline. real sector. If the real sector continues to suffer large * Incentives favoring private sector growth and losses, the banking sector cannot be isolated from competition. such losse, and may be a culprit in continuing to - Cadre of professional bankers (domestic or such losses, and may be a culprit in continuing to foreign). finance real sector losses. Successful bank restructur- * Effective bank supervision and enforcement. ing calls for a recognition of these losses and imple- Bank Restcuring Techniques 47 from large external debt overhangs that developed wipe away the fiscal burden. The government had to in the early 1980s. Ghana was able to embark on a remove its debt burden by transferring its banking path of macroeconomic adjustment after 1983 debt into dollar-denominated long-term debt, which with substantial external aid. The success of its fis- reduced the fiscal debt-servicing burden and passed cal reforms generated a primary surplus that, the losses to the future. Finally, massive privatization together with external aid, allowed a carve-out to and efforts in fiscal, trade, and enterprise reform- be executed in the banking system to remove bad including a major currency reform in 1991-turned debt from the banks without generating any mone- the tide of confidence and restored stability in the tary pressure. Argentine economy. Chile and Colombia both suffered from the These eight cases demonstrate the importance of effects of rapid liberalization in the financial sector, political will, or in the language of structural adjust- resulting in overlending and the shocks of the cutoff ment loans, "ownership" of the adjustment program. of external resources after the 1982 debt crisis. Since much of bank debt is in domestic currency, Colombia was able to maintain growth and absorbed external aid and technical assistance are neither neces- the losses within the central bank. The shock to sary nor sufficient conditions for successful bank Chile was significantly larger, and the carve-out restructuring. But where there are large overhangs of imposed an extremely large burden on the central external debt, external debt restructuring can relieve bank that has yet to be eliminated, despite the the government's debt-servicing burden, and hence restoration of strong growth. its capacity to absorb further internal debt arising Argentina and Yugoslavia demonstrate the prob- from bank restructuring. lems of attempting bank restructuring without deal- The importance of a single effective restructuring ing with structural issues. In Yugoslavia the experi- agency-whether in the central bank, ministry of mentation with self-managed enterprises that both finance, or another agency-is demonstrated by the owned and borrowed from banks was a structural gaps and conflicts in supervision in a number of failure. Banks that were owned by worker-managed cases. The creation of deposit insurance schemes with enterprises could not impose financial discipline on insufficient resources or legal powers to deal with the loss-making enterprises and instead became the trans- problems can be disastrous. These institutions give mission mechanism of massive resource misalloca- the illusion of a responsible agency without the sub- tion. Even though the federal government did not stance. Deposit protection agencies in Kenya and the ostensibly have any fiscal deficits, the quasi-fiscal loss- Philippines were not provided with sufficient es of the national bank and the losses of enterprise- resources to deal with the rising level of bank prob- owned banks were huge by any standard. Such losses lems, and in the end the rescuer had to be rescued. could not even be removed through hyperinflation. This was also the fate of the U.S. FSLIC. The political difficulties of dealing with such losses This chapter has demonstrated the importance of were inherent in the regional fragmentation that diagnostics and loss allocation. Having determined occurred in 1990. the sources of losses of the banking system, an appro- Argentina began and ended the 1980s with eco- priate mix of policies and measures has to be found nomic and banking crises. In both cases the credibili- to deal with the losses. Where the state has preferred ty and sustainability of fiscal discipline was at stake. to preserve the stability of the banking system and With the cutoff of external financing, the govern- the nominal value of bank deposits, banking system ment resorted to internal financing and borrowed losses are ultimately state obligations. Resources--in heavily from the banking system. The central bank the form of renewed growth, savings, and fiscal rev- incurred large losses from its net foreign liabilities, enue-will have to be found to pay for the losses. and attempts to impose monetary stability failed The rewriting of the national balance sheet has to be because monetary policy could not overcome fiscal undertaken in such a way that the incentives struc- indiscipline. In the end the banking system chan- ture does not permit a repeat of bank failure. Such a neled more than half its resources to funding the fis- rewriting involves fundamental change in the effi- cal deficits of the government and the quasi-fiscal ciency and productivity of the real sector. Unless the losses of the central bank. Disintermediation was real sector returns to growth and stability, there may massive despite real interest rates exceeding 30 per- be insufficient resources to remove the bad debts cent a year. In the end even hyperinflation could not from the banking system. 48 BANK RESTRUCTURING Notes Lamberte, Mario. 1989. "Assessment of the Problems of the Financial System: The Philippines Case." Background ]. Financial transactions to relieve enrerprises or banks of bad paper prepared for World Development Report 1989. World debts also have incentive problems. If relief of the debt burden Bank, Washington, D.C. hardens the budget constraint for enterprise and bank manage- Meinecke, Hans-Dieter. 1991. "The Restructuring of the East ment, this is a positive development. This appears to have been German Banking System." Deutsche Bundesbank, Berlin. the case in the Republic of Korea. If debt relief encourages moral Minsky, Hyman P 1977. "A Theory of Systematic Fragility." In E. hazard behavior, as appeared to have occurred in many Latin I. Altman and A. W Sametz, eds., Financial Crises: Institutions American countries, bailouts may have very high costs. and Markets in a Fragile Environment. New York: Wiley. 2. Recent U.S. experience suggests that banks are economi- Numachi, T 1993. "The Cooperative Credit Purchasing Co., cally insolvent (negative capital) when accounting (capital) sol- Ltd." World Bank, Financial Policy and Systems Division, vency falls below 2 percent of risk assets. Accordingly, the 1992 Washington, D.C. Federal Deposit Insurance Improvement Act mandates that Sasaki-Smith, Mineko. 1994. "The Bank of Japan's New Bailout supervisory authorities take statutory action againsr banks when Scheme." Morgan Stanley, Tokyo. their capital adequacy falls below 2 percent. Schwartz, Anna J. 1985. "Real and Pseudo-Financial Crises." In 3. Fiscal revenue from bank profit taxes in a number of Forrest Capie and Geoffrey Wood, eds., Financial Crises in Eastern European economies in transition was as much as 15 the World Banking System. New York: St. Martin's Press. percent of total revenue. But such profits were calculated before Sheng, Andrew. 1992. "Bad Debts in Transitional Socialist adequate provisions were made for bad debt, so that banks were Economies." World Bank, Financial Policy and Systems overtaxed and decapitalized by bad accounting. Division, Washington, D.C. Sheng, Andrew, and Yoon Je Cho. 1993. "Risk Management References and Stable Financial Structures." Policy Research Working Paper 1109. World Bank, Washington, D.C. Bank of Japan. 1994. "How U.S. Commercial Banks Dealr with Siebert, Horst, Holger Schmieding, and Perer Nunnenkamp. the Asset-Quality Problem." Quarterly Bulletin (May). Tokyo. 1992. "The Transformation of a Socialist Economy: Barnes, Guillermo. 1 992. "Lessons for Bank Privatization in Lessons of German Unification." In Georg Winckler, ed., Mexico." Policy Research Working Paper 1027. World Central and Eastern Europe: Roads to Growth. Washington, Bank, Washington, D.C. D.C.: International Monetary Fund. Brewer, Elijah, and Thomas Monschean. 1992. "Ex Ante Risk and Solheim, Jon A. 1990. "The Banking Crisis in Norway." Paper Ex Post Collapse of S&Ls in the 1980s." Ec&nomic Perspectives presented at the Economic Development Institute seminar (July/August). Federal Reserve Bank of Chicago. on financial sector liberalization and regulation, June, Caprio, Gerard, Izak Atiyas, and James Hanson. 1993. Boston. "Financial Reform Lessons and Strategies." Policy Research Sundararajan, Vasudevan. 1992. "Financial Sector Reforms and Working Paper 1107. World Bank, Washington, D.C. their Appropriate Sequencing." International Monetary Chamely, Christophe, and Patrick Honohan. 1990. "Taxation of Fund, Monetary and Exchange Affairs Department, Financial Intermediation." Policy Research Working Paper Washington, D.C. 421. World Bank, Washington, D.C. Sundararajan, Vasudevan, and Tomas Balifno, eds. 1991. Banking de Juan, Aristobulo. 1987. "From Good Bankers to Bad Bankers: Crises: Cases and Issues. Washington, D.C.: International Ineffective Supervision and Management Deterioration as Monetary Fund. Major Elements in Banking Crises." Economic Sundaravej, Tipsuda, and Prasarn Trairarvorakul. 1989. Development Institute Working Paper. World Bank, "Experiences of Financial Distress in Thailand." Policy Washington, D.C. Research Working Paper 283. World Bank, Washington, - 1991. "Does Bank Insolvency Matter? And What to D.C. Do About It?" Economic Development Institute Working Tenconi, Roland. 1989. "Restructuring of the Banking System Paper. World Bank, Washington, D.C. in Guinea." Background paper prepared for World Diaz-Alejandro, Carlos. 1985. "Goodbye Financial Repression, Development Report 1989. World Bank, Washington, D.C. Hello Financial Crash." Journal ofDevelopment Economics. U.S. Comptroller of the Currency. 1988. "An Evaluation of the Fischer, Stanley, and William Easterly. 1990. "The Economics of Factors Contributing to the Failure of National Banks." the Government Budget Constraint." World Bank Research Washington, D.C. Observer 5(2):127-42. U.S. FDIC (Federal Deposit Insurance Corporation). 1991. GBIF (Guarantee Bank Insurance Fund). 1992. Annual Report. Quarterly Banking Profile, Fourth Quarter. Washington, D.C. Norway. U.S. Treasury. 1991. "Modernizing the Financial System." Glaessner, Thomas, and Ignacio Mas. 1991. "Incentive Structure Washington, D.C. and Resolution of Financial Institution Crises: Latin World Bank. 1993. "Bank Supervision: Suggested Guidelines." America Experience." Latin America Technical Department Financial Policy and Systems Division, Washington, D.C. Technical Paper. World Bank, Washington, D.C. Zahler, Roberto. 1993. "Financial Sector Reforms and Kane, E. J. 1985. The Gathering Crisis in Deposit Insurance. Liberalization: Welcome Address." In Shakil Faruqi, ed., Cambridge, Mass.: MIT Press. Financial Sector Reforms in Asian and Latin American Kindleberger, Charles P. 1978. Manias, Panics, and Crashes: A Countries. Economic [)evelopment Institute Seminar History of Financial Crises. New York: Basic Books. Series. Washington, D.C.: World Bank. CHAPTER 3 Resolution and Reform: Supervisory Remedies for Problem Banks Andrew Sheng Banking is essentially a service business of intermedi- banks transmit funds on behalf of their customers- ation. Banks fail because of human errors or faults, the payments system function. Bank failure is funda- whether fraud, mismanagement, poor policies, or mentally the contractual failure of a bank to honor its inadequate supervision. The human element in obligations to its depositors and results primarily from banking looms so large that bank supervisors often the related failure of borrowers to honor their contracts blame bank failures on bank management-and in to the bank (counterparty risks) or from the impact of certain cases they may be absolutely right. But the market changes on the bank's net worth (market risks). simultaneous eruption of so many bank failures inter- Moreover, since any financial transaction involves nationally in the 1980s suggests that there are larger payment and settlement, there are risks associated macroeconomic reasons for widespread bank distress. with settlement and operations. Liquidity risk is a It is no coincidence that bankers worldwide engaged major problem in any market that dries up in times in a credit frenzy, creating the environment for and of uncertainty. Legal risks occur when property conditions of loss. As chapters I and 2 indicated, the rights are uncertain or undefined. A computer failure roots of bank failure are multifaceted, with structural, may delay payment or accounting, thus subjecting systemic, policy, and managerial dimensions. some party to loss. Finally, systemic, aggregation, or At the most basic level, banks intermediate interconnection risks occur when the failure of one between two sets of contracts-deposit contracts party triggers systemwide failure through contagion. between depositors and the bank and loan contracts Financial market risks are defined in box 3.1. between borrowers and the bank. In addition to these All financial products are essentially contracts savings mobilization and credit allocation function, involving the exchange of property rights or value. Box 3.1 Financial market risks Counterparty credit risks-the risk that the counter- Liquidity risk-losses that result if forced to sell party will fail to fulfill the (credit) contract. The size of under illiquid market conditions. the loss is the replacement cost of the contract in the Legal risk-losses caused by uncertainties in the legal market. definition of obligations or court reversals of commonly Market risks-risks arising from market price understood obligations, such as the legal obligations of changes, such as interest rate risk, exchange rate risk, multilateral netting. and commodity price risk. Aggregation risk-sometimes called systemic or inter- Settlement risk-the risk that one party (or agent connection risk. Failure of one party triggers failure else- bank) will not settle or deliver final value when settling where in system (for example, contagion). a contractual obligation. Operating risk-losses due to inadequate internal Source: Federal Reserve Board, Federal Deposit Insurance controls, procedures, and operating equipment, soft- Corporation, and U.S. Comptroller of the Currency ware, and systems. 1992. 49 50 BANK RESTRucrURlNG Financial markets trade these contracts between the system as a whole. Without such mechanisms in different parties. Bank supervision, in a narrow sense, place, bank crises may last much longer than neces- involves ensuring that banking contracts are honored sary (Zahler 1993). and enforced according to legal guidelines. Since mar- In many countries the rising demand to protect ket transactions involve the risks listed above, market small depositors and ensure systemwide stability participants (banks and their clients) and bank supervi- has been extended-through the "too big to fail" sors must be concerned with risk management. principle-to an implicit or explicit guarantee of the This chapter explores the techniques of bank fail- entire deposit base. This need to protect depositors, ure resolution and restructuring. Specifically, it looks and consequently the banking system, generates the at damage control at the microeconomic level, the dangers of moral hazard, the socialization of losses, need to rebuild bank profitability and create the right and the privatization of gains. Hence the drive to incentives for monitoring a sound financial base, and strengthen supervision of the financial sectoi to finally the need to address the critical link between avoid the high costs of bank failure and disruption. bank and enterprise restructuring. This push for stronger supervision has paralleled the trend toward financial sector liberalization and seems Controlling Damage through Supervision contradictory to a key lesson of the 1980s-that even the most sophisticated bank supervisors have Bank supervision has four objectives: been unable to prevent bank failures. * Competition and operational efficiency. What then, is the role of the modern bank super- * Safety and soundness. visor? The supervisor's primary role is damage con- * Monetary policy and allocative efficiency. trol and getting the incentives right. Banks are in the * Protection of small depositors (equity). business of risk intermediation, but the deposit Problem banks develop for a variety of reasons, macro- insurance schemes that are in place around the world economic as well as microeconomic. Overcapacity in have led to bank risk-taking that is asymmetric: banking, excessive competition, operational inefficien- speculative risks are taken for private gain at public cies, repressive taxation, structural rigidities, portfolio cost. Where financial sector liberalization programs concentration, speculation, mismanagement and have failed to take account of this asymmetry, fraud, external shocks, policy mistakes, and inadequate reforms have generally failed. Banking crises have supervision all contribute to bank insolvency and poor convinced policymakers of the need to level the play- performance. ing field and to evenly match risks with rewards or Some objectives of bank supervision are conflict- punishments. A corollary of this objective requires ing: many government-directed credit policies with bank supervisors to minimize moral hazard behavior, welfare objectives impose costs, risks, and perverse connected lending, conflicts of interest, fraud, and incentives on banks with disastrous impacts on their mismanagement. Accordingly, financial sector liber- net worth. Similarly, tight monetary policies that alization is going through a phase of re-regulation, require banks to maintain large non-interest bearing with broader coverage extending not only to the statutory reserves tend to erode banks' profitability. banking sector but also to nonbank financial inter- There is a tradeoff between safety and efficiency. mediaries. International cooperation in financial sec- Some bank supervisors allow the banking system to be tor supervision also has become necessary as finan- concentrated in banking cartels, which generate high- cial services become globalized. er profits and capital through monopoly profits that Bank supervision in the 1990s may focus on can cushion the system against periodic external national risk management: setting the rules of the shocks. Smaller and less-capitalized banks could easily game for financial intermediaries in such a way that be absorbed by the stronger banks during a recession. risks are taken with an adequate level of capital Bank solvency is protected at the expense of competi- (Sheng and Cho 1993). In other words, financial tion and efficiency. intermediaries should be free to engage in asset liabil- Supervisors who are concerned with competition ity management that allows them to diversify their and efficiency may have exit mechanisms already risks by region and by sector while matching their built into the system. The deposit insurance scheme interest rate and exchange rate risks, so long as they in the United States, for example, permits the order- are supported by a level of capital sufficient to cush- ly exit of inefficient institutions without disrupting ion them against their risk profile. Resolution and Reforn: Supervisory Remediesfr Problem Bankl 51 Thus bank supervision helps banks (and other is inadequate, high nominal capital-asset ratios are financial intermediaries) define and better manage deceptive. Thus the correct cushion against nonearn- their risks, given their capital adequacy. To supple- ing or nonperforming assets is non-interest bearing ment the safety and soundness of the system, supervi- liabilities. A simple rule of thumb to ensure stability is sion should be augmented by a lender-of-last-resort that total core capital (excluding interest-bearing sub- system or some form of deposit insurance (Talley and ordinated loans) plus total provisions against nonper- Mas 1990). Damage control of ailing financial insti- forming loans should at least equal the total stock of tutions begins with diagnostics-the assessment and nonperforming loans (Sheng 1991). evaluation of problem institutions. Asset quality. The greatest danger to capital ero- sion is poor asset quality. Asset quality is based on the Assessingproblem banks quality of credit evaluation, monitoring, and collec- tion within each bank. What assets are at risk? Are Most bank supervisors have broadly adopted the U.S. assets too concentrated in one geographical or eco- CAMEL method of assessing bank performance: nomic sector? Is there connected lending? Are there capital adequacy, asset quality, management quality, hidden losses "below the line?" Are losses hidden in earnings, and liquidity. The Canadian Office of the bank subsidiaries and affiliates? Are there unreported Superintendent of Financial Institutions has refined fraud losses? Have apparently diversified loans been this system to CAMELOT, adding the quality of collateralized on a single large asset? Have adequate operations and treasury management (McPherson provisions been made against potential losses in the 1992). asset portfolio? These are some of the questions that Capital adequacy. Since the Basle Committee on supervisors should ask bank management. Banking Supervision agreed on a uniform standard of Asset values are subject to greater potential losses risk-weighted capital adequacy of 8 percent (adopted in a volatile price environment, not simply from the internationally at the end of 1992), most developing changes in supply or demand, but also from the illiq- countries are moving toward the 8 percent target. uidity in markets of certain assets during times of There is, however, a tendency to regard this standard stress. For example, when markets are thin, as in a as a maximum rather than as a minimum. Most newly introduced government bond market, bond developing country banks require a substantially prices can fall sharply if there are no market makers higher capital base, together with higher standards of or if the economy is uncertain. Thus, even if the loan or asset provisioning, because of their higher credit risks on a particular financial instrument are risks. It should be remembered that until deposit low, the holder may still be subject to significant liq- insurance was introduced in the 1930s, U.S. banks uidity and interest rate risks. had capital-asset ratios well above 20 percent in the Bank supervisors should improve the accounting nineteenth century and between 15 and 20 percent treatment or value measurement of various assets and until the 1930s. (The substantial variation in the liabilities, both tangible and intangible. Values (and capital-asset ratios of international banks can be seen thus risks) of derivatives are also conceptually diffi- in annex 1.2.) cult to quantify and measure. As Goldstein and There is not only significant international varia- Folkerts-Landau (1993, p.23) emphasize, "rapid tion in capital-asset ratios, but also varying defini- expansion of, and concentration in, a particular tions of capital. Some regions, for example Latin banking activity often signals both a weakening of America, report much higher capital-asset ratios than internal controls and an underassessment of credit the OECD average, but provisions against bad loans risk." may be considerably lower. In other countries hidden Management quality. In a world of uncertainty reserves are not disclosed in published balance sheets, and asymmetric information, bankers' judgment on so international comparisons may be misleading. The risks is as important as bank supervisors' judgment on Basle Committee is still working on the proper level the competence and integrity of bank management. of capital adequacy for market and interest rate risks, The bank supervisor's job is not to second-guess the including the difficult area of derivatives and off-bal- banker, since any involvement in credit decisions ance sheet liabilities. (direcdy or indirectly through credit allocation guide- The quality of capital is only as good as the quality lines, for example) would compromise the bank of asset or loss provisioning. Where asset provisioning supervisor's objectivity. Moreover, good behavior 52 BANK RESTRUCTURING cannot be legislated (Lin 1992). Still, bank supervi- Table 3.1 Performance indicators of U.S. national sors should ensure that the managers of public savings banks, 1986-91 have professional training and experience, that sound (percent) procedures and controls are in place, and that succes- Ind4ator Performance ratio sion plans exist for key management. For example, Return on equity 7.7 boards of directors should be fully informed of the Return on assets 0.5 operations and performance of their banks and be Net interest income on assets 3.5 held responsible for management's compliance with Noninterest income on assets I.6 Noninrerest expense on assets 3.5 banking laws and regulations. There should be good Loss provisioning/assets 1.0 internal audit and controls, and timely and accurate Loss reserves/loans 2.5 information should be provided to management to Loss reserves/noncurrent loans 71.4 Net loan loss/loans 1.3 assess the risks of the bank. Loss provisioning/net loan loss 131.7 Recent developments that hinder management's Real estate loans/total loans 35.2 ability to monitor risks include the growing area of Noncurrent loans/total loans 3.5 derivatives, which in nominal terms are almost the Noncurrent real estate loans/real estate loans 3.7 size of the on-balance sheet assets of the global bank- Equity capital/assets 5.9 ing system. A number of banks have made substantial Source: U.S. Comptroller of the Currency 1992. profits from such trading, but bank supervisors and industry experts share a common concern that some should ensure that minimum standards are followed institutions do not fully understand how and why and that banks do not engage in "cosmetic account- such profits were made, and worry that the institu- ing," such as the elaborate rescheduling of nonper- tions did not have the proper management tools and forming loans to make the loans appear to be internal controls to monitor the risks involved. performing. In the early 1980s in Colombia, for Earnings. The potential problems of financial example, banks opened subsidiaries in Panama to institutions often are reflected in their earnings per- bypass domestic lending guidelines so that problem formance. Many developing country banks run into loans would escape supervisory attention (chapter 6). difficulties because of excessive operating expenses. Liquidity Liquidity is generally not a major prob- In Ghana, for example, the operating expenses of lem for sound banks in a reasonably competitive bank- state-owned banks were almost 75 percent of gross ing system, and weak banks often can replenish liquid- interest income in the late 1980s (chapter 8). Banks ity by bidding up interest rates. The recent banking were overstaffed and maintained unprofitable crises suggest that in many cases liquidity crises have branches for reasons of prestige or under govern- their roots in solvency problems. No international ment direction. In OECD countries operating costs standard exists for measuring liquidity, since liquidity are 65 to 70 percent of gross income (including non- depends on an individual bank's financial position, the interest income), with provisions for loan losses set depth of domestic money and capital markets, and the between 15 and 20 percent and pretax profits willingness of the central bank to supply liquidiry to between 12 and 20 percent. OECD banks earn banks. The Basle Committee has developed a frame- about a third of their gross income from noninterest work for liquidity that attempts to measure and man- sources, particularly foreign exchange trading and age net funding requirements, market access, and con- fees. The average return on assets for the top fifty tingency planning (Basle Committee on Banking U.S. banks was 0.95 percent in 1988, nearly four Supervision 1992; Musch 1992). times that of the top fifty Japanese banks and rough- In many developing countries-including many ly twice that of the top fifty European banks. Return post-centrally planned economies-problems within on equity of U.S. banks in 1988 was 18.4 percent, the national payments system have created large compared with 12.5 percent for the Japanese banks "floats" (items being cleared) that tie up huge and 10.3 percent for the European banks (American amounts of bank liquidity. In early 1993 float and Banker 1990). The performance ratios of U.S. clearing items in the Russian banking system totaled national banks are shown in table 3.1. 55 percent of the monetary base. This created a credit Of particular importance to bank earnings is the crunch for enterprises that could only be relieved by income accruals policy on nonperforming loans and central bank credit. Supervisors may be able to bolster the stringency of loan loss provisioning. Supervisors bank liquidity by improving the efficiency of national Resolution and Reform: Supervisory Remedie jfr Problem Banks 53 check clearinghouses or automating the large-value yields, duration (maturity of instrument), and coun- interbank transfer systems. terparty exposure. Bank supervisors must understand Operations. Because banks operate the national the risk management systems that are in place and payments system, they are vulnerable to the systemic ensure that these systems are fully understood at the risks of failures of banks within the payments system treasury, chief executive, and board levels of the bank. or even from technical operating system failures. Many banks today are so dependent on computers Early warning signals and telecommunications that they cannot tolerate more than a day-long shutdown of their computer A common complaint about banks is that their per- equipment. When the computer system of the Bank formance is not transparent to the public, and often of New York failed in 1990, for example, it threat- not even to the bank supervisor. Formal bank super- ened to delay settlement and payment of all transac- vision entails both off-site surveillance and analysis of tions in New York. The New York Federal Reserve bank reporting and performance and regular on-site Bank had to inject additional liquidity into the mar- inspection (World Bank 1993). Irregular inspection ket until the technical failure was resolved. Bank may result in a failure to detect early warning signals. supervisors must ensure both that technical backup During 1985-86, for instance, when many banks in facilities are in place for individual banks and that the U.S. state of Texas began to fail as a result of systemwide operations for the payments system are overlending, the number of inspections was halved protected from technical failures. due to budgetary and staff constraints (chapter 4). Banking has become highly dependent on tech- Deregulation, globalization, and financial innova- nology-not only for global operations, but also for tion have allowed banks to evolve into complex orga- information processing, funding arrangements, and nizations thar span the world geographically and market arbitrage. Accordingly, bank management across different product lines. Banks have become must pay close attention to the robustness of operat- involved in insurance, securities firms, and fund ing systems and management information systems, management, including commodities and futures which are critical to risk management. A system fail- trading. While such diversification of assets could ure that coincides with large market volatility could reduce risks, in a number of cases-notably the U.K. destroy a bank's capacity to avoid losses in the interim. Bank of Credit and Commerce International (BCCI) Treasury management. Partly because of the need case-the inadequacy of coordinated supervision has to "save" on capital, and also because of financial perpetuated fraud. The ability of bank supervisors to innovations and market deepening, banks have supervise financial conglomerates across legal and increasingly relied on non-credit based income to geographical boundaries will be a major challenge in boost earnings. For example, income from treasury a global banking environment (Corrigan 1992). operations and noninterest income accounts for as Bank supervisors will have to better understand the much as half of gross bank income in Switzerland. financial commitments of the banking core to the The risks in treasury operations are mainly interest nonbanking (and probably unregulated) wings of the rate and exchange rate risks. But with the increasing financial conglomerates. Some of these new activities role of derivatives (options and futures)-which could be driven by "regulatory arbitrage" to avoid involve off-balance sheet exposures-bank supervi- detection of overleveraging or fraud. Thus what sors are becoming increasingly concerned with expo- appears to be a minor new item about the involve- sure to third parties (reputation and counterparty ment of a bank in a distant offshore cornpany could risks), which may unravel when one party in the chain turn out to be a major source of losses. of market transactions fails to honor its contracts. Early warning systems can be improved through Some money-center banks have transformed their the installation of commercial software for off-site operations from a lending function to a market trans- surveillance and bank analysis. Still, such programs action function, generating revenue from money are no substitute for alert analysis of bank indicators market instruments, bonds, derivatives, and foreign from a variety of sources (box 3.2). "Soft" indica- exchange trading rather than from credit generation. tors-such as market talk or anonymous letters In doing so, these banks have transformed their inter- reporting fraud and mismanagement-are particular- nal risk management procedures, concentrating on ly indicative that a bank may be at risk. The BCCI the return on capital that factors in benchmark case is a classic example of the gaps in international 54 BANK RESTRUcruRING speculative risks to cover up such losses. Where fraud is suspected, management should be changed, since * Weak or uninformed board of directors and a dom-- untrustworthy managers have incentives to escalate the inant chief executive. fraud and run before supervisory action is taken. Many . Poor lending practices, bank supervisors hesitate to change management * Late submission of bank returns and reports. * Problems in affiliates, related companies, or large because current managers have the best nformation customers. on the state of the bank, whereas new managers will * Rapid staff turnover and changes in top management. need time to learn. Moreover, there is often a punitive * Change in auditors. desire to make the existing management responsible * Rapid growth in assets. for collecting the bad debts that they made. * Liquidity problems. Bank supervisors tend to forget that, where fraud * Deposit rates higher than market. or corruption in credit decisions is present (such as . Market talk. taking commissions on loans or connected lending), * Public complaints. * Ostentatious spending. managers have every incentive to hide or escalate loss- * Use of political influence. es. This is partly because jail sentences are not com- mensurate with the scale of bank theft. Managers go to jail whether they have stolen $10,000 or $100 million, supervision that delay detection of extensive fraud and the larger the losses, the greater the need for public and mismanagement over a number of years. rescue-thus implicating the supervisors in some form Although off-site surveillance is important and of coverup. The Malaysian central bank governor suc- the use of external auditors can strengthen supervi- cinctly expressed the Malaysian experience with prob- sion of financial intermediaries, there is no substitute lem bank management: "Never let monkeys look after for on-site inspection of banks-if only to provide bananas" (Bank Negara Malaysia 1987). direct evidence to bank supervisors of problems on Double your estimate of losses. Bank accounting is the ground. Moreover, external auditors may face not transparent to the public or to the bank supervi- conflicts of interest since they are paid by banks, and sor, mainly because losses can stem from a variety of may not be trained to detect fraud. Bank examiners causes that may not be uncovered until stern supervi- can assess the validity of market talk about cosmetic sory action has been taken. Aristobulo de Juan, a bank accounting, connected lending, or speculative losses. supervisor who helped resolve Spanish banking prob- Examiners also can gather evidence that may be used lems in the early 1980s, asserts that final losses are in later legal actions. In a number of instances the lag always double the previous estimates (de Juan 1991). between reports of potential fraud by auditors and His experience, borne out in a number of other coun- subsequent action to obtain evidence has resulted in tries, is that external auditors' estimates of losses are the destruction of such evidence. almost always double those of the bank management, bank inspectors' estimates are at least double those of Approaches to damage control the auditors, and final losses on liquidation are double those estimated by the bank inspectors. The purpose of damage control in supervisory action Delayed loss recognition is partly the result of is to minimize bank losses once they are detected. In wishful thinking-the hope that things will recover in almost all cases inspection teams should be sent in as the next quarter or accounting period. The underre- soon as possible to determine the extent of the dam- porting of losses in the previous accounting period by age. Where laws prevent direct bank inspection, bank management or auditors undercuts their credi- external auditors should be commissioned to provide bility, and hence there are incentives not to report an independent view of the losses. Four rules should large variations. In addition, the evaluation of nonper- guide damage control efforts. forming loans is a judgment, and actual writeoffs Strengthen or change management. Because of the could be substantially higher. Experience in Finland high leverage in the banking business, shareholders and Norway suggests that writeoffs of nonperforming face lower risks as losses increase, and consequently the loans ultimately could be as high as three-quarters of greatest risks are borne by depositors or by the deposit book value. As conditions deteriorate, and the net guarantee fund. Thus shareholders and management worth worsens, the bargaining position of the have incentives to disguise losses and to take problem bank weakens and sale of assets under time Re.olution and Refrm: Supervisory Remee.iefr Problem Bank. 55 pressure forces further losses. Recent U.S. experience insurance schemes or implicit central bank support of suggests that even under conservative accounting, liq- the banking system are present, commercial bank loss- uidation losses averaged as much as 17 percent of es in excess of bank capital are quasi-fiscal deficits. book value for failed banks and thrifts. Consequently, action taken by bank supervisors must BCCI is an example of the escalation of losses. be carefully explained both to the public and to poli- Initial press reports suggested that the bank's capital cymakers. Bankers who are likely to lose their jobs was totally eroded. Later estimates escalated to $5 bil- and shareholders who are likely to lose the value of lion of total assets of $20 billion, and recent estimates their shares are quick to blame inadequate or misguid- suggest that more than $10 billion may have been ed bank supervision. Thus it is important for the pub- lost. Another example comes from Norway, where lic to understand that the primary objective of super- problems in the banking system were detected in visory action is to protect the safety of public deposits 1988 but losses continued to escalate until 1992. The and that losses often are caused by poor bank manage- banks only began to recover in 1993. Asset deteriora- ment and inadequate action by shareholders- tion occurs partly as a result of management deterio- particularly where shareholders have been unwilling to ration and breakdown of controls. In cases of fraud augment capital to save the problem bank. involving banker-borrower collusion, borrowers may Damage control inevitably means that rhe bank abscond with or deny the validity of collateral. When supervisor has to intervene in problem financial insti- banks get into trouble, borrowers have incentives to tutions. The degree of intervention depends on the seek legal loopholes to delay or renege on payment seriousness of the situation, and also on the legal because they see an end to their debt-servicing com- powers available to the supervisor. A typical sequence mitment. Thus, where collateral documentation has of events is listed in box 3.3. been weak, the bank (or its liquidator) can collect Depending on the legal powers available to bank only after protracted court proceedings-which supervisors, some countries use moral suasion to get result in further losses. bank management to take corrective action. In the Act sooner, not later. Experience suggests that the United States, where legal powers are clearly defined, a sooner supervisory action is taken, the lower will be memorandum of understanding about corrective the costs incurred. The U.S. thrifts crisis illustrates action is signed by the bank supervisor and the prob- the enormous costs of delayed action, in this case the lem bank. Failure to comply with the conditions of result of reluctance to absorb the losses in the loosely the memorandum can trigger a series of supervisory regulated environment of the early 1980s. The eco- actions, such as cease and desist provisions. These may nomic insolvency of the thrifts might have been only require the problem bank to cease certain loss-making about $100 billion if all assets had been marked to activities, such as foreign exchange trading or buying market in 1982 (U.S. Treasury 1991). But after more junk bonds. In other countries central banks can pro- than 1,400 thrifts disappeared from the market as a hibit further branching or the granting of new loans. result of liquidations and mergers in the 1980s, gov- Access to the money market can be denied, and cen- ernment efforts to bail out the industry cost at least tral bank discounting facilities can be withdrawn. $315 billion (U.S. GAO 1989). These actions control the problem bank's access to The larger losses due to delayed action are not funds and limit the damage that the bank can inflict attributable solely to the time cost of initial losses on other financial institutions and depositors. (the stock problem). In many cases-particularly in post-centrally planned economies-flow problems Box 3.3 Bank supervisors' damage control options are more to blame, arising from the structural diffi- Guide management action and reporting. culties of transforming loss-making state enterprises. * Draft memorandum of understanding with board As long as banks continue to lend to these loss-mak- of supervisors. ers, the deterioration in net worth will compound. * Cease and desist action. This failing can only be resolved through determined * Restrict central bank lines and trading limits. damage control and radical restructuring. Since 1990 * Change management and board of supervisors. Hungarian authorities have engaged in four rounds * Appoint adviser or conservator. of recapitalizing loss-making state banks. * Assume control. Obtain support for supervisory action. Problem * Increase capital. banks hold public funds. Where explicit deposit * Initiate liquidation. 56 BANK RESTRUcrURING If the provisions of the memorandum of under- restrict dividend payments, limit asset growth, face standing or any legal provisions are violated, U.S. curbs on activities, and file a plan to increase capital regulators may remove or suspend specific persons (table 3.2). from participating in the affairs of the bank. Alternatively, a penalty can be levied on banks or Rehabilitate or close? bank managers for violations of regulations or orders. A U.S. bank manager found guilty of certain viola- One of a bank supervisor's most difficult tasks is cal- tions can be disbarred permanently from employ- culating the costs and benefits of bank rehabilitation ment in any bank. In other countries this punish- or closure. In the mid-1980s governments rescued ment is enforced largely through moral suasion, large problem banks because of the "too big to fail" where licenses are withheld if unfit persons are found principle, thus extending deposit insurance coverage managing a deposit-taking institution. to almost 100 percent of the deposit base. Now there is consensus that only banks that are critical to the Criteriafor intervention payments system, and whose failure could trigger sys- temic failure, should be intervened in and perhaps Once a bank supervisor intervenes in a financial insti- not allowed to fail. This approach suggests that bank tution, whether by appointing an adviser (in most supervision should focus on a core of high-quality countries) or conservator (in the United States) or by "failproof" banks that are directly involved in the assuming control directly, the public considers the high-value, time-critical payments transfer and clear- government responsible for the deposit liabilities. ing system. Smaller banks and nonbank financial Thus intervention is not a step to be taken lightly. In intermediaries whose failure would not jeopardize the the past the decision to close or intervene in a bank payments system should be allowed to fail, to rid the has been left largely to the bank supervisor. The 1991 market of inefficient institutions. Indeed, the current U.S. Federal Deposit Insurance Corporation view is that supervisors should not aim to prevent Improvement Act, however, put in place a series of bank failures (Quinn 1991). measures that would compel regulatory action when There is, however, a body of opinion that func- bank capital falls below certain levels. tioning institutions may have a higher market value The act mandates prompt corrective action when (and thus a lower cost of bailout) than immediate a bank becomes critically undercapitalized, defined as closure would produce. Different countries use differ- the point at which its capital and reserves fall below 2 ent institutional mechanisms to make this difficult percent of risk assets. This threshold is both a general decision. In the United States the Federal Deposit rule of thumb about levels of "near insolvency" and an Insurance Corporation (FDIC) decides whether to empirically verifiable danger threshold. According to intervene. The FDIC cannot choose a method that is the 2 percent criteria, the margin of error in the mea- more expensive than liquidating the bank and paying surement of accounting solvency (according to gener- off the depositors. A purchase and assumption route ally accepted accounting principles) should not exceed is taken if it will save the FDIC substantial time and 2.5 percent if risk assets account for 80 percent of administrative costs over liquidation and deposit total assets. In other words, undermeasurement of repayment and if the sale price of a failed bank will losses of 2.5 percent of risk assets would wipe out the be higher than its liquidation value. capital of a bank with only a 2 percent capital-asset In Belgium an Intervention Fund Committee ratio. Minor falls in the return on assets of a bank due comprising bank representatives has to satisfy itself to nonperforming loans could quickly decapitalize a that offering support would be less costly than reim- bank according to market-value accounting. bursing depositors after a winding up. In Malaysia an Under the 1991 act significantly undercapitalized advisory panel to the central bank-comprising banks-defined as those having less than 3 percent members of the central bank, ministry of finance, core capital and 6 percent risk-weighted capital- attorney general, a representative of the bankers asso- would have to restrict their pay and bonuses, and ciation, and another private sector member-advises bank regulators could force the election of new direc- whether the central bank can invoke certain powers tors and the sale of units of the problem bank. Banks to rescue a failing institution. with core capital of less than 4 percent and risk- The decision to rehabilitate or close depends on weighted capital of less than 8 percent would have to the institutional framework, the legal powers and Resolution and Reforr: Supervisoty Remediesfor Problem Banks 57 Table 3.2 U.S. regulatory action under the 1991 Federal Deposit Insurance Corporation Improvement Act Risk-weighted Core Capital status capital, capital' Mandatory and discretionary actions Well capitalized 10 percent 5 percent * Cannot make any capital distribution or pay a management fee to a controlling person that would leave the institution undercapitalized. Adequately capitalized 8 percent 4 percent * Same as above. Undercapitalized Less than 8 percent l,ess than 4 percent * Subject to increased monitoring. * Must submit a capital restoration plan. * Asset growth is restricted. * Approval required for acquisitions, branching, and new lines of business. Significandy Less than 6 percent Less than 3 percent * Must increase capital. undercapitalized * Interest rates paid on deposits restricted to the prevailing rates in the region. * Asset growth is restricted; reduction of total assets may be required. * May be required to elect a new board of directors. Critically 4 percent Less than 2 percent * Must be placed in conservatorship or receivership within undercapitalized ninety days. * Prohibited from the following without FDIC approval: - Entering into any material transaction other than in the usual course of business. - Extending credit for any highly leveraged transaction. - Making any material change in its.accounting methods. - Paying excessive compensation or bonuses. a. Ratio of capital to assets. Source: Nakajima and Taguchi 1993. resources available to intervene, and the ultimate cost remaining assets, with those interested performing a to the government. The steps involved, however, "due diligence" examination of the failed institution broadly include the following: (box 3.4). i Determining the size of bank losses (the degree of A failed bank has four basic pieces: a deposit fran- insolvency). Conservative estimates should be chise; market assets-such as bonds and cash-that made according to the "double losses" rule can be disposed of easily; banking assets-such as described above. loans-that cannot be valued easily; and problem i Calculating the net deficit, estimating the time assets-such as nonperforming loans and real frame of recovery and assuming that new capital estate-that are worth less than book value and have is available. high carrying costs. The franchise value of banks is * Determining loss allocation-to shareholders, the difficult to assess. But the scarcity of bank licenses in a need for (and costs of) staff retrenchment, and number of countries and the existence of large branch the burden to be borne by the restructuring networks often generate significant franchise value. In agency (central bank, deposit insurance agency, or Spain, for example, foreign banks were willing to ministry of finance). acquire problem banks from the Deposit Guarantee * Estimating the fiscal impact of depositor bailout, Fund in order to gain entry into the Spanish market. assuming that government will engage in either a The Bank of Thailand was able to recapitalize prob- carve-out or flow solution. lem banks by selling off shares to well-capitalized * Is there a franchise value? If so, this can be added nonfinancial institutions that wanted to acquire a to the capital of the insolvent bank. strategic stake in banks. In Norway the Government * Is there a market for bank shares with other banks Bank Insurance Fund effectively "nationalized" the or new shareholders? banks by injecting capital into problem banks. Bank Recent U.S. experience indicates that the resolu- privatization had to await the restoration of the banks tion process starts with a diagnosis, typically involv- to health, and by early 1993 a number of them had ing an information package and asset valuation begun to generate small profits. review. After this, an information meeting is arranged Experience from a number of countries suggests with potential acquirers of the failed bank or the that the net insolvency of problem banks should be 58 BANK RESTRUCTURING Box 3.4 Lessons from the U.S. experience The U.S. experience is a useful example of a compre- banks will solve problems and preserve value. But the hensive banking and bankruptcy law covering a large public supervisory agency should design the provisions pool of competing banks that, with the right incentives, and structure of the assistance agreement, with built-in can take over the assets of failed institutions. Such a arbitration and other dispute resolution procedures. framework is missing in many developing countries. * The ability to design a structure and to persuade pri- The following guidelines should be kept in mind to vate banks to take part depends on the quality of reduce the losses arising from the resolution process: available information. -Assets are the problem. Someone has to manage . Negotiations should be open-competition keeps them, someone has to bear the risks associated with acquirers honest. them, and some entity has to carry them on its bal- * The tax consequences of any structure-and what ance sheet. These functions can be separated or com- they mean in terms of net cost to the government- bined in a number of ways, and no single structure is should be understood, but they should not dominate inherently superior. the deal. The private sector should be a partner in any resolu- tion or restructuring. With the right incentives, private Source: Hove 1994. determined in a transparent and legally enforceable high real interest rates create uncertainty and large manner. In the Philippines shareholders of failed wealth losses in the economy, which worsen financial banks taken over by the Central Bank have success- distress. Open capital accounts permit volatile capital fully disputed the valuation of the net capital short- flows and make monetary and macroeconomic man- fall, impeding the efforts of bank supervisors working agement even more difficult. on the failing banks. Once the net losses of the prob- The massive banking problems of the 1980s drive lem bank have been established, the appropriate tech- home the message that inappropriate macroeconomic nique of bank restructuring should be chosen based and regulatory policies contribute to bank fragility. on the initial conditions, an assessment of the flow of As many banking crises showed, the competitive future losses, and the need to sequence reforms. pressures unleashed by financial liberalization, while increasing efficiency, also carry risks, becouse banks The Longer View: Rebuilding Bank Profitability and other financial institutions alter their behavior to ward off institutional downsizing (IMF 1993). Asset Prevention is obviously better than a cure. Financial price bubbles-and their subsequent deflation-had distress is not a historical inevitability-it can be their origins in excessively lax monetary and supervi- avoided through a combination of supervision, sound sory policies during the early phase of the boom, and macroeconomic policies, and effective financial insti- in excessive restrictiveness thereafter. tution-building. As the financial liberalizations of the As discussed in chapter 2, the banking system can 1980s show, liberalization in the absence of a consis- generate losses that are ultimately quasi-fiscal deficits tent and adequate macroeconomic policy framework because of the implicit or explicit state guarantee on may be unsuccessful or even counterproductive (Gelb the deposit base. In macroeconomic terms the sus- and Honohan 1989). Drastic surgery-carving out tainability of such losses is based on the sustainability bad loans or rehabilitating banks-will not be sus- of the total public deficit and debt. In highly indebt- tainable if the underlying causes of bank distress are ed countries that had external financing cut off, not removed. A number of conditions must be met if bank restructuring is to he successful (box 3.5). Box 3.5 Conditions for rebuilding bank profitability Stable macroeconomic environment * Stable macroeconomic environment * Relationship with enterprise restructuring A stable macroeconomic and financial environment is * Removal of tax and regulatory distortions crucial to financial reform. Countries with high infla- * Legal and accounting reform tion often witness severe disintermediation from the * Market and institutional deepening banking system and distorted allocation of resources. * Getting the incentives right * Professional competence and ethics High inflation, the accompanying devaluation, and Rtsolution and Reform: Supervisory RemediesforIProblem Banks 59 reliance on internal debt financing resulted in money On the other hand, rapid trade liberalization in creation and high inflation. Distressed borrowing- previously protected markets such as Ghana and first by enterprises, then by banks, and ultimately by Poland removed the high rents earned by inefficient the state-forced up real interest rates beyond sus- state-owned enterprises, creating large losses as firms tainable levels. Where such real rates paid by the state struggled to adjust to the new competitive environ- exceed the real growth rate and the growth rate of ment. State banks that had lent heavily to these noninterest tax revenue, the total public debt grows enterprises were unable to avoid large losses. exponentially and becomes unsustainable, leading to Combining bank restructuring with enterprise hyperinflation or debt default (see box 2.6; Fischer restructuring will be particularly important in the and Easterly 1990). post-centrally planned economies because the enter- In the end, a number of countries achieved prises will require significant resources during the macroeconomic stability (together with restoration of transition to a market-based environment. The stability in the banking sector) only when the govern- adjustment costs during this period, which could ment was able to generate a primary surplus that severely decapitalize the banks in these countries, allowed the total public debt ratio (including the may be unavoidable (Sheng 1992). stock of recognized or unrecognized bank losses) to stabilize or decline over time. The anchor of macro- Removal of tax and regulatory distortions economic stability is fiscal discipline, where the need for public borrowing does not crowd out funding for Tax and regulatory distortions in the financial sector private investment or overextend the nation's external worsen resource allocation. In Turkey, for example, debt capacity. Sound domestic and external debt 26 percent of the lending rate in 1983 was attribut- management strengthens monetary control, deepens able to a direct tax on interest or reserve costs on the capital market, and reduces inflationary pressures. banks, thus pushing the cost of intermediation higher (Hanson and Rocha 1986). In Ghana the tax rate on Relationship with enterprise restructuring banks was higher than that for nonfinancial enter- prises (chapter 8). A financial system cannot remain profitable over the Discriminatory tax rates between banks and non- long run at the expense of the real sector. In many bank financial intermediaries also have created unfair countries-and especially in the post-centrally competition. For example, development banks, planned economies-bank distress is a direct conse- finance companies, and money market funds may quence of enterprise insolvency. By failing to address not have to maintain statutory reserves with the cen- bank distress, governments worsened resource alloca- tral bank, and therefore their cost of funds may be tion and macroeconomic stability, as banks continue significantly lower than that of commercial banks. to pump resources into loss-making enterprises to Such sectors of the market enjoy significant competi- keep themselves afloat. Consequently, bank restruc- tive advantage, and yer the commercial banks would turing must be accompanied by a credible package of bear the brunt of the credit risks if they were to fund real sector reforms that address the causes of enter- these nonbank financial institutions excessively. The prise and borrower losses. recent exposure of Japanese banks to stocks and prop- For example, countries that suffered from erty was relatively limited, but if the exposure to non- the "Dutch disease" of overvaluation of the bank financial institutions is included, this exposure exchange rate, such as Chile (1976-80), Colombia was considerably greater. (1980-84), and Malaysia (1980-84), saw the prof- In essence, tax and regulatory reforms entail itability of domestic industries erode as imports removing financial repression and desegmenting the grew, experienced a decline in the profitability of financial system. This requires the removal of exces- exports, and saw investment turn toward nontrad- sive regulation of interest rate and credit allocation ables such as real estate and share speculation. The policies, including dismantling of interest rate con- correction of the exchange rates in these countries trols, freeing entry into investment banking and helped restore enterprise profitability, since domes- capital market activities, reducing quasi-fiscal bur- tic agriculture and industries became able to com- dens such as excessive reserve requirements, liberal- pete once again with imports and were able to cre- izing exchange rate controls, and removing credit ate new markets for exports. quotas, targets, and ceilings. Taxation on domestic 60 BANK RESTRUCTUIUNG intermediation of funds should be harmonized to payments system (such as check clearing and elec- create a level playing field for competition. tronic fund transfer mechanisms). Improvements in money and capital markets Legal and accounting reform strengthen macroeconomic management because the central bank is able to use more efficient indirect Legal and accounting reforms should ensure that instruments of monetary management and the bud- financial transactions and financial institutions are as get can be financed in a more transparent manner. transparent as possible, with laws and court proce- Better money and foreign exchange markets also dures that enforce financial discipline. Domestic improve the market skills of banks, particularly their accounting standards should be brought in line with short-term liquidity management. With deeper capi- international accounting standards so that domestic tal markets, the budget gains access to longer-term bank efficiency can be compared with international funds at market rates, reducing the need to repress standards. Accounting disclosure and audit standards the banking system for funds. Development of a gov- should facilitate public and supervisory understand- ernment bond market creates a benchmark yield ing of the performance trends of financial institutions curve for domestic interest rate determination while so that the market can appraise the true value of these providing a stable and noninflationary source of institutions' net worth. financing for the budget (Emery 1993). Prudential standards must be simple, compre- Development of the securities market provides hensive, and educational (Butsch 1992). They long-term equity financing for the corporate and should not be a substitute for internal controls of banking sectors, reducing their gearing and therefore banking institutions, but should educate financial strengthening their capacity to absorb higher risks institutions and the public to be aware of their risks. and external shocks. As such markets develop, spe- Moreover, legal impediments to debt recovery, cialized funds such as venture capital, mutual, and including bottlenecks in court proceedings or property funds can be established to mobilize funds defects in the law, should be remedied so that banks for higher-risk investments. can impose financial discipline on borrowers. In Improving the timeliness and certainty of pay- Hungary courts were unprepared-both technically ments transmission, clearing, and settlement gener- and in terms of capacity-to deal with the growing ally reduces the liquidity pressures in an economy number of bankruptcies and liquidations required (Summers 1992). It also imposes financial and under the new banking and commercial law of accounting discipline on participants in the finan- 1992. As of September 1992, the debt of companies cial system and allows the bank regulator to detect under liquidation or bankruptcy proceedings by quickly those institutions that have difficulties financial institutions amounted to 110 billion meeting their payments or that are operationally forints, or nearly 90 percent of the classified bad inefficient. The inefficient payments and settle- debt reported by banks (National Bank of Hungary ment system that was used in the post-centrally 1992). In many countries where debt recovery was planned econornies prevented central banks from delayed by cumbersome court procedures, special engaging in more active monetary measures to sta- tribunals were established to expedite the process. In bilize inflation. In active, global, and twenty-four- 1993 Pakistan took the unusual step of publishing hour trading of financial markets, central banks the list of major defaulters to state-owned banks to have become more aware of the systemic risks aris- induce them to repay. ing from the possible disruption or failure of money and capital market payment, clearing, and Market and institutional deepening settlement systems. Another area of financial deepening is the devel- Market and institutional deepening calls for basic opment of nonbank financial intermediaries, which improvements in the financial infrastructure. This help to spread risk and transform maturities. involves two major areas: the creation of active Encouraging the development of contractual savings money and capital markets (to allow markets to institutions, such as life insurance companies, pen- function smoothly and price the market value of sion and provident funds, and social security financial assets at a competitive level) and the schemes, would promote long-term finance that enhancement of the operating efficiency of the strengthens the capital market. Resolution and Reform: Superviory Remediesfir roblem Banks 61 Getting the incentives right sional competence and ethics of the people who operate and supervise the system. Even the best bank In recent years policymakers have become aware of supervisor cannot prevent the failure of a system how distorted incentive structures can lead to massive where bankers cannot tell a good borrower from a waste and inefficiency. A policy framework that bad borrower, as in the highly distorted price envi- attempts to alleviate bank distress must take into ronments of the post-centrally planned economies. account the incentives of all parties concerned-bank Bad bankers cannot become good bankers if they do managers, bank equity and liability holders, and regu- not receive training in credit skills and the account- lators. Such a framework should strive to preserve ing and legal framework does not allow them to financial discipline and induce cooperative solutions enforce debt recovery. And good supervisors can (Glaessner and Mas 1991). In other words, the incen- become bad supervisors if they supplement their tives structure should induce bank management and income by "taxing" or borrowing from bankers. shareholders to take preventive action before bank dis- Strong institution-building therefore depends on tress occurs, and also force timely action on the part of devoting significant resources to building skills and regulators to act before distress becomes too costly. strengthening controls within individual banks. The In principle this logic is irrefutable. In practice, bank regulator or central bank can facilitate this by because laws are enacted periodically, the incentive encouraging private involvement in the development structures built into laws and institutions may quickly of selected institutions, such as bankers associations become obsolete because of rapid market changes. and training centers, credit rating agencies, and credit Therefore the design of incentive and institutional information sharing bureaus (Gibeault 1993). As structures should take into consideration cyclical fac- international supervisors recognize, "the self- tors and random shocks. For example, deposit insur- discipline of banking institutions must be reinforced" ance may be useful in preventing bank runs, but it also (Butsch 1992). has proven quite costly in terms of moral hazard costs. A central dilemma of financial supervision in a Thus a fundamental principle in the design of a global financial market with twenty-four-hour trad- safe and sound financial system is the avoidance of ing is that the regulatory framework is a patchwork asymmetry between risk and rewards. Bankers must of national policies. Global financial conglomerates always have their own capital at risk so that they will are evolving that can engage in significant interna- not engage in speculative behavior at large social costs. tional regulatory and tax arbitrage, moving funds to Bank supervision should not supplant bank manage- take advantage of information asymmetry and mar- ment judgment, and overregulation runs the twin risk ket opportunities. Most of these movements have of bank management and depositor complacency and legitimate market reasons, but predatory measures- risk aversion, so that high-yielding projects may not be such as acting in concert-are increasingly being funded, leading to lower long-term growth. Moreover, taken in small markets by big players that are outside a high level of bank supervision is equated with a full the jurisdiction of any single regulatory authority. or extensive government guarantee of the deposit base. Much has been achieved through international Another lesson from recent experience is that cooperation in banking supervision. Difficulties nat- bank supervision is useless if there is insufficient urally arise in cooperation across functional lines, enforcement. Policy credibility influences public such as cooperation with securities and insurance reg- behavior. Despite the extensive failure and high losses ulation. Nevertheless, the long-term health and sta- of many African banks due to mismanagement or bility of the global financial system depends on the fraud, few bankers have been charged or jailed. Thus integrity and fairness of the market operators-the bank supervision has bite but no teeth. Incentives intermediaries themselves. alone will not create a sound financial system, but strong disincentives should be built in so that rewards Bank and Enterprise Restructuring in the are evenly matched with the risks of punishment. Post-Centrally Planned Economies Professional competenceand ethics World Bank operations to assist the post-centrally planned economies with bank restructuring are cur- The soundness and efficiency of the banking system rently under way in a number of countries, including ultimately is only as good-or bad-as the profes- Bulgaria, Hungary, Poland, Romania, Slovenia, and 62 BANK RESTRUCrURING the Baltic states. As discussed in chapter 1, the inter- them. Not all price reforms are negative. For example, mingling of roles between the budget, state-owned trade and exchange rate reforms were principal ele- banks, and state-owned enterprises suggests that the ments of the reform package that allowed Polish, move toward a market-oriented financial system will Czech, and Hungarian firms to adjust to export mar- require a careful sequencing of the disengagement of kets, thus reducing the losses that would have occurred public roles from private functions. The crucial ele- had access to such markets been denied. ments of bank and enterprise restructuring are listed Much of the debate on the correct sequencing of in box 3.6. bank and enterprise restructuring in these countries is First, state-owned enterprises have to shed their colored by the controversy over privatization of banks social welfare functions (housing, child care, and and enterprises. While privatization is not an end in other benefits for workers and their households) and itself, it is a powerful means of improving efficiency in concentrate on efficient production. the production of goods and services and of strength- Second, state banks have to disengage themselves ening social and democratic institutions (Snoy 1992). from the quasi-fiscal function of providing funding- There is no question that privatization helps in two both working capital and investment-to the budget key areas: adjusting the incentives structure (moving and to state-owned enterprises, especially loss-making toward more efficient production) and improving the enterprises. As long as banks are burdened with such financial balance (capital adequacy) for economic quasi-fiscal roles, they cannot intermediate savings agents. Still, whether mass privatization should be and investment efficiently, and controlled interest pursued all at once or through intermediate steps, rates do not reveal the true market price of credit and such as corporatization, is an empirical question that savings. can only be answered by each country's legal, institu- Third, the budget must not rely on the taxation tional, political, and historical precedents. of banks and state-owned enterprises to finance itself While the process of bank restructuring will vary (McKinnon 1991). Neither should the budget con- from country to country depending on initial condi- tinue to protect inefficient state-owned enterprises by tions and financial structure, several principles and providing massive subsidies-including monetary objectives should guide the design of bank restructur- creation-to maintain economic growth. The rapid ing programs: rise in inflation resulting from monetary creation to * The financial condition of the banks needs to be finance loss-making enterprises or rescue failing improved so that they can efficiently intermediate banks has been the bane of many governments in the funds. Bank losses often reveal themselves in large post-centrally planned economies. spreads, leading to both negative real deposit rates Fourth, the time and pace of price reform is impor- and high real lending rates. Banks should be tant. Rapid price reforms immediately cause huge loss- "cleaned" so that spreads narrow to revive saving es and reveal inefficiencies in state-owned enterprises, and investment. The cleaning-up process would resulting in massive losses to the banks that finance facilitate the privatization of banks, including meeting minimum capital adequacy requirements. Box 3.6 Preparing for reform Since bank losses generally stem from lending to loss-making enterprises, the debt burden of enter- The following stcps must guide bank and enterprise prises must be relieved to improve their recovery restructuring efforts: and income generation capacity and facilitate pri- * Determine the size of losses-the stocks and flows. vatization of state holdings and general economic * Choose a centralized or decentralized debt restruc- turing solution. restructuring. Related to the problem of bank * Reduce flow losses resulting from continued expo- restructuring is the issue of "pass through"-how sure to loss-making enterprises and thereby improve debt relief for the banks can be passed through to intermediation. the enterprises to facilitate their restructuring and * Determine whether the writeoff of enterprise debts recovery. Restructuring also must reduce the will be done by banks or by the state. backlog of enterprises in liquidation or bankrupt- * Determine whether to restructure banks before pri- cy, since these tie up scarce economic resources. vatization and restructuring of enterprises. The burden of past losses of state enterprises and * Determine the appropriate role of banks in enter- prise restructuring. banks should be shared by the state, nonstate bank shareholders, borrowers, and depositors. Resolution and Refrrm: Supervisoty Remediesfor Problem Banks 63 The rationale for using state funds to relieve The inability of supervisory authorities to banks of their stock of nonperforming loans is to impose damage control in many post-centrally improve the allocation of scarce resources. The planned economies stems from the limited resources higher the level of state assistance, the greater the and capacity of these authorities. Bank supervisors burden of losses that is shifted to taxpayers, have been preoccupied with the need to draft bank- instead of to future borrowers or depositors or ing legislation, develop off-site surveillance capacity, current shareholders. The "carving out" of bad and produce basic banking regulations and stan- debt through swaps of bad debts for long-term dards. Examination capacity is limited to a few bonds spreads the costs over several years. inexperienced inspectors whose ability to regulate * It must be determined how the losses can be the rapidly growing banking sector is at best rudi- absorbed by the budget without threatening fiscal mentary. In many cases damage control measures discipline and macroeconomic stability. In many are limited to the gradual withdrawal of central post-centrally planned economies it is unclear bank funding to loss-making state banks and to how the state can finance such losses given these canceling licenses when fraud or wrongdoing is sus- countries' tenuous fiscal situation. This is further pected. In the absence of a comprehensive bank complicated by governments' heavy reliance on supervision framework, many supervisors face the tax revenues from state-owned banks. daunting task of resolving large loss-making banks, * The choice of restructuring options may depend which are structurally linked with loss-making not only on the fiscal cost, but also on time and enterprises. In many cases these supervisors have administrative costs. For example, pushing neither the financial nor the human resources to numerous failed debtors into bankruptcy pro- deal with the problems. ceedings without established court procedures and trained personnel will create bottlenecks that Absorbing quasi-fiscal losses allow asset values to deteriorate as banks await court decisions. Thus solutions should foster Since the state generally provides an explicit or competition and transparency and avoid overly implicit guarantee on deposits in the predominantly bureaucratic measures. state-owned banking system, losses in these banks are • Finally, bank restructuring must address incen- quasi-fiscal deficits. The ability of the budget to bear tives in such a way as to prevent excessive debt such losses can be determined using the Fischer- leveraging, avoid weakening credit discipline, Easterly equation for the change in the government allow market forces to operate on a level playing debt to GNP ratio (d) (see box 2.6): field, and improve competition, resource alloca- tion, and risk management. Any scheme that pre- Change in d= (primary deficit/GNP) serves monopolies or oligopolies (state or private) - (seigniorage/GNP) + d(real interest rate will only perpetuate existing distortions and - GNP growth rate). increase future costs of resolution. The restructur- ing scheme should improve incentives that reward The primary deficit is worsened by the quasi- competition and efficiency and punish agent fiscal losses of the banking system, which can be behavior that raises social costs. financed through seigniorage. Consider a post- The diagnosis process requires clear identification centrally planned economy at the early stage of of the causes of the losses in the banking system, reform: including the proper measurement of losses-using international accounting standards-in both banks Primary deficit/GNP and enterprises. Recent experience in Eastern Europe before banking losses 2.0 percent of GNP suggests that portfolio audits of banks do not reveal Credit to state-owned the true size of losses because the fundamental prob- enterprises 70.0 percent of GNP lems stem from enterprise losses, which are difficult Average nonperforming 25.0 percent of total to measure in the absence of complete price reform loans credit and enterprise restructuring, particularly the lack of Equivalent "carve out" clear ownership of property rights. Thus losses will bond to stabilize banking continue to flow until these issues are resolved. system (70 times 0.25) 17.5 percent of GNP 64 BANK RESTRUCTURING Quasi-fiscal losses of banking system: The budget will be reformed by introducing a value- Carve-out bond at x added tax and other direct forms of taxation to interest rate on government reduce the overreliance on enterprise taxes. debt, say 10 percent a year 1.8 percent of GNP Seigniorage 2.5 percent of GNP Diagnosis and prognosis Real interest rate 0.5 percent a year Growth rate -1.0 percent of GNP The first point to recognize is that banks are dealing Inflation 9.5 percent a year not only with historical (stock) losses, but also with flow losses arising from continued loans to loss-mak- Using the equation above, ing enterprises. In many post-centrally planned economies nonperforming losses occur partly because Change in d= [2.0 + 1.8] - [2.5] of the lack of accounting standards to measure such + d [0.5 - (-1.0)] = 2.75 percent x d activity. Guidelines used by auditors to assess loans should be consistent with the guidelines issued by The total sustainable debt ratio d increases with both bank supervisors and tax authorities. Auditors the primary deficit (including quasi-fiscal losses of often may agree with the bank supervisors, but with- the banking system), with the real interest rate paid out the agreement of tax authorities-and tax on debt, and when growth becomes negative. If inter- allowance for bad debt provisions-the banks are est rates are liberalized during this process, and dis- unwilling to provide for nonperforming loans. tressed borrowing raises the real interest rate to, say, Second, losses mount because there is no legal 10 percent a year, then change in d would increase framework for debt recovery. In a number of coun- correspondingly to 13.9 percent. In other words, a tries authorities were preoccupied with drafting a rising real interest rate would make the public sector banking law and neglected to strengthen court proce- debt burden unsustainable. dures to deal with debt recovery. Moreover, the A central issue in bank and enterprise restructur- administrative and technical capacity of the courts to ing has been the reluctance of governments to liberal- deal with massive liquidation and bankruptcy pro- ize interest rates for fear that the fiscal condition ceedings is extremely limited. Bankers, lawyers, bor- would run out of control. At the same time, reluc- rowers, and judges have little experience with com- tance to deal with bank and enterprise losses- plex debt resolution, particularly where losses have to because of the politically unacceptable levels of be allocated. Banks are reluctant to write off debt unemployment that would result-has worsened the because doing so could lead them to become severely macroeconomic environment and delayed the adjust- capital impaired. Tax authorities are reluctant to ment process. allow banks to make extensive provisions for fear of If the quasi-fiscal (flow and stock) losses are too further eroding the tax base. Courts and lawyers have large for the budget to absorb in a single carve-out, few legal precedents for adequate decisionmaking. how should the losses be allocated? The troika model Enterprise managers cannot preserve the value of (chapter 1, figure 1.4) suggests that bank restructur- assets so long as decisions on resolution are pending. ing cannot be undertaken without corresponding Finally, policymakers cannot make decisions unless reforms in the fiscal and enterprise sectors, including they are empowered to act. accompanying changes in the legal, accounting, and Thus the bad debt problem is likely to worsen regulatory frameworks. One approach is spelled out unless there are institutional and policy solutions to in the recent Chinese reforms, announced in the problem enterprises and borrowers. In essence, banks October 1993 Third Plenum of the Chinese and enterprises are suffering from a lack of clarity in Communist Party. At the banking level all policy- ownership and management of real sector resources. based loans will be removed from the banking system As privatization was delayed, enterprises and borrow- and transferred either to the budget or to specialized ers having liquidity problems became accountable to policy-based banks. This frees the banks to become no one, since management did not have the authority profit-motivated institutions. At the enterprise level or resources to correct liquidity or solvency issues and state-owned enterprises will be corporatized (made banks could not exercise financial discipline since into legal entities with equity capital, initially entirely they did not have sufficient capital or managerial state owned) and run according to market principles. skills to write off bad debts or convert loans to equity. Resolution and Reforin: Supervisory Remediesfr Problem Banks 65 Consigning problem borrowers to bankruptcy and and give central planning powers back to the public liquidation courts only worsened the process of prob- sector. Proponents of centralization argue that banks lem resolution since the courts also had no experience are not equipped to deal with enterprise restructuring with these issues. This led to gridlock, and in the and that decentralized restructuring would give a blank meantime assets deteriorate in value while awaiting check to banks to waste funds on behalf of the state. decisions on resolution. Both arguments have some validity, and the appropri- The design of the enterprise resolution problem ate solution depends on the conditions prevailing in requires a good database on the structure of the non- each country. A summary of the major advantages and performing loans. Data on classified loans should be disadvantages of each approach is shown in table 3.3. made available, broken down by class of enterprise, size of enterprise, economic sector, and the level of Damage control and the needfor sound intermediation provisioning for each class. There should also be an industrywide study on whether bank losses are due to Any restructuring scheme must address how to limit poor internal credit procedures or excessive overhead the flow of future losses in the banking system. expenditure. A cost-of-intermediation study would Centralizing bad debts in a debt restructuring agency is also reveal what caused the rise in intermediation one way of moving bad debts off the books of the spreads and whether banks can respond by reducing banks, but this approach would require fiscal resources overhead costs in the face of loan deterioration and that may not be available to many post-centrally declining business. planned economies. The Polish and Hungarian experiences suggest Centralized and decentralized solutions that giving senior bank management greater autono- my and discretion seems to improve bank perfor- Loan losses often are highly concentrated in a few mance, although the results are not uniform. Where large loss-making enterprises and economic sectors state banks are excessively burdened by inherited bad (box 3.7). If such concentration exists, the design of loans and no solutions are in sight, some banks may the bank and enterprise restructuring program will be engage in greater speculative lending, with further very different from one where loan losses are widely losses borne ultimately by the state. dispersed. Concentration in a few large enterprises or Corporatization is not enough if the underlying sectors suggests that the problems be dealt with at a weaknesses of credit procedures and lack of internal specialized or centralized level, whereas a widely dis- controls have not been corrected. Supervision of persed bad debt problem may require decentralized banks must be tightened. Without clear changes in solutions. the ownership of banks-whether to the private sec- There has been lively debate over centralized and tor (domestic or foreign) or to a specially formed state decentralized solutions to the bad debt problem holding company-no single institution will be able throughout Central and Eastern Europe. Opponents to exercise full shareholder rights to control losses. of centralized solutions fear that centralization would An institution-building program to strengthen reimpose bureaucratic solutions, delay privatization, internal bank controls, credit procedures, debt skills, Box 3.7 Concentration of nonperforming loans Recent evidence suggests that bad debt in post-centrally were exposures to state-owned enterprises. The twenty planned economies tends to be highly concentrated. In largest borrowers in each bank accounted for about 40 Romania in 1992, four sectors accounted for almost percent of their total assets. The fifty-eight largest three-quarters of nonperforming loans and about half of companies in the classified loans category account for interenterprise arrears. In three sectors, twenty companies 11 percent of total sales, 5 percent of total budget rev- accounted for nearly two-thirds of the total loans and bad enue, 7 percent of total subsidies, and 6 percent of debt in their sector. The largest enterprise in each sector total employment. Thus resolving these loss-making accounted for more than a third of that sector's bad debt. enterprises would have a considerable impact on Moreover, about 100 companies accounted for more growth, employment, fiscal expenditure, and future than three-quarters of bad loans. bank profits. In Poland in 1991, more than 80 percent of the bad and doubtful loans of the nine state-owned banks Sourre: World Bank sector studies. 66 BANK RESTRUCTURING Table 3.3 Centralized and decentralized approaches to bank and enterprise restructuring Centralized Decentralized Comments Bad debts transferred to one agency Banks responsible for bad debts Depends on banks' capacity to absorb bad debts Focused approach-restructuring Bad debt resolution integrated with banks' Depends on skill levels in banks skills are concentrated normal operations Ideal where bad debts are concentrated Ideal for dispersed small-borrower problems in a few large enterprises or sectors Frees banks to concentrate on serving Scale of enterprise problem cannot be divorced healthy customers from normal bank operations-carve-out would leave banks as empty shells Legal change required to transfer No legal change required-banks operate Transfers of claims to agency could be property rights to agency under debt recovery law legally problematic, and information could be lost on transfer of debt Budget bears brunt of losses Banks are more likely to be more careful in Neither solution may have a hard debt recovery, since their fiunds are involved budget constraint as long as ownership rights are unclear Could create a bureaucratic bottleneck Banks could be paralyzed because the size of Incentives structure must be clear- bad debts overwhelms their capaciry enterprise management, banks, and budget must find a solution that is practical and entails the least social cost and overall management-complemented by interna- shares in Hong Kong. In Russia and Yugoslavia liber- tional banking expertise wherever possible-should al bank licensing laws allow entcrprises to own banks be developed for each country. This program should to finance their own businesses, creating several prob- strengthen audit skills and disclosure requirements, lems for the future. such as the need to disclose accounts according to The use of interenterprise credit presents a diffi- international accounting standards. In Poland, for cult challenge to bank and enterprise restructuring. example, the "twinning" of nine state-owned banks Such credit can rise to the level of money supply dur- with foreign banks to strengthen bank management ing a period of tight money. Significant efforts to net appears to have helped prepare some of the banks for such credits were attempted in China and Romania. eventual privatization. Indigenous banking skills also Because of the large potential losses in such lending, were developed by licensing foreign banks to operate interenterprise credit can bring down good enterpris- in the domestic market, adding skills, techniques, es as well as loss-making ones. and technology, as well as competition. Enterprise restructuring and capital market development were Loss allocation-shouM the budget bear it all? assisted by the creation of specialized development institutions, such as the Polish Development Bank, The interlocking nature of the post-centrally which were better equipped to engage in corporate planned bank and enterprise problem requires that a finance work than the commercial banks. restructuring strategy be formulated by a high-level It is wrong to assume that the banking sector is so interagency committee comprising representatives of critical to intermediation that it is the only source of the budget ministry, the ministry in charge of privati- credit. In many post-centrally planned economies zation, the enterprise sector, the law ministry, the the banking sector has been constrained from granti- central bank, supervisory authorities, the banks, and ng new credit to state-owned enterprises because of the private sector. management caution and official credit restraint. In Hungary a banking supervision committee- Enterprises have relied instead on internally generat- comprising representatives of the State Banking ed earnings, interenterprise credit, export sales, and Supervision, the National Bank, the Ministry of joint ventures with foreign enterprises. In China, the Finance, the State Holding Company, and representa- Czech Republic, Hungary, and Poland foreign direct tives of the banking and enterprise sectors-was investment has been a major source of capital and established to advise the government and undertake know-how for corporate transformation. China also measures that would coordinate and strengthen bank has boosted corporate investment by developing supervision during restructuring. Such a committee stock markets in Shenzhen and Shanghai and floating focuses attention on the fact that bank and enterprise Resolution and Reform. Supervisory Remediesor Problem Banks 67 restructuring has to be an across-the-board effort. No Czech and Slovak Federal Republic had to recapital- bank and enterprise restructuring program can be suc- ize the Bank for Consolidation at least twice since its cessful without consensus from all parties concerned. formation. Hungary had to engage in a loan consoli- The early debate in some countries over whether dation scheme in 1992, despite guaranteeing 50 per- the banks or the budget should bear bad debt losses cent of the inherited bad loans as late as 1991. was unproductive since all banks and enterprises With respect to coverage of the carve-out, there belonged to the state at the early stages of transforma- should be only one uniform criteria for nonperform- tion. Delays in the removal of the bad debts only ing loans, to be agreed on by the banks, the auditors, worsened the problem. Governments initially tried to the supervisory authorities, and the tax authorities as avoid absorbing losses in the budget by allowing early in the process as possible. This makes the banks to widen their spreads and generate profits process more transparent and equitable in terms of from the inflation tax. This worsened incentives for treatment by different banks. Having different defini- debt recovery because banks found it easier to gener- tions of coverage for different bad debts of different ate credit than to recover bad debts. As bank posi- periods would generate different incentives for the tions deteriorated with the collapse of the Council for banks. Banks would be tempted to overstate certain Mutual Economic Assistance market, many Central categories of classified loans in order to receive prefer- and Eastern European economies had no choice but ential treatment. to act. Buyback. The second issue is whether the state Requiring the government to issue bonds to carve should reduce its burden by requiring a "buyback" out (or at least isolate) the nonperforming loans of feature or imposing a fee or tax on the banks that the banks raises two issues that are usually hotly benefit from recapitalization. Banks or their share- debated: the coverage issue and the fee or buyback holders would be required to buy back from the state issue. Both essentially involve the size of the state's the after-tax value of the bonds, since the bonds must share of the debt burden. be repaid from after-tax profits. Coverage. The first problem, determining the size If the fee or buyback is levied on the banks, then of the nonperforming loans, depends on the criteria the cost of the recapitalization effort is reduced by the chosen. Conceptually, all classified loans should be net present value of the fee or buyback feature. in fully provisioned, and if banks are undercapitalized other words, the banks have been given an asset that the government should recapitalize the banks by issu- is also a liability, the net effect of which may be zero ing bonds at the market interest rate sufficient to depending on the size of the buyback. This will not cover the capital shortfall. But since the definition of have the effect of recapitalizing banks in order to classified loans is still debated (since the value of col- improve their intermediation, since the banks may lateral and the flow of future losses are by no means pass the full extent of their burden onto borrowers or certain), the true size of the bond issue is subject to depositors by widening spreads. considerable variation. Budget officials are likely to The net effect is the same if the fee or buyback is argue on the basis on "affordability"-that increasing levied on the bank shareholders, except that the state debt during the transformation will likely sub- banks' balance sheet would improve at the expense of stantially worsen the fiscal position. the shareholders. But to the extent that the share- On the other hand, the banks and state holding holders force the banks to generate higher dividends companies that wish to privatize the banks argue that to repay the fee or the buyback, the "intermediation the banks should be cleaned up as much as possible. improvement" effect is correspondingly reduced. The state can avoid making an early decision by Imposing the buyback on the shareholders also cre- offering a blanket guarantee on nonperforming loans ates a new class of shares because all existing shares during the transition period, an approach that creates become subject to the "lien" of the buyback. This no immediate cash flow implications. Many govern- would hamper future efforts at privatization because ments avoid this path, however, because of the moral new shareholders will not want to share in the loss hazard risks implied. and old shareholders will want the newcomers to bear Given that the flow of losses will continue until some of the burden of past losses. the enterprises are restructured or privatized or until A more common method of avoiding the buy- macroeconomic and price stability has been achieved, back is for the state to increase the capital of banks there will be a continuing burden on the state. The (through ordinary or preferred shares) to the extent 68 BANK RESTRUCrURING of estimated losses. The increased share capital dilutes The role of banks in enterprise restructuring the equity of existing shareholders, and the bad debts can be written off against the enlarged capital base of To what extent should banks be involved in the the banks (or against the preferred share element) on restructuring of their problem borrowers? Banks have a parri passu basis. The state benefits from the bank perhaps the best concentration of skills-however recapitalization through future dividends or from the limited-and information on enterprise performance sale of such shares. The state can also use this oppor- and behavior, and through their long association with tunity to reduce the extent of bank shareholdings of enterprises have considerable supervisory power enterprises that are also substantial borrowers from because of their control of credit. their banks. Because enterprises in these economies have large It should be recognized that the budget cannot employment and social welfare responsibilities, how- bear the complete losses of the banking system all at ever, the banks do not have the political backing to once. Nor is the true size of the losses fully deter- engage in massive restructuring that would entail mined at this stage since it depends on other factors, retrenchment of labor, nor do they have the resources such as the macroeconomic environment. Still, if the to finance new long-term investments, since their government wants to improve intermediation in the own deposit base has very short maturity. banking system, it has to decide what level of net In addition, commercial banks in the West- resources it wishes to allocate to bank recapitaliza- unlike investment banks-have never been signifi- tion. The net effect can be reduced if the government candy involved in financing corporate restructuring. decides to issue bonds at a below-market interest rate, Supervisors in the post-centrally planned economies since these bonds will have to priced at market value. who enforce the traditional separation between com- There is therefore a clear tradeoff between the mercial banking and the corporate sector also have extent of state contribution (in the stock of bonds) built into the banking law safeguards against banks and the benefits in the improvement of bank inter- taking too much equity in enterprises. mediation. Since different banks price their spreads As discussed earlier, bank and enterprise restruc- differently depending on the assistance they receive, turing requires a rewriting of the national balance the method of recapitalization is critical. In other sheet, redistributing gains and losses. Successful words, the method of recapitalization will have incen- restructuring strengthens financial balances in all tive effects that may defeat the whole objective of domestic sectors. The issue is whether short-term recapitalization. bank funding is the correct mechanism to resolve fundamental issues of enterprise inefficiency and Privatization before restructuring? hence national losses. A related question is whether the state should absorb the losses of both the enter- Should state-owned banks be privatized before or prise and banking sectors through debt carve-outs or after restructuring? Experience with privatization, their equivalent. particularly where foreign investors are involved, The failure of the enterprise and banking sectors indicates that such exercises should be as credible and in these economies has much to do with the grossly transparent as possible. If banks are privatized (or overleveraged financial structures of central planning, minority shares sold to the public) before a clean-up, which centralized all risks (and ownership) in the the state may have to pay for larger clean-ups at a hands of the state. Given the distorted and inefficient later stage. pricing structure that prevails in these economies, Poland successfully privatized some banks after a requiring the banks to provide more debt to nascent period of restructuring and twinning, where foreign private sector enterprises or overleveraged state enter- banks advised state banks on privatization and inter- prises cannot improve their overall financial position. nal reorganization. Other post-centrally planned The long-run solution must come from redressing economies have had problems privatizing banks the debt-equity ratios of key sectors, moving them to before restructuring because foreign banks are cau- a more viable and sustainable position. This implies tious about investing in banks with uncertain asset that the domestic economy requires more capital- quality. In such cases the state would gain more from namely, savings, foreign investment, or retained earn- privatization if the banks were restructured before ings-rather than more debt. The banking system flotation. cannot provide long-term investment capital for Resolution and Refrm. SuptrvisoryI Remtdies for Problm Banks 69 enterprises, only short-term working capital. But it how difficult the solutions. There has never been a can help the restructuring process by imposing finan- single, quick solution to financial distress. What this cial discipline, reducing transaction costs through volume hopes to show is that there are common intermediation, and helping to weed out inefficient approaches and integrated processes, taking into enterprises. account the need to be consistent in rewriting the The solutions to enterprise restructuring have to national balance sheet, along with important institu- come from the generation of domestic equity, either tional, legal, and commercial changes in markets, from privatization or from removing trade and distri- products, and human skills, that could lead to the butional barriers to market entry, which would foster design of better restructuring programs. domestic savings in the form of private entrepreneur- ial retained earnings. Encouraging foreign direct References investment also helps. Fostering capital and equity markets clearly supplements this effort. American Banket: December 27, 1990. The development of restructuring enterprises Bank Negara Malaysia. 1987. Annual Report 1986. Kuala should be left to investment banks, which can be cre- Lumpur. .. . . . ~~~~~~~~Basle Committee on Banking Supervision. 1992. Internzational ated through joint ventures with foreign investment Developments in Banking Supervision. Basle, Switzerland. banks with expertise in such restructuring. Resolving Butsch, Jean-Louis. 1992. "The Role of Banking Supervisory the real sector will require supply-side solutions, Authorities in the Prevention of Systemic Risk." Paper pre- which cannot be undertaken by inducing the bank- sented at the seventh international conference of banking ing system to direct credit to large loss-making ven- supervisors, October 7-9, Cannes. tures. Such credit would put the short-term deposits Corrigan, E. Gerald. 1992. "Strengthening International Cooperation between Banking Supervisors." Paper present- of the banking system at risk levels that would desta- ed at the seventh international conference of banking bilize the economy in the long term. supervisors, October 7-9, Cannes. The transformation of previously state-owned deJuan, Arisrobulo. 1991. "Does Bank Insolvency Matter? And factors of production toward market orientation What to do About It?" Economic Development Institute requires that economic decisions ultimately be Working Paper. World Bank, Washington, D.C. Emery, Robert. 1993. "Developing a Government Bond decentralized to the marketplace. As stated earlier, Market." Financial Sector Development Department bank restructuring is ultimately about building Technical Paper. World Bank, Washington, D.C. sounder financial structures and clearer property Federal Reserve Board, Federal Deposit Insurance Corporation, rights. The process of bank restructuring, diagnosis, and U.S. Comptroller of the Currency. 1992. "Joint Report damage control, loss allocation, and getting the to the Senate Banking Committee." Washington, D.C. incentives right is essentially about the definition of Fischer, Stanley, and William Easterly. 1990. "The Economics of incentives right is essentially abOUt the definition of the Government Budget Constraint." World Bank Research property rights and obligations. Hence, rewriting the Observer5(2):127-42. national balance sheet will require a closely coordi- Gelb, Alan, and Patrick Honohan. 1989. "Financial Sector nated effort of reforms in all sectors, in which com- Reforms in Adjustment Programs." Policy Research mercial banking is only one component. And Working Paper 169. World Bank, Washington D.C. because many institutional changes are involved, the Gibeault, Jean. 1993. "Credit Risk Bureaus." Financial Sector process will necessarily take time. Development Department Technical Paper. World Bank, Washington, D.C. Glaessner, Thomas, and Ignacio Mas. 1991. "Incentive Structure Conclusion and Resolution of Financial Institution Crises: Latin America Experience." Latin America Technical Department We are witnessing a dramatic integration of the Technical Paper. World Bank, Washington, D.C. post-centrally planned economies with the inte. Goldstein, Morris, and David Folkerts-Landau. 1993. "The post-centrally planned economies with the interna- Growing Involvement of Banks in Derivative Finance: A tional market economy. This transformation will Challenge for Financial Policy." In International Capital bring with it enormous challenges of bank and enter- Markets, Part I: Systemic Issues in International Finance. prise restructuring. As stated earlier, this volume is Washington, D.C.: International Monetary Fund. not a definitive guide for bank supervisors, encom- Hanson, James, and Roberto Rocha. 1986. "High Interest Rates, passing all the ideas and experiences related to the Spreads, and the Costs of Intermediation." Industry and phenomenon of financial distress. A wide range of Finance Series 18. World Bank, Washington, D.C. Hove, Andrew C. 1994. "Resolving Failed Banks." Paper pre- ideas have been brought together to show how com- sented at the eighth international conference of banking plex the origins and effects of the problems are, and supervisors, October 12, Vienna. 70 BANK RESTRUCTURING IMF (International Monetary Fund). 1993. International . 1992. 'Bad Debts in Transitional Socialist Economies." Capital Markets, Part II: Systemic Issues in International World Bank, Financial Policy and Systems Division, Finance. Washington, D.C. Washington, D.C. Lin See Yan. 1992. 'Financial Reform in Malaysia: The Sheng, Andrew, and Yoon Je Cho. 1993. "Risk Management Institutional Perspective." Bank Negara Malaysia, Kuala and Stable Financial Structures." Policy Research Working Lumpur. Paper 1109. World Bank, Washington, D.C. McKinnon, R. 1991. "Taxation, Money, and Credit in Liberalizing Snoy, Bernard. 1992. "Privatization and Financial Sector Reform Socialist Economies: Asian and European Experiences." in Central and Eastern Europe." Paper presented at the Stanford University, Department of Economics, Stanford, Colloquium of Association Internationale de Droit Calif. Economique (AIDE), Budapest. McPherson, Donald. 1992. "Reformation of the Regulatory Summers, Bruce J. 1992. "The Role of the Central Bank in Framework for Financial Institutions in Canada." World Developing Payment Systems." Paper presented at the nine- Bank, Financial Policy and Systems Division, Washington, teenth SEANZA central banking course, November, Tokyo. D.C. Talley, Samuel H., and Ignacio Mas. 'Deposit Insurance in Musch, Erick. 1992. "A Framework for Measuring and Managing Developing Countries." Policy Research Working Paper Liquidity." Paper presented at the seventh international con- 548. World Bank, Washington, D.C. ference of banking supervisors, October 7-9, Cannes. U.S. Comptroller of the Currency. 1992. "Remarks of William Nakajima, Zenta, and Hiroo Taguchi. 1993. "Toward a More P. Bowden, Jr. on National Banks in the United States." Stable Financial Framework: Long-Term Alternatives." Quarterly Journal I 1(): 51-61 . Washington, D.C. Bank of Japan, Institute of Monetary and Economic U.S. GAO (General Accounting Office). 1989. Thrift Failures. Studies, Tokyo. Washington, D.C. National Bank of Hungary. 1992. "Monthly Report, U.S. Treasury. 1991. "Modernizing the Financial System.' September." Budapest. Washington, D.C. Quinn, Brian. 1991. "Techniques for Dealing with Problem World Bank. 1993. "Bank Supervision: Suggested Guidelines." Banks." In P. Downes and R. Vaez-Zadeh, eds., The Financial Policy and Systems Division, Washington, D.C. Evolving Role of Central Banks. Washington, D.C.: Zahler, Roberto. 1993. "Financial Secror Reforms and International Monetary Fund. Liberalization: Welcome Address." In Shakil Faruqi, ed., Sheng, Andrew. 1991. "Bank Supervision: Principles and Financial Sector Reforms in Asian and Latin American Practice." Economic Development Institute Working Countries. Economic Development Insiture Seminar Series, Paper. World Bank, Washington, D.C. Washington, D.C.: World Bank. CHAPTER 4 The United States: Resolving Systemic Crisis, 1981-91 Andrew Sheng The United States has the largest banking system in the system, the deposit insurance and regulatory regimes world, with 29,000 mostly small, regionally restricted created in the 1930s were designed not only to pro- deposit-taking financial intermediaries and more than mote stability, but also to facilitate orderly exit. $5 trillion in assets. With 12,500 commercial banks, Despite the advanced supervisory system, however, the United States has about ten times as many banks there were 1,100 commercial bank failures and 630 per capita as the other countries in the Group of Ten thrift (savings and loan institutions) insolvencies combined. Japan has only 150 commercial banks, and requiring official intervention during the 1980s. The the United Kingdom 600 licensed deposit-takers number of thrift institutions fell by more than 30 per- (Miller 1991). The United States also has what is prob- cent and banks by about 14 percent through market- ably the most complex statutory regulatory regime in driven consolidation within the industry, officially existence (table 4.1). Deposit insurance in the U.S. assisted resolutions, and occasional liquidation. This banking system is comprehensive-covering all record of instability was unparalleled since the banking deposits under $ 100,000-and practically universal. crisis of the 1930s, when 4,370 banks failed between Because the United States traditionally has favored 1933 and 1940-4,000 of them in 1933 alone. By a unitary, rather than an interstate, branch banking 1991 U.S. taxpayers had contributed at least $80 Table 4.1 Regulatory mecharisms for U.S. depository institutions as of June 30, 1990 Share of total assets of Assets depository (billions of institutions Institution Number U.S. dollars) (percent) Rulatry agency Commercial banks 12,502 3,360.0 66.7 Office of the Comptroller of the Currency: 'National" banks Federal Reserve Bank: Bank holding companies and state- chartered members of the Federal Reserve System Federal Deposit Insurance Corporation: National and state- chartered, federally insured, nonmembers of the Federal Reserve State agencies Savings banks insured by the Bank Insurance Fund 461 233.4 4.6 Federal Deposit Insurance Corporation Thrifts and other savings banks' 2,878 1,251.7 24.8 Office of Thrift Supervision: All institutions insured by the Savings Association Insurance Fund, induding most state- chartered thrifts State agencies Credit unions 13,102 195.3 3.9 National Credit Union Administration Total 28,943 5,040.4 100.0 Five federal and fifty state agencies a. As ofDccember31, 1989. Source: U.S. GAO 1991; U.S. Treasury 1991. 71 72 BANK RESTRUCrURING billion to "cleaning up" the thrift industry, with still few hundred bank failures prior to 1989 were more demands to follow. In addition, the 1991 FDIC resolved comfortably within the resources of the Improvement Act approved $70 billion to recapitalize FDIC and had no discernible real economic or the Bank Insurance Fund operated by the Federal financial impact. It was not until the end of the Deposit Insurance Corporation (FDIC). How could a decade-when the magnitude of large bank failures debacle of this magnitude occur in a system of this size threatened to bankrupt the deposit insurance fund and sophistication, where no less than five federal and for banks-that uncomfortable parallels with the fifty state regulatory agencies oversee deposit-taking early thrift crisis arose. Moreover, the decrease in activities, and where comprehensive deposit insurance profitability and increase in the domestic loan losses has existed for more than fifty years? of the banks grew as the economic recession that began in 1990 continued. This provoked a very real The Origins of Failure fear of a domestic credit crunch and further defla- tionary consequences. Many of the reasons for the exceptional number of There are also important similarities in structural failures are common to other banking systems in dis- weakness, which came to the fore in the second half tress. Nevertheless, the U.S. case merits special com- of the 1980s. Unlike their European counterparts, ment because of its sheer cost in nominal terms and the U.S. bank and thrift systems are highly segment- the lessons it holds for developing countries with less ed functionally, geographically, and in terms of regu- sophisticated banking systems. This chapter focuses latory oversight. Furthermore, until the mid-1980s on two main groups of institutions: the roughly 3,000 deposit interest rate ceilings under Federal Reserve thrifts with one-quarter of total depository system Regulation Q constrained the ability of banks and assets and the 12,000 commercial banks with two- thrifts to attract personal savings.' These ceilings thirds of system assets. Although their problems heightened the competition between the two types of evolved differently-and the thrift industry crisis institutions for domestic savings and encouraged the unfolded earlier-they share a legacy of cyclical and growth of money market funds and other deposit structural weaknesses. Specifically, they faced drastic substitutes. Specifically, the banks lost their franchise changes as technological innovation fostered competi- value in the 1980s and sought to regain market share tion, they operated in a regulatory environment within the confines of regulatory corsets. Even replete with overlaps and gaps, their operating free- though Regulation Q was eventually removed, the dom was segmented both functionally and geographi- built-in structural rigidities impeded the banks' and cally, and they suffered the moral hazard risks of a thrifts' ability to make a smooth adjustment to a comprehensive deposit insurance umbrella. These world of unstable relative prices or to accommodate structural weaknesses were exacerbated by cyclical fac- to competitive changes in domestic and international tors when the U.S. economy entered a recession in markets, including the rapid pace of technological 1980-81, emerged with a period of strong growth, innovation that began in the 1970s. and then entered another recession in 1990. During the 1970s and 1980s the banks lost their There were, however, important initial qualitative franchise on the payments mechanism, low-interest differences between the problems of thrifts and those deposits, and loans to commerce and industry due to of banks. First, structural differences affected their technological innovation and competition from cred- ability to adapt to changing market conditions. it card companies, money market funds, insurance Thrifts were historically fixed-rate mortgage lenders companies, mortgage pools, and pension funds. with variable deposit rate liabilities and thus were From 1960 to 1994 the depository institutions' share subject to massive interest rate risks, whereas banks of financial intermediary assets fell from 39 percent faced less severe maturity mismatches and had to about 29 percent (Edwards and Mishkin 1995). In greater investment flexibility. As a result the thrift 1960 bank loans to commerce and industry were industry suffered a systemwide technical insolvency nearly ten times the size of the commercial paper of truly crisis proportions during 1980-83, when market. By 1989 that ratio had fallen to just under U.S. domestic interest rates rose to unprecedented 1.5 times. Demand deposits of banks fell from 60 levels. By contrast, the commercial banking system, percent of total deposits to less than 20 percent over despite more than 1,000 bank failures during the the same period. Competition also came from foreign 1980s, remained both solvent and profitable. The banks. Between 1983 and 1990 the percentage of The United States: Resolving Systemic Crisii, 1981-91 73 commercial and industrial loans provided by foreign ment and ultimately guaranteed by it as well. banks rose from 20 percent to just over 29 percent. Historically, thrifts were organized primarily as mutu- Another factor contributing to the problems of al associations, and they provided long-term, fixed- banks and thrifts was the extensive financial deregula- interest, predominantly residential mortgages for tion in the early 1980s, which coincided with an their local communities. Accordingly, their asset base increase in deposit insurance coverage. These develop- was more specialized than that of commercial banks ments allowed greater bank gearing and expanded and their overall interest rate risk significantly higher. portfolio diversification. They also encouraged new The legislative foundations for this system were laid entrants and promoted rapid growth. These sudden in the 1930s in response to the failure of more than changes also encouraged higher risk-taking and specu- 4,000 U.S. banks. Despite the evolution of bank and lation by some institutions in highly leveraged trans- thrift holding companies as vehicles for circumvent- actions and real estate. For most of the decade this ing statutory restrictions on functional and geograph- speculative expansion was underpinned by an unusu- ic diversification, until the 1990s no nationwide ally long period of domestic economic growth pro- branching system had evolved. moted by the free market, supply-side economic phi- losophy and policies of the Reagan administration. At Regulatory overlaps and gaps the same time, deposit insurance coverage was extend- ed widely through the "too big to fail" principle when The supervisory system allows for regulatory overlap regulatory authorities intervened to save a number of and regulatory arbitrage because banks and thrifts large institutions. Brokered deposits and "pass- can be chartered by federal or state authorities. Over through coverage" also overextended insurance cover- time, however, the balance of supervision has fallen age. Consequendy, extensive moral hazard risks arose on the federal government, largely because of the fed- at a time when regulatory oversight-particularly, but eral guarantee backing the industry-financed deposit not exclusively, of the thrifts-was relaxed due to insurance. Each of the four principal federal supervi- budgetary cutbacks.2 sory agencies has an extensive off-site surveillance sys- These cumulative factors-loss of franchise, tem backed by selective on-site examination process- deregulation and intense competition, geographical es. These agencies employed some 6,800 examiners and functional segmentation, moral hazard, and gaps at the end of 1989, 1,500 more than at the low point in regulation-reinforced losses rooted in macroeco- in 1984. nomic and microeconomic factors. The combined The Office of Thrift Supervision audhorizes nation- effects of regional economic downturns, poor bank al thrifts and savings banks, whereas the Office of the management, imprudent lending practices in foreign Comptroller of the Currency, an agency of the U.S. debt and then real estate, and, in the case of thrifts, Treasury, charters national banks. The term national extensive fraud, generated unprecedented losses in bank refers to a bank chartered by federal agencies. the U.S. banking system. A structural risk asymmetry Since 1989 the Office of Thrift Supervision has regu- existed: the government had an open-ended obliga- lated virtually all thrifts.3 The Board of Governors of tion to protect deposit liabilities during a period the Federal Reserve System (the central bank) supervis- when banking asset values were exposed to volatile es bank holding companies; its regional Federal market forces and private sector mismanagement. Reserve Banks are also responsible for state-chartered banks that are members of the Federal Reserve System The System of Depository Institutions and that have access to the central bank discount win- dow and lender-of-last-resort facilities.4 The United States has no national banking system, Deposit insurance falls under the purview of the making it unique among industrial countries. Federal Deposit Insurance Corporation (FDIC). The Reflecting the populist mood of the country against FDIC also supervises, often joindy with state charter- the concentration of wealth and power, the banking ing agencies, its state-chartered member banks that and the thrift systems were designed with two aims: do not belong to the Federal Reserve System. Until to prevent market concentration by preserving August 1989 there was also a Federal Savings and regional and functional markets and to ensure safety Loan Insurance Corporation (FSLIC), which all fed- and soundness under industry-financed deposit pro- erally chartered thrifts were obliged to join and to tection, which was mandated by the federal govern- which most state-chartered ones also belonged.5 A 74 BANK RESTRUCTURING Savings Association Insurance Fund, maintained sep- Glass-Steagall provisions to permit limited securities arately within the FDIC and financed by the thrift activities. Moreover, some state-chartered banks industry, was created in 1989 to replace the old have enjoyed more freedom anyway. The U.S. deposit insurance fund beginning in August 1992. In Treasury's 1991 reform proposal would have abol- the interim, thrift deposit insurance was provided by ished Glass-Steagall restrictions. the newly created Resolution Trust Corporation (RTC). The RTC was financed partly with federal Deposit insurance encouraged moral hazard behavior funds and was charged with taking over and resolving thrifts that failed after August 1989. The two deposit insurance funds created in the early Geographical segmentation in the system is 1930s, FSLIC (for thrifts) and FDIC (for commer- enforced through limitations on interstate banking cial banks), virtually eliminated bank runs in the and limitations on intrastate branching. The United States, but their intrinsic flaws emerged dur- McFadden Act of 1927 allows states to restrict or pre- ing the 1980s. For example, banks pay deposit vent branching, and since most states do not permit insurance premiums on a flat-rate basis, regardless entry into their territory through branches of banks of the different risks inherent in each bank. from another state, interstate branching is effectively Furthermore, the insurance coverage per account prohibited. The need to attract out-of-state capital to was raised from $40,000 to $100,000 in 1980, a assist in resolving local banking problems has led to fourfold increase in real terms from the original some relaxation of these restrictions, however. By $2,500 set in 1933.6 By relaxing the definition of 1991 almost all states permitted interstate banking by deposit coverage to include "brokered" deposits, bank holding companies owning local banks through thrifts and banks were able to finance growth from state-chartered subsidiaries. To enable banks and access to savings nationwide.7 Furthermore, the thrifts to grow out of their problems, most states also deposit insurance scheme was implicitly extended to relaxed their limits on intrastate branch banking by cover 100 percent of the deposit liabilities of thrifts banks headquartered in their states. Nevertheless, and banks, with few exceptions, under the "too big laws still prevent interstate bank branching through- to fail" principle. Regulators decided not to allow out the United States. Reform proposals introduced some larger banks to fail and guaranteed all by the U.S. Treasury in 1991 would have allowed deposits-the bailout of Continental Illinois in nationwide branching by banks. Most of these highly 1984 is a notable example. And in the case of offi- controversial recommendations, however, failed to cial thrift resolutions, it was also common practice receive the necessary congressional approval. to protect all deposits. Functional segmentation is mandated by the The FSLIC actually had to be closed in late 1989 Banking Act of 1933 (the Glass-Steagall Act), which after a failed attempt to recapitalize it under the 1987 prohibits all Federal Reserve members from engag- Competitive Equality Banking Act. It wound up with ing in investment banking activities, particularly the largest corporate deficit ($87 billion) ever report- from dealing directly in underwriting or purchasing ed (U.S. GAO 1990). At the same time, the com- securities (except for certain government issues). mercial banks' insurance fund in the FDIC was Moreover, the act prohibits any person or company renamed the Bank Insurance Fund. Consequently, at that issues, underwrites, sells, or distributes securi- the end of 1989 the FDIC and RTC together were ties from taking deposits from the public. U.S. providing deposit insurance coverage to 12,706 com- money center banks therefore cannot engage in mercial banks, 476 state-chartered and 20 federally investment banking and securities activities that are chartered savings banks, and 2,878 federally and open to European banks. While this functional state-chartered thrifts (U.S. Treasury 1991). segregation made sense when banks had an exclusive "franchise" in deposit-taking activities, this profit Mlacroeconomic Developments base was eroded substantially in the 1970s. Money market funds and other deposit-substituting innova- The structural weaknesses in the U.S. banking tions offered by investment banks and nonbanks system were exacerbated by cyclical factors. During now also capture domestic savings. Consequently, the 1970s inflation rose as a result of successive oil functional segmentation also has eroded in recent shocks. As part of the recycling of petrodollar sur- years as federal authorities relaxed or reinterpreted pluses, banks began their aggressive lending to The United States: Resolving Systemic Crisis, 1981-91 75 developing countries. Deregulation in the securities, end of the decade. Nevertheless, different sectors and telecommunications, airline, and natural gas markets regions were affected at different times with varying generated more competitive pressure in the econo- severity by changes in the business cycle and relative my. As competition from savings banks, real estate prices. Regional economic recessions, in turn, trusts, mutual funds, and pension and insurance adversely affected banks and thrifts in those areas. In funds began to erode the banks' market share, they particular, declining prices and exports of U.S. farm responded by diversifying out of traditional bank products resulted in a sharp deterioration in agricul- lending activities, mainly through the creation of tural land values, leading to the failure of Midwest bank holding companies. farm banks in the early and mid-1980s. Similarly, The escalation of thrift failures in the early 1980s from 1982 to 1988 nominal domestic oil prices fell was a consequence of the 1979-82 recession. In late by 57 percent, with severe impact on oil-producing 1979 the Federal Reserve severely tightened mone- states in the South and Southwest. Thrifts and banks tary policy in an effort to curb inflation. Treasury bill had expanded aggressively in these areas on the basis rates rose to nearly 16 percent a year in 1981, com- of oil wealth accumulated in the late 1970s. In Texas, pared with an average of 10 percent in 1979 (5.5 per- for example, thrift assets increased from $35 billion cent in 1977). These events, combined with a sud- in 1982 to $85 billion in 1985 (U.S. Treasury 1991). den, sharp increase in world oil prices, acted as a In addition, tax incentives stimulated real estate brake on the economy. Concurrently. the steep construction in 1982, with an across-the-board appreciation of the dollar during 1981-84 affected impact on residential and commercial property devel- the international competitiveness of U.S. firms, opment. The speculative boom in property was which were already hurt by rising financial and fuel financed by liberal lending by thrifts, which sought to costs. Dollar appreciation also had an adverse impact grow out of their problems. But when nominal inter- on the U.S. agricultural sector. When the economic est rates fell to their lowest levels in 1987, a serious slowdown in 1979-80 was followed almost immedi- general overcapacity was exposed. This oversupply, on ately in 1982 by the worst recession since the 1930s, top of regional economic recessions, created pressure unemployment rose and business bankruptcies on bank asset and collateral values. First in the South reached a post-World War 11 high. Growth in real and Midwest and subsequently in the Northeast and gross national product (GNP) dropped from 1.9 per- Southwest, each episode of localized economic down- cent in 1981 to -2.6 percent in 1982 (table 4.2). turn had a direct impact on local banks and thrifts, The U.S. economy was sufficiently large and which were virtual captives given their inability to diversified to absorb these shocks, and it rebounded diversify risks through nationwide branch operations. to enjoy steady economic growth from 1983 to the In 1987, for example, 90 percent of bank failures Table 4.2 Macroeconomic indicators, 1979-90 (percent unless otherwise specified) Fxchange rate Real Current Certificate of Prime (U S. dollarsl growth accountl Terms of trade Inflation Government deposit (CD) lending .Special Drawing Year in GDP GDP (1980=100)" rateb balance/GDP deposit rate rate Rights) 1979 2.1 0.0 105.5 11.3 -1.4 11.2 12.7 1.3 1980 -0.1 0.0 100.0 13.5 -2.8 13.0 15.2 1.2 1981 2.0 0.2 110.4 10.3 -2.6 15.9 18.9 1.1 1982 -2.5 -0.2 110.0 6.2 -4.3 12.3 14.9 1.1 1983 3.7 -1.2 121.2 3.2 -6.0 9.0 10.8 1.0 1984 7.0 -2.7 123.1 4.3 -4.7 10.4 12.0 0.9 1985 3.6 -3.1 124.3 3.6 -5.3 8.0 9.9 1.1 1986 3.0 -3.4 135.7 1.9 -5.0 6.5 8.3 1.2 1987 3.8 -3.6 130.2 3.7 -3.3 6.9 8.2 1.4 1988 4.4 -2.7 146.8 4.0 -3.2 7.7 9.3 1.3 1989 2.5 -2.1 150.1 4.8 -2.8 9.1 10.9 1.3 1990 0.8 -1.7 98.5 5.4 -4.0 8.2 10.0 1.4 - Not available. a. Ratio of the index of average export prices to the index of average import prices. b. Consumer price index. Source: World Bank dara; IMF, various years. 76 BANK RESTRUCTURING occurred in states that still prohibited or at least limit- gages until 1980. With their variable lending rates, ed intrastate branching (U.S. GAO 1987). commercial banks were not as seriously affected by Beginning in 1983 the national economy grew interest rate risks, except for a few mutual savings steadily by borrowing. The supply-side economic banks in the Northeast.8 For the most part, however, philosophy of the period brought cuts in income tax bank failures were confined to specific geographic rates and led to rising fiscal deficits. These budget regions, involving mainly small banks in agricultural shortfalls averaged 5 percent of GNP during and oil-producing states.9 They did not have systemic 1982-86, nearly double the level in the previous five ramifications and were easily managed by the well- years. At the same time, household savings declined capitalized FDIC. Local banks with generally conserv- from 5.0 percent of GNP to 3.8 percent in the ative management and adequate capital and reserves 1980s. As corporate savings also fell in response to were able to weather their temporary distress. increased competition, the net savings rate dropped In the face of wholesale thrift insolvencies, and to only 3.0 percent in the 1980s, less than half of its with an increasing number of small bank failures, pres- historical average of 7.5 percent (Harris and Steindel sure mounted to deregulate. Encouraged by the popu- 1991). lar antipathy toward government interference charac- In the face of this relaxed fiscal stance, tight mone- teristic of the Reagan administration, legislators and tary policy sought to keep real interest rates high- regulators complied in the early 1980s. Consequently, they averaged 4.5 percent a year from 1983 to the end the general outlook for banks and thrifts changed of the decade-compared with the negative rates that markedly in the mid- and late 1980s. Sustained had prevailed during the late 1970s. Short-term rates growth nationwide and growing competition prompt- remained above those of the country's key competitors ed banks and thrifts to expand credit rapidly, relax until the closing years of the decade, which induced credit assessments, and engage in imprudent leveraging capital inflows that financed the rising fiscal and cur- and outright moral hazard. rent account deficits. As domestic consumption rose Many banks, especially the larger ones, had prob- with tax cuts and a favorable exchange rate, the lems arising from a particular exposure. In the late United States experienced the longest nninterrupted 1970s banks had enjoyed a large expansion in petrodol- economic expansion in postwar history from 1983 to lar deposits from oil-exporting countries following their the end of the decade. Unemployment and inflation windfall oil price increases. A significant amount of this fell to levels not seen since the early 1970s. Prosperity wealth was placed with U.S. money center banks in the was accompanied by a prolonged bull market in com- Eurodollar markets. It financed large-scale lending to pany shares, rapid appreciation of property values, developing countries at a time when U.S. domestic securitization of debt, and a boom in corporate merg- lending opportunities were diminished by the impact ers and acquisitions. Many of these were financed by of the U.S. recession. By 1983 international business high-leverage transactions and innovative junk bonds. accounted for one-third of U.S. banks' net income. The high interest rate environment also intensi- Between March 1981 and September 1982, when the fied competition for new deposits from nonbank and Mexican debt crisis occurred, U.S. banks' exposure to new financial institutions. To create a more level developing countries rose 76.1 percent, to $220.7 bil- playing field, Regulation Q retail deposit rate ceilings lion. Their exposure to twelve highly indebted coun- were relaxed beginning in 1980 and eventually elimi- tries more than doubled, rising from $43.4 billion to nated. Although these moves improved liability man- $91.1 billion. After 1982, howevcr, this debt became a agement, the endowment effect of cheap deposit lia- drag on large banks' profitability as loan arrears and bilities was permanently eroded for both banks and losses mounted. Banks were driven back to the domes- thrifts. They saw a steady decline in aggregate finan- tic market in search of higher-yielding investments and cial strength throughout the decade. fee-based income to boost earnings, support loan loss There were, therefore, two broad phases to the provisioning, and replenish eroding capital. U.S. bank and thrift crisis. In the early 1980s the A 1988 study by the Office of the Comptroller of thrift industry, whose assets were primarily long-term the Currency revealed that of 171 bank failures fixed-rate mortgages, became technically insolvent fol- between 1979 and 1987, the poor economic environ- lowing the sudden jump in U.S. interest rates. Their ment was a contributing factor in only 35 percent of costs of funds increased sharply, yet most of them did the cases. And in no case was it the principal cause of not have the freedom to offer adjustable rate mort- failure. Poor bank management, on the other hand, The United States: Resolving Systemic Crisu, 1981-91 77 was apparent in 90 percent of the failures. In the thrift Table 4.3 Selected data for federally insured thrifts, industry, where macroeconomic externalities played a 1980-90 greater role in early failures, failures after 1985 were Deposits Earnings attributed primarily to fraud, gross mismanagement, (billiom of (billions of and extensive speculative investments. Failures Year Number US. dollars) US. dollars) increased sharply and were qualitatively different dur- 1980 4,005 503.2 - ing the second half of the decade. 1981 3,785 519.9 -6.2 ing ~~~~~ ~~~~~~ ~~~~~1982 3,349 560.5 -5.9 1983 3,183 671.1 2.6 The thrifi crisis, 198083 1984 3,136 784.5 1.0 1985 3,246 843.9 3.7 1986 3,220 890.7 0.1 With continued higher costs of funds against an asset 1987 3,147 932.6 -7.8 base that was tied to fixed-rate lending, the thrift 1988 2.949 971.5 -13.4 1989 2,878 945.7 -17.6 industry become insolvent virtually overnight. 1990 2,699' 885.1 6.5' Between 1980 and 1983, 470 thrifts failed. The net - Nocavailable. insolvency of the thrift industry was estimated at a. As ofJune. $100 billion in 1982 if the assets were marked to mar- Source: U.S. Treasury 1991; U.S. OTS, various years. ket (U.S. Treasury 1991). The FSLIC had insufficient resources to cope with across-the-board depositor pay- aggregate earnings resulted between 1983 and 1985 offs and liquidations. Because the technical insolvency (table 4.3). of the thrifts was systemic, a stock solution would have placed too high a cost on the budget (and on Bank and thrifi problens, 1985-89 taxpayers) in a political climate that favored budget cutting. Congress consequently adopted a flow reme- As late as June 1987, nine U.S. money center banks dy to allow thrifts to grow out of their problems. The still held two-thirds of all U.S. banks' claims on fifteen Depository Institutions Deregulation and Monetary heavily indebted countries, an amount equivalent to Control Act of 1980 gradually eliminated Regulation 113 percent of their total capital. From 1985 to 1989 Q interest rate ceilings, raised the deposit insurance U.S. banks managed to reduce the ratio of direct loans ceiling to $100,000, and permitted federally chartered to developing countries from 5.9 percent of total loans thrifts to offer adjustable rate mortgages."0 In 1982 ($96.6 billion) to 3.3 percent ($66 billion), and they the Garn-St. Germain Act accelerated the pace of lib- began large-scale loan loss provisioning in 1987. Data eralization, authorized certain capital forbearance on U.S. banks' loan losses on developing country debt measures, and permitted emergency interstate acquisi- is not available, but in 1987 the large money center tion of failed thrifts (U.S. GAO 1987). banks made loss provisions of $20.7 billion on their Federally chartered thrifts obtained greater access developing country portfolios, causing the first annual to funding through (insured) brokered deposits, and loss for the U.S. banking system as a whole. they were allowed to convert from mutual status to The banks are still charging off their developing limited liability companies. In addition, authorities country debt, a large component of which was sold pressured troubled but solvent thrifts into "superviso- in the secondary market, rescheduled, or forgiven ry" mergers without financial assistance from the under Brady debt initiatives. In 1991, 13.4 percent of FSLIC, while market forces induced more than 400 net loans to foreign governments and institutions was voluntary mergers in 1981 and 1982 alone. Regulators charged off, so that by the end of 1991 such loans also granted an array of special capital and regulatory totaled $26.3 billion, or less than 1.3 percent of total forbearance measures that permitted thrifts to stay in loans. business, sometimes for years, even though they were The developing country debt burden had a long- insolvent."' Nevertheless, between 1980 and 1983 run impact on the banking industry as a whole. more than 850 thrifts were closed or merged, mostly When the international debt crisis erupted in 1982, through industry consolidation and a few officially U.S. banks turned homeward for growth and earn- managed liquidations (Barth, Bartholomew, and ings prospects in order to build reserves against Bradley 1990). When nominal interest rates dedined prospective losses on their developing country loans. after 1982, the positive impact on thrift profitability In the face of heavy competition, they tried to maxi- appeared to justify the liberalization. Modest positive mize returns on equity by engaging in off-balance 78 BANK RESTRUCTURING Table 4.4 Selected financial ratios for commercial than 190 percent of their net income. This occurred banks, 1930-89 despite widening net interest margins, which reached (percent) their highest level (3.41 percent) ever during the sec- Ratio 1930-39 1970-79 1980-89 1980-84 1985-89 ond half of the 1980s. Equity/ Of particular concern were debt overhang and real assets 11.88 6.39 6.11 5.96 6.22 estate loan exposure. During the 1980s the U.S. cor- Loans/ porate sector engaged in widespread overleveraging. assets 29.43 53.73 57.75 54.71 59.94 With the liberalization of several big industries, merg- Loans/ equity 2.48 8.41 9.45 9.18 9.64 ers and acquisitions rose sharply as industries simulta- loans - 1.33 1.78 1.14 2.20 neously engaged in massive competition and consoli- Loans/ dation. Debt-financed acquisitions through leveraged deposits 33.92 65.00 74.15 69.67 77.42 buyouts and the issue of junk bonds (subordinated Return on debt) became fashionable. Tax incentives favored debt assets 0.46 0.77 0.61 0.69 0.55 Return on creation, while defense against takeovers resulted in equity 3.84 12.09 9.94 11.65 8.77 corporations retiring their own equity. Net inerest. 3.32 3.0 3.41 Between 1984 and 1990 U.S. corporations built Net loan up $400 billion in debt, raised $650 billion in secu- chargeoffs/ rities markets, and retired $640 billion of equity. loans and leases - 0.39 0.82 0.57 0.99 This pushed corporate debt-to-assets ratios to 32 Net loan percent, double the leverage ratios in the early 1950s chargeom3s/ (Frydl 1991). Similarly, household debt increased as net i ehousehold savings declined and consumption SNot available, increased. The ratio of household debt to household Sourceo: U.S. Treasury 1991. assets rose from 38 percent in 1980 to more than 52 sheet and contingent liability transactions. They also percent in 1990. Home mortgages rose by $1.3 tril- increased consumer lending, high-leverage transac- lion. tions, and domestic real estate lending. By the end of For the banks, however, the greatest exposure was 1990 their exposure to domestic real estate had risen in real estate loans. By the end of 1991 real estate to about 36 percent of total loans, compared with 31 accounted for 41.5 percent (by value) of total loans, percent in 1980. Noninterest income also doubled compared with 2.7 percent for highly leveraged trans- between 1982 and 1989 to the equivalent of 15 per- actions. Noncurrent real estate loans amounted to 4.6 cent of interest income, while off-balance sheet activ- percent of total real estate loans, compared with 7.6 ities doubled in dollar terms from 58 percent of bank percent for highly leveraged transaction debt and 12.3 assets to 116 percent. percent for developing country debt (table 4.5). The financial strength of banks has steadily erod- The real estate overhang was serious because ed since the 1930s (table 4.4). Average equity-asset excessive capacity was pervasive. The commercial ratios, for example, declined from 11.9 percent in the office vacancy rate in 1991 reached 18 percent, com- 1930s to 6.1 percent in the 1980s. Regulatory efforts pared with less than 4 percent in 1980, while returns to raise and harmonize international capital adequacy on commercial property turned negative. The failure standards prompted a small increase in the second of large developers such as Olympia and York contin- half of the 1980s. Returns on assets and equity also ued to threaten the profitability of the large banks. declined as net loan chargeoffs rose to the unprece- Although the number of bank failures declined to dented level of 109 percent of net income between 108 in 1991 (table 4.6), the assets of failed banks 1985 and 1989. During 1990-91 U.S. banks provi- jumped to $66 billion, compared with $19 billion in sioned $66 billion on loan losses, equivalent to more 1990. At the same time, the assets of problem banks Table 4.5 Noncurrent loan rates at year-end, 1983-93 (percentage of total assets) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1.96 1.97 1.87 1.94 2.46 2.14 2.30 2.94 3.02 2.54 1.61 Sourse: U.S. FDIC, various issues. The United States: Resolving Systemic Crisis, 1981-91 79 Table 4.6 Problem indicators, 1985-91 Indicator 1985 1986 1987 1988 1989 1990 1991 Unprofitable FDIC-insured commercial banks(asapercentageoftotal) 17.1 19.8 17.7 14.6 12.5 13.4 10.8 Problem commercial and savings banks 1,140 1,484 1,575 1,406 1,109 1,046 1,069 Failed or assisted FDIC-insured commercial banks 118 144 201 221 206 159 108 Assets of failed banks (billions of U.S. dollars) - - 10.1 49.8 30.2 18.7 66.2 Assets of problem commercial and savings banks (billions of U.S. dollars) 237.8 335.5 358.5 352.2 235.5 408.8 611.1 Noncurrent loans and leases plus real estate owned/ total assets of FDIC-insured commercial banks 1.87 1.95 2.46 2.14 2.26 2.90 2.99 - Not available. Sourrce: U.S. FDIC, various issues; Federal Reserve Bank of Kansas City 1992. rose by 50 percent to $611.1 billion, accounting for largest Texan bank holding companies received 17.8 percent of total bank assets. FDIC assistance. Furthermore, in 1988-when 221 The thrifts, for their part, tried to expand out of FDIC-insured banks with assets of $53.7 billion their residential mortgage enclave during the mid- failed-the corporation reported its first-ever operat- 1980s. Deregulation, easy access to funds, and a ing loss. The incidence of large bank failures was ris- federal deposit guarantee encouraged an influx of ing, and bank holding company failures were begin- entrepreneurs to take over ailing thrifts. often with ning to drain the resources of the FDIC. little requirement to commit new capital. The Similarly, the number of thrifts requiring deposit thrifts' ownership structure also changed. By the protection and official resolution increased markedly mid-1980s more than half the thrifts were full- as accumulated credit problems materialized. These fledged, profit-motivated, limited liability compa- failures were qualitatively different from the earlier nies; in 1980, 73 percent had been mutual institu- episode since they could not be attributed to the tions with home ownership as their sole communal interest rate mismatches of the early 1980s. Nearly goal. They advanced aggressively into direct real three-quarters of the 631 official thrift resolutions estate investments-especially commercial and recorded from 1980 through 1989 occurred after industrial real estate-and industrial and commer- 1984 (table 4.8). The FSLIC was virtually insolvent cial loans, in which they had little expertise. Once- by early 1989 and was wound up in August. conservative thrifts invested heavily in junk bonds in search of higher yieids without fullv understand- Response to the Crisis ing the risks involved.'2 Thrift failures were conspic- uously high among the new entrants and where In the face of rising failures, the regulatory agencies aggressive new management abandoned conserva- adopted market-based and innovative measures to tive lending policies. restructure ailing institutions. By mid-decade the ovetall situation among depository institutions was deteriorating rapidly despite continued economic growth. The combina- Table 4.7 Bank failure resolutions, 1980-89 tion of eroding margins, higher-risk lending, regula- Purchase Insured Insured Open tory forbearance, and intense competition among and deposit deposit bank ' Ycar 0~~~~~~~~~~asumption transfir payoff assistance Total banks, thrifts, and other financial institutions began 1980 7 0 3 1 1t to take its toll. Regional economic factors also con- 1981 7 0 2 3 13 tinued to play an important role, especially in oil- 1982 35 0 7 8 50 producing regions as oil prices fell. The FDIC 1983 36 0 9 3 48 1984 62 1 2 4 2 80 reported 322 bank failures during 1980-85; there 1985 87 7 22 4 120 were nearly 800 during 1986-89 (table 4.7). In the 1986 98 19 21 7 145 energy-producing region of the Eleventh Federal 1987 133 40 6I 19 203 1988 164 30 6 21 221 Reserve District, which includes Texas, 265 banks 1989 175 22 9 1 207 failed between 1988 and 1989, nearly one-fifth the Total 805 130 94 69 1,098 number of banks in the district.'3 Nine of the ten Sorce: U.S. Trcasury 1991. 80 BANK RESTRUcruRING Table 4.8 Thrift failure resolutions, 1980-90 than 100 a year in 1984-85, the FSLIC shifted Mergers Management emphasis. and consignment Conservator- The FSLIC tried to keep thrifts open long Year Liquidations acqusitions program, ships Total enough to arrange mergers or acquisitions. Keeping 1980 0 11 0 0 11 some troubled thrifts afloat required regulatory for- 1981 1 27 0 0 28 bearance, including capital augmentation and income 1983 5 31 0 0 36 maintenance techniques as enticements to new own- 1984 9 13 0 0 22 ers.'5 However, these techniques became very costly 1985 9 22 23 0 54 tcnqeeycsl 1986 10 36 29 0 75 because some of the acquisitions required additional 1987 17 30 25 0 72 assistance later.'6 The thrifts were simply too insol- 1988 26 179 18 0 223 vent to recover a fact that was particularly evident 1989 7 1 0 37 47t' v t r 1990 0 0 0 315 315 among some of the resolutions arranged in December Total 85 413 95 352 946 1988. New investors often were not required to put a. Only lasted from 1985-88. up much cash, while the FSLIC filled the net worth b. Includes ten resolutions by the FSLIC before August 9, 1989, when the hole with special financing instruments. Although Financial Institutions Reform, Recovery, and Enforcement Act was passed, although only cight are accounted for under liquidationis and mergers and these arrangements conserved the FSLIC's cash and acquisitions. permitted the new investors to borrow future Sourre: U.S. OTS, various years. against revenue streams, they frequently left the agency with FSLIC and FDIC intervention contingent liabilities. Hurried transactions have been criticized for the competitive advantage they gave to The regulatory response to the thrift crisis, however, thinly capitalized thrifts and for their failure to pro- showed an initial preference for deferring realization mote sound management of acquired assets, which of market value losses. Officials urgently needed to were sweetened with government guarantees covering avoid liquidations or costly restructuring because the yields and capital values (Fogel 1990). insurance fund (FSLIC) had a book net worth of With larger resources, and a relatively lower level only about $6 billion in 1981. This would have of problems in the banking sector, the FDIC had completely evaporated if the fund had continued to authority to provide "open bank assistance" to avert provide financial assistance to failed institutions. On failures. This assistance involved preventive interven- the banking front, however, circumstances were dif- tion before closure, which usually included coopera- ferent. Liquidation of large banks was abjured largely tive assistance from the Federal Reserve in the form on a "too big to fail" basis, exemplified early in the of emergency credits for temporary liquidity purpos- decade when federal authorities took control of es."7 Banks were kept open if the FDIC felt that their Continental Illinois Bank (1984). By contrast, small- operations were essential to the community. Where er bank failures, although regionally disruptive, were failures did occur, FDIC law required bank regulators resolved comfortably with the then-ample resources to choose a method of resolution that was less costly of the FDIC, and generally without liquidation.'4 than liquidating the institution and paying off depos- In general, the FSLIC preferred low-cost or no- itors. However, the method chosen did not necessari- cost interventions that conserved its cash. These ly have to be the least costly remedy. The Garn-St. included the Management Consignment Program, Germain Act relaxed even this requirement and which placed troubled and poorly run thrifts under expanded the authority for open bank assistance. The the management of other institutions or individuals law also empowered the FDIC to permit emergency at no cost to the FSLIC. Introduced in 1985, the interstate acquisitions of failed institutions. program was discontinued in 1988 because franchise Most often, however, the FDIC arranged pur- values tended to deteriorate. Hired management did chases and assumptions, similar to thrift mergers and not have enough financial incentive to restructure the acquisitions (see tables 4.7 and 4.8). Purchases and institutions quickly and adequately for resale. In the assumptions involved cash contributions covering the early years, the Federal Home Loan Bank Board also shortfall of assets over liabilities (less the premium merged groups of insolvent thrifts into new institu- paid by the acquirer). In later years the agency some- tions-"phoenixes"-with new management and times substituted interest-bearing notes. Although provided some recapitalization (Silverberg 1990). But acquirers recapitalized the banks, the FDIC might with the number of thrift insolvencies rising to more contribute some new capital as well. Beginning in The United State!: Resolving Systemic Crisis, 1981-91 81 1986 the fund pursued "whole bank" purchases and agricultural and energy sectors. It extended the scope assumptions more aggressively because "clean bank" of the program in 1987. carve-outs of all bad assets were draining its resources. Technically insolvent thrifts were able to stay in When whole-bank transfers became more difficult to business for years because they could rely on historic arrange, the FDIC introduced "small loan" purchases book-value rather than marked to market, generally and assumptions, whereby acquirers agreed to take a accepted accounting principles. Similarly, developing package of smaller loans, including some nonper- country debt was not marked to market for the forming ones. By the early 1990s, with the increasing money center banks. Thrift regulators made an addi- number of large bank failures, the FDIC was running tional mistake by introducing special regulatory out of resources to undertake even these resolutions. accounting principles. Among their many shortcom- At this point the General Accounting Office, the ings, the principles permitted consistent overvaluation auditing and investigative arm of the U.S. Congress, of delinquent or nonperforming loans and property in recommended establishing a formal early closure pol- depressed real estate markets and greatly enhanced the icy. The agency was keen to reduce the average rate of value of intangible goodwill on balance sheets. Under loss (16 percent of total assets) incurred by banks that these special accounting rules, apparent income could failed between 1985 and 1989. be boosted so much that insolvent thrifts appeared to In March 1990 FDIC assistance to ailing institu- be highly profitable. More than half the failed thrifts tions became subject to three key criteria: all propos- that were resolved in 1988, when the principles were als had to be considered with a competitive bidding suspended, had been accounting insolvent for more process, institutions requesting assistance had to agree than three years, and more than 70 percent had been to unrestricted "due diligence" review by all potential insolvent on a more stringent "tangible" capital basis acquirers cleared by the FDIC, and bidders had to (Benston and Kaufman 1990). establish quantitative limits on guarantees so that the FDIC could assess the cost of each proposal. Regulatory reform Regulatoryforbearance The concept of official "conservatorship" gained sup- port in the 1987 Competitive Equality Banking Act, Outright capital and regulatory forbeara.ice for ailing which empowered the FDIC to establish short-life depository institutions was another important feature "bridge" banks to operate banks already in receiver- of bank restructuring. It supported the efforts to ship. Bridge banks allowed early settlement of losses avoid liquidation, even though resolution costs ulti- against uninsured depositors and creditors, preserved mately rose as a result. Regulators and supervisors not some franchise value (although holding company only allowed forbearance of capital requirements, but creditors and shareholders lost their investments), and relaxed accounting standards as well. In addition, permitted a clean sweep of management (Mires and budgetary cutbacks reduced the number, frequency, Spong 1988). By the end of 1989 the FDIC had and quality of examinations in the earlier years, become sole shareholder in five bridge banks that suc- although vigilance increased toward the end of the cessfully reprivatized formerly unmarketable banks decade. This laxity ultimately increased the scope and (including two large Texas bank holding companies). magnitude of losses. It also made the task of resolving The Financial Institutions Reform, Recovery, and failures more difficult because franchise values deteri- Enforcement Act of 1989 expanded conservatorship orated with the increasing severity of distress. authority for both banks and thrifts to allow reorgani- In the case of thrifts the FSLIC actually reduced zation and restructuring of mostly insolvent institu- net worth requirements from 5 percent, to 4 percent, tions not yet in receivership. Although outright and then to 3 percent in 1982. Beginning in 1983 nationalization is not considered a viable option in the the agency relied increasingly on outright exemp- United States, these techniques place control under tions. Furthermore, the FSLIC used capital augmen- the temporary stewardship of federal authorities and tation instruments such as net worth certificates and in some cases allow a temporary public equity stake income maintenance certificates to boost accounting when institutions are sold to the private sector. capital. In 1986 the FDIC adopted formal capital In late 1989 the U.S. Congress was forced to forbearance measures, mainly to assist the temporari- overhaul the insurance and regulatory system for ly distressed Midwest banks heavily exposed to the thrifts because the FSLIC was bankrupt. The 82 BANK RESTRUCTURING Financial Institutions Reform Act liquidated the in the quality of corporate debt securities they could FSLIC, established a new oversight agency-the purchase. Office of Thrift Supervision-and reorganized the Costly bank failures also were mounting, a prob- deposit insurance fund. The legislation also created lem that only increased with the depth of the reces- the Resolution Trust Corporation (RTC) to handle sion. Significantly, just over half the failures in 1989 imminent thrift failures. The RTC, which was given were among national banks, rather than state- six years to clean up the thrift industry mess, became chartered ones, and nearly all of those were in the an interim deposit guarantor. To finance the bailout, Southwest (65 percent in Texas alone). The 12,593 the RTC could borrow $20 billion through general FDIC-insured banks earned $1 billion less in the first Treasury bond financing and contributions from fed- quarter of 1990 than they had in the first quarter of eral home loan banks. It could also issue $30 billion 1989. This drop resulted mainly from higher provi- of special Resolution Finance Corporation bonds, sioning for domestic credit losses (U.S. FDIC Annual using zero-coupon Treasury bonds as collateral. Report 1990). Net loan chargeoffs were nearly 50 per- Congress optimistically expected the RTC to use cent greater than a year earlier, and the ratio of trou- the proceeds of rapid asset disposition to cover the bled assets continued. Aggregate performance was underwriting of new thrift conservatorships and the affected disproportionately by banks in the warehousing of the bad assets it acquired. Income Northeast, where Mexican loan chargeoffs and real from future thrift resolutions and bond sales were estate loan losses took a heavy toll. As the economy also supposed to support the RTC's program. slipped into recession at the end of 1990, there were Between August 1989 and the end of 1990, 531 1,012 problem banks (10 percent of the total), and thrifts, with $271 billion in assets, failed. At the end II percent of all banks--including 20 percent of the of April 1991 the RTC still had 219 thrifts in conser- largest with assets in excess of $10 billion each-were vatorship and 375 in receivership, with more likely to unprofitable. follow. Through May 1991 the RTC reported total On November 27, 1991, Congress passed the estimated losses of $55 billion incurred from the dis- FDIC Improvement Act, a compromise that posal of 396 insolvent thrifts, for which asset sales addressed few of the reforms the administration and collections totaled $163 billion. requested. The act included the following provisions: With the onset of a recession in 1990 and a deep * Authority for the FDIC to borrow $70 billion slump in real estate markets throughout much of the from the Treasury to replenish the Bank country, failures were costing more and resolutions Insurance Fund. and asset disposition were harder to arrange. The * Rules enabling regulators to act quickly when a RTC returned to Congress in 1990 for more fund- bank's core capital falls below 2 percent of risk ing, finally receiving authorization for $78 billion in assets and to replace management and limit the early 1991. By the end of 1991 the RTC had seized asset growth of "critically undercapitalized" banks. $357 billion in assets, had sold and recovered $228 * Rules requiring ailing banks to suspend dividend billion, and had remaining assets of $129 billion to payments. be sold. * Restrictions on the Federal Reserve's ability to The Financial Institutions Reform Act also re- provide credit to ailing banks. regulated in some important areas, and it increased * Some rollback of deposit insurance coverage on insurance premiums for both thrifts and commercial brokered deposits and deposits of U.S. banks banks to generate additional billions for the Savings abroad. Association Insurance Fund and the Bank Insurance * Requirement that foreign banks taking insured Fund by the mid-1990s. The minimum tangible deposits be locally incorporated. capital standard for thrifts was set at 3 percent of * Reduction of the number of state-chartered banks total assets, a standard that was achieved by the end authorized to underwrite and sell insurance. of 1994. Also, thrifts were directed to increase their The legislation failed to address the Treasury's pro- level of housing assets, and the ratio of commercial posals to allow banks to underwrite securities and real estate loans to capital was restricted. State-char- thereby remove the functional segmentation between tered thrifts were limited to the same type and investment banking and commercial banking estab- amount of real estate equity investment as federally lished by the Glass-Steagall Act (The Economist, 30 chartered ones, and thrifts in general were restricted November 1991). The United States: Resolving Systemic Cisis, 1981-91 83 Conclusion credit risks can be "reduced" to an interest rate risk since the higher the credit risks, the lower (or even When the Financial Institutions Reform Act was first negative) the interest rate yield on bank loans, which proposed in February 1989, the administration reduces the value of bank loans. Opponents of mar- claimed that the thrift cleanup would cost the taxpay- ket-value accounting argue that the uncertainties (and ers $60 billion."8 A subsequent independent estimate errors) of measuring such risk are high, and therefore of deposit fees and taxes for 1988-99 placed the fig- time should be given to allow more accurate assess- ure at $150 billion, or about 3 percent of gross ment of such valuation. The weight of evidence defi- domestic product for 1988 (Hill 1990). Failed bank nitely points toward market-value accounting, howev- resolutions cost the industry, through the FDIC, a er, as financial markets converge and price volatility mere $17 billion during 1985-89, although large increases. A clear case for market-value accounting is outlays continued in 1990 and 1991. The 1991 the need for higher capital for financial institutions to FDIC Improvement Act provided the FDIC with cushion against risks, not whether such risks are mea- another $70 billion in borrowing powers to deal with sured accurately under such accounting. the problem banks. On the positive side, public attention has focused Apart from these costs, there were long-run mar- on the fundamental structural problems of the U.S. ket consequences as well. Delays in the resolution of system. Restrictions on geographic and functional failing financial institutions slowed the pace of asset diversification have been eased, if not eliminated. disposition and shifted risk-first from individual Industry consolidation continues: since 1985, 5,400 institutions to the industry through the mechanism of banks, thrifts, and credit unions have been absorbed the insurance fund, and then to the government. In or eliminated. Roughly 10 percent of thrifts account addition, selective sale of interstate acquisition privi- for about 70 percent of total industry assets, whereas leges in a structurally segmented system transferred 11 percent held 60 percent of the total in 1979. costs to competing institutions. Regulatory forbear- Similarly, 70 percent of commercial bank assets are ance prevented normal market shrinkage by inhibiting held by less than 3 percent of banks, and the trend of exit. Solvent and comparatively healthy institutions mergers of large money center banks has continued, were forced to compete with undercapitalized or vir- the most recent being the merger of Chase tually insolvent ones that could afford to price liabili- Manhattan and Chemical Bank. ties and assets unrealistically. Tax benefits to acquirers The U.S. case demonstrates the difficulties of shrank the pool of prospective buyers to those to resolving failures in a system consisting of a large whom they were an attractive incentive. The extent of number of geographically and functionally segment- deposit insurance encouraged moral hazard behavior ed financial institutions. While a federally guaranteed by depositors, managers, and directors, and effectively and industry-financed deposit insurance scheme was substituted public capital for private capital. by and large successful for more than fifty years in The U.S. case raises important issues regarding preventing bank runs, the large losses of recent years the timing of loss recognition and failure resolution. show that quasi-public capital (in the form of the Proponents of market-value accounting, the federal deposit guarantee) is not a substitute for Securities and Exchange Commission among them, sound private capital. argue that marking the value of assets to market Clearly, after a long history of growth, banking forces discipline on financial institutions because loss- capital in the United States became severely eroded es are recognized more promptly, thereby ensuring relative to the risks and the competition from new prompt management and regulatory action and ulti- players. The franchise for low-cost deposits in operat- mately reducing the cost of failure resolution. ing the payments mechanism was eroded by the inva- Historic accounting, by contrast, gives banks more sion of credit cards and the evolution of deposit sub- time to recognize losses on bonds and loans, since it stitutes from mutual funds, insurance companies, can always be argued that markets for bank loans are and securities houses. There was also greater competi- not perfect and the valuation of bank loans should be tion from foreign banks in the U.S. market. The judgmental. banks and thrifts responded by taking higher risks, This debate is particularly interesting because including derivatives and new, off-balance sheet bank balance sheets are clearly subject to both interest activities that incurred higher costs and losses when rate risks and credit risks, which are interrelated. Even the economy began to slow down. 84 BANK RESTRucruRING Several observations can be made: during the 1970s, Regulation Q proved a serious constraint on * Even the most sophisticated bank regulation and both types of institutions, resulting in disintermediation. examination did not prevent large-scale bank fail- 2. For example, between 1983 and 1985, when many Texas Banks are s to structural flaws and banks were building up bad assets, bank examinations were ures. BanKs are subject to structural ttaws ana halved (U.S. FDIC Annual Report 1990). cyclical changes in the economy. If tax incentives 3. The Office of Thrift Supervision was created in 1989 push the economy toward real estate, banks are under the Financial Institutions Reform, Recovery, and likely to make mistakes in both industrial and Enforcement Act to replace the Federal Home Loan Bank Board, developing countries, which had been responsible for national chartering, regulation, * Bank losses are quasi-fiscal deficits, even in highly and oversight. market-based economies. Even though 4. Although the Federal Reserve is the primary regulator of private, market-based economies. Even though bank holding companies, their subsidiary banks are usually the U.S. banking system remains almost entirely supervised by the Office of the Comptroller of the Currency or in private hands, a large percentage of bank the FDIC. losses was financed by budget appropriations or 5. The FSLIC was financed by the Federal Home Loan Bank borrowings by the insurance funds under System, thus linking the insurance and oversight functions. The Federal Home Loan Bank Board had an inherent conflict of Treasury guarantee. . objectives since it was both a regulator of thrifts and a promoter * Inadequate resources and powers postponed time- of the housing industry and widespread home ownership. ly decisionmaking, and in the end the state suf- 6. By 1990 the FDIC covered 75 percent of deposits in fered from the inefficient actions of an insolvent insured banks, compared with 65 percent in 1970 and only 45 rescuer (FSLIC) trying to rescue insolvent inter- percent in 1934. mediaries. 7. Specialist brokers arranged deposits nationwide for thrifts for a fee. Deposits were broken down into insurable amounts so * Ill-timed and ill-conceived deregulation con- that they could be covered by FDIC's $100,000 limit per tributed substantially to the problems or exacer- account. "Pass through" coverage brought pension and insurance bated difficulties precipitated by changes in busi- funds into the deposit network as well. ness cycles. The relaxation of supervisory vigi- 8. Mutual savings banks were insured by the FDIC bur had lance and sound bank accounting principles con- interest risks similar to the thrifts and suffered the same conse- ibuted to the accumulation of losses greater quences of tight monetary policy. Failures among a few state-char- tri tered mutual savings banks in the Northeast following a property than those initially reported. price slump tapped FDIC resources, but their loan loss and net From these observations, two lessons can be drawn: worth problems proved more short-lived than those of the thrifts. * During economic liberalization, anticompetitive Fourteen mutual savings banks with $17.4 billion in assets regulations should be removed to promote com- received FDIC assistance between 1978 and 1983. petition, but regulations against conflicts of inter- 9. Banks in a twelve-state, predominantly agricultural area petition, accounted for more than half of bank failures in 1985, following est, fraud, and moral hazard should be strictly the collapse of farm land prices and widespread foredosures, but enforced. these Midwest farm banks rebounded in 1987 (Hefferman 1988). * The costs of delayed action demonstrate that the 10. California had already allowed its state-chartered thrifts sooner action is taken, the better. To some, delays this privilege. This new freedom had an immediate positive in action reflected the gridlock of overlapping impact on the structure of thrift balance sheets and consequently in action reflected the gridlock oonoearnings. supervisory authorities and the absence of political on earnings. I1. These included lower capital and liquidity requirements, will to confront the issues. But in the U.S. case, liberal lending limits on real estate loans, deferrals on loan losses, once the will was manifested, considerable national large tax-free transfers to loan loss reserves, special trearment of resources were harnessed, backed by professional tax loss carryforwards, revaluation of premises and other proper- skills, to deal with the problems. Few developing ry, recording mutual fund assets at cost rather than at marker ,i.n.g. value, liberal valuation of intangible assets, less frequent use of countries have either the will or the capacity to deal vu,lbrlvlaino nagbeast,ls rqetueo countries have either the will or the capacitytodeal enforcement action, and ready access to Federal Home Loan with banking problems in this manner. Bank advances (ABA 1988). 12. By) mid-1990 federal rescuers held a portfolio of $3.7 Notes billion in junk bonds acquired from failed thrifts. 13. The banks in the district lost $1.0 billion in 1986, $2.7 1. Regulation Q interest rate ceilings for deposits in commercial billion in 1 987, and $2.2 billion in 1988. banks were introduced in the 1930s, and were imposed on thrifts 14. Only eighty-five thrifts were liquidated berween 1980 in the mid-1960s to dampen competition among them. The and 1989, nearly half of those in 1987 and 1988. Similarly, the deposit rate ceiling for thrifts, however, was set at a margin above FDIC paid off depositors in only 244 cases between 1971 and the level for commercial banks to preserve thrifts' competitive 1989, and 94 of those cases occurred during 1980-89. In the advantage in collecting small savings deposits. As inflation rose entire history of the FDIC only $2.8 billion had been spent for The United States: Resolving Systemic Crisis, 1981-91 85 this purpose. In a limited number of bank and thrift cases, and Regulatory Policy." Federal Reserve Bank of New York insured deposit transfers were used when depositor payouts were Economic Policy Review (July): 27-42. imminent but a healthy institution was willing to pay a premi- Federal Reserve Bank of Kansas City. 1990. Banking Regulation, um for the insured deposits, and perhaps some assets. Its Purposes, Implementation, and Effects. 15. The Garn-St. Germain Act authorized net worth cer- . 1992. "Commercial Bank Performance, 1991." tificates, issued by thrifts in exchange for FSLIC promissory Financial Industry Trends. notes, which were used to boost thrift capital to the minimum Felgran, Steven D. 1988. "Bank Participation in Real Estate: requirement. Conduct, Risk, and Regulation." New England Economic 16. Thrift-assisted mergers involved some, but not necessar- Review (November-December): 57-73. ily all, of the following additional measures: capital loss prorec- Fogel. Richard L. 1990. "Failed Thrifts: Resolution Trust tion of a negotiated "covered" portion of assets; yield guarantee Corporation and 1988 Bank Board Resolution Actions." subsidies or income maintenance agreements, which shifted Statement of the Comptroller General, General interest rate risk to the FSLIC; and indemnification for legal Government Programs, before the Committee on Banking, expenses incurred from lawsuits or contingencies arising from Finance, and Urban Affairs, U.S. House of Representatives, the failed thrift. They also might include loss-sharing of asset April 2, Washington. D.C. disposition, gain sharing, tax allowances, buyout options allow- Frydl. Edward J. 1991. "Overhangs and Hangovers: Coping ing the FSLIC to purchase covered assets, warrants for shares in with the Imbalances of the 1980s." Annual Report. Federal assisted thrifts, and marked to market payment for covered assets Reserve Bank of New York. or for goodwill booked for covered assets (Seidman 1990). Gregorash, George, Eileen Maloney, and Don Wilson. 1987. 17. The FDIC's open bank assistance programs incurred "Crosscurrents in 1986 Bank Performance." Federal losses of $11.2 billion from inception in 1933 to the end of Reserve Bank of Chicago Economic Perspectives (May/June): 1988. The greatest costs were in the 1980s, with cases like 23-35. Continental Illinois (1984), BankTexas Group (1987), and First Harris, Ethan S., and Charles Steindel. 1991. "The Decline in City Bancorporation (1988). U.S. Savings and its Implications for Economic Growth." 18. By the time the legislation passed in August, the initial Federal Reserve Bank of New York Quarterly Review 15 program provided $166 billion over ten years to resolve 191 (Winter): 1-19. thrifts, at a direct nominal expense to taxpayers of $90 billion. Hefferman, Peter J. 1988. "Nothing is Forever: Boom and Bust In April 1990, however, the General Accounting Office estimat- in Midwest Farming." Federal Reserve Bank of Chicago ed that the projected nominal budget outlays would amount to Economic Perspectives (September/October). at least $315 billion (U.S. GAO 1990). This estimate subse- Hill, Edward W 1990. "The S&L Bailout: Some States Gain, quently rose to $400 to $500 billion. Many More Lose" Challenge 33 (May/June): 37-45. Holrhus, C.G. 1989. Statement of the President-Elect of the References American Bankers Association before the Subcommittee on Financial Institutions Supervision, Regulation, and ABA (American Bankers Association). 1988. "The FSLIC Crisis: Insurance of the Committee on Banking, Finance, and Principles and Issues." Washington, D.C. Urban Affairs, U.S. House of Representatives, September Amel, Dean F. 1989. "Trends in Banking Structure Since the 20, Washington, D.C. Mid-1970s." Federal Reserve Bulletin 75 (March): 120-33. IMF (International Monetary Fund). Various years. Barth, James R., Philip F. Bartholomew, and Michael G. International Financial Statistics. Washington, D.C. Bradley. 1990. "Determinants of Thrift Institution Johnson, Manuel H. 1989. Testimony of the Vice Chairman, Resolution Costs." Journal of Finance 45 (July): 731-54. Board of Governors of the Federal Reserve System, before Benston, George J.. and George G. Kaufman. 1990. the Subcommittee on Financial Institutions Supervision, "Understanding the Savings-and-Loan Debacle." Public Regulation, and Insurance of the Committee on Banking, Interest 99 (Spring): 79-95. Finance, and Urban Affairs, U.S. House of Representatives, Brumbaugh, R. Dan, and Robert E. Litan. 1989. Joint testimo- September 19, Washington, D.C. ny before the Subcommittee on Financial Institutions Kaufman, George. 1990. "Crisis Spreads: Make FDIC Insurance Supervision, Regulation, and Insurance of the Committee Redundant." Challenge 33 (January/February): 13-17. on Banking, Finance, and Urban Affairs, U.S. House of Melichar, Emanuel. 1986. "Agricultural Banks Under Stress." Representatives, September 19, Washington, D.C. Federal Reserve Bulletin 72 (July): 437-48. Brumbaugh, R. Dan, Andrew S. Carron, and Robert E. Litan. Miller, Randall. 1991. "Testimony on Mergers." U.S. Office of 1989. "Cleaning Up the Depository Institutions Mess." the Comptroller of the Currency Bulletin (December). Brookings Papers on Economic Activity 1. Washington, D.C. Washington, D.C. Corrigan, Gerald. 1989. "A Perspective on Recent Financial Mires, Ralph E., and Kenneth Spong. 1988. "The Death of a Disruptions." Federal Reserve Bank of New York Quarterly Bank: Assuring an Orderly Transition." Banking Studies Review 14 (Winter): 8-15. special issue, "Problem Banks," Federal Reserve Bank of - 1990. "Reforming the U.S. Financial System: An Kansas City. International Perspective." Federal Reserve Bank of New York Seidman, L. William. 1990. Testimony of the Chairman, FDIC, Quarterly Review (Spring). on FSLIC Resolution Fund Appropriations for Fiscal 1991, Edwards, F. R., and F. S. Mishkin. 1995. "The Decline of before the Subcommittee on Housing and Urban Traditional Banking: Implications for Financial Stability Development, Veterans Administrations, and independent 86 BANK RESTRUCrURING agencies of the Committee on Appropriations, U.S. Senate, . 1991. "Deposit Insurance and Strategy for Reform." May 23, Washington, D.C. Report GGD-91-26. Washington, D.C. Silverberg, Stanley C. 1989. "U.S. Banking Problems in the U.S. OCC (Office of the Comptroller of the Currency). 1988. 1980s." World Bank, Financial Policy and Systems Bank Failure:An Evaluation of the Factors Contributing to the Division, Washington, D.C. Failure of National Banks. Washington, D.C.: U.S. Treasury. . 1990. "The Savings and Loan Problem in the United e 1990. "Condition of the Banking System 1989." U.S. States." Policy Research Working Paper 351. World Bank, Treasury, Division of Jndustry and Financial Analysis, Washington, D.C. Washington, D.C. Spong, Kenneth. 1988. "Assistance for Problem Banks." Banking U.S. OTS (Office of Thrift Supervision). Various years. Annual Studies special issue, "Problem Banks," Federal Reserve Report. Washington, D.C. Bank of Kansas City. U.S. RTC (Resolution Trust Corporation). 1990. U.S. FDIC (Federal Deposit Insurance Corporation). Various "Conservatorship Institutions Post Slight Decline in First years. Annual Report. Washington, D.C. Quarter Operating Losses." Press release. Washington, D.C. Various issues. Quarterly Banking Profile. Washington, . 1991. RTC Review (April). D.C. U.S. Treasury. 1991. Modernizing the Financial System: U.S. GAO (General Accounting Office). 1987. "Thrift Industry Recommendations for Safer, More Competitive Banks. Forbearance for Troubled Institutions, 1982-86." Washington, D.C. Washington, D.C. White, Lawrence J. 1990. "Problems of the FSLIC: A Former - 1990. Statement by the Comptroller General of the Policy Maker's View.' Contemporary Policy Issues 8 (April). United States before the Committee on Banking, Housing, . 1991. The S&-L Debacle: Public Policy Lessons for Bank and Urban Affairs, U.S. Senate, April 6, Washington, D.C. and Thrift Regulation. New York: Oxford University Press. CHAPTER 5 Bank Restructuring in Spain, 1977-85 Andrew Sheng The Spanish banking crisis was the most serious to percent) was borne by either the Deposit Guarantee hit the OECD countries after the U.K. secondary Fund or the Bank of Spain (Cuervo 1988). banking crisis of 1974-76. The crisis spanned As in other cases, Spain's banking crisis followed a 1977-85 and affected 52 of the country's 110 banks; period of strong economic growth in the 1960s and most were small and medium-size firms that together early 1970s. The death of General Francisco Franco accounted for more than 20 percent of total deposits. in 1975, however, unleashed several years of social, Twenry banking institutions related to the Rumasa political, and economic transformation. The transi- industrial conglomerate had to be resolved through a tion to democracy did not permit decisive adjust- special nationalization and reprivatization program ment to the oil shocks. Between 1975 and 1979 the introduced in 1983. The rest were resolved through peseta appreciated by more than 35 percent in real the Deposit Guarantee Fund, and most were terms, the stock of external debt doubled, and real absorbed by domestic and foreign banks. Twenty-one wages increased by 9 percent. Although the banking of the non-Rumasa banks had to be recapitalized by crisis peaked in 1982 and 1983, losses had been accu- the fund before they could be resold, and only three mulating throughout the 1970s and early 1980s, small banks were actually liquidated. In only a few exacerbated by the impact of rising oil prices, double- cases did private banks absorb troubled institutions digit inflation, and rigidities in the labor market and without extensive official intervention. The seven productive sectors, which undermined the profitabili- largest Spanish banks survived with limited damage ty and solvency of enterprises. and played an important role in absorbing a number Problems in the banking system began to surface of the failed institutions. in 1977 when the Bank of Spain noticed that rising As with other banking crises, Spain's banking liquidity problems in some cases hid much larger sol- problems resulted from a combination of the oil vency problems. One of the most striking features of shock of 1973-74, inappropriate policy responses to the crisis was the extent to which failed banks had these shocks, and rapid liberalization and expansion of exceeded permissible levels of loan concentration to the banking sector without adequate regulation and related parties and risk concentration to single per- supervision. Institutional factors such as the extensive sons or entities. The initial response to the emerging ownership of banks by industrial conglomerates and problems was slow, partly because the legal frame- bad bank management were also responsible. The cri- work for banking was antiquated. The central bank sis was cosdy to resolve. For example, the government initiallv had neither the legal power to intervene nor expected to be repaid only 25 percent of a 400 billion adequate powers to impose and enforce sanctions peseta ($2.5 billion) interest-free loan that it extended against wrongdoers. The Deposit Guarantee Fund to the Rumasa holding company as part of the plan to was established in 1977, but this too proved inade- salvage the Rumasa banking group (de Juan 1985). quate to deal with the unfolding crisis. One estimate placed the net cost of the crisis at 1,581 The Banking Corporation was created in 1978 to billion pesetas (in constant 1985 prices), or 5.6 per- take control of and, where possible, reorganize trou- cent of 1985 GDP, of which 1,216 billion pesetas (77 bled institutions. Financed half by private banks and 87 88 BANK RESTRUCrURING half by the Bank of Spain, the corporation lacked the shocks of 1973 and 1979. Although the country legal means to recapitalize insolvent institutions and enjoyed one of the highest growth rates in Europe in proved an inadequate mechanism for dealing with the early 1970s-annual GDP growth peaked at impending crises. As a result the Deposit Guarantee nearly 8 percent in 1972-real growth fell to 0.5 per- Fund was reconstituted as a legal entity in 1980. The cent in 1975 and the economy entered a period of new fund, still jointly financed, provided not only stagnation during which growth did not exceed 2 managerial and administrative oversight of failed banks percent a year. Reaction to the first oil shock was but also direct recapitalization. The amount of deposit delayed by the political transition following Franco's insurance coverage was also increased. By law fund death. Real wages were allowed to rise to avert social ownership had to be divested within one year of unrest, and the real exchange rate appreciated by takeover, promoting prompt resolution and resale of more than 35 percent between 1975 and 1979. As five of the fund-controlled banks to foreign institu- fiscal deficits began to widen, inflation exceeded 24 tions and twenty-two to domestic banks by late 1985. percent in 1977 and did not fall below double digits This process inaugurated a prolonged period of struc- until 1985 (table 5.1). tural reform in Spanish banking that continued into The dramatic changes in the economy coincided the 1990s. The fund has become an important model with a period of liberalization and rapid expansion in for bank restructuring in many other countries because the banking sector, which financed the growing of its flexibility, efficiency, fairness, and realistic indebtedness of enterprises. Spanish businesses, approach to protecting the public interest in the buoyed by the steady growth of the 1960s, faced banking system. major structural adjustments in the 1970s, with ris- ing energy costs, increasing competition, and rigidi- Structural Adjustment and the Financial System ties in the labor market structure impeding their abil- ity to reduce real labor costs. In addition, enterprises In the 1960s Spain embarked on a series of indicative began to face higher financial intermediation costs as development plans along the French model, with a a result of tighter monetary policies introduced to heavy emphasis on industrialization. This energy- curb inflation. Higher real interest rates, declining intensive investment was seriously affected by the oil liquidity, and slower growth all led to a deterioration Table 5.1 Macroeconomic indicators, 1970-89 (percent unless otherwise specified) General Exchange Realgrowth Current Terms oftrad Inflation governmentfinancial Bank nominal rate Year in GDP account/GDP (1980=100)' rateb balance/GDP interest rate' (pesetas/U.S dollar) 1970 4.1 2.2 153.1 5.7 0.7 n.a. 70.0 1971 4.6 2.0 159.8 8.2 -0.6 n.a. 69.4 1972 8.0 1.1 155.8 8.3 -0.2 n.a. 64.2 1973 7.7 0.8 146.4 11.4 0.8 n.a. 58.2 1974 5.1 -3.6 108.2 15.7 -0.3 n.a. 57.6 1975 0.5 -3.3 117.9 16.9 -0.5 n.a. 57.4 1976 3.3 -3.9 114.0 14.9 -1.0 n.a. 66.9 1977 3.0 -1.6 108.0 24.5 -1.3 14.0 75.9 1978 1.4 1.1 114.9 19.8 -2.3 15.0 76.6 1979 -0.1 0.6 114.7 15.7 -2.1 15.7 67.1 1980 1.2 -2.4 100.0 15.6 -2.5 17.1 71.7 1981 -0.2 -2.6 85.1 14.5 -3.8 16.7 92.3 1982 1.2 -2.4 91.3 14.4 -5.4 16.4 109.8 1983 1.8 -1.6 86.8 12.2 -4.6 16.7 143.4 1984 1.8 1.3 89.3 11.3 -5.3 16.9 160. 1985 2.3 1.7 89.8 8.8 -6.9 15.8 170.0 1986 3.3 1.7 104.6 8.8 -6.0 14.6 140.0 1987 5.5 0.0 98.3 5.3 -3.2 15.7 123.4 1988 5.0 -1.1 103.0 4.8 -3.3 15.3 116.4 1989 4.8 -2.9 105.5 6.8 -2.8 16.3 118.3 a. Ratio of the index of avcrage export prices to the index of average import prices. b. Consumer price index. c. Does not include regulated or special rates for privatc banks. Rates are averages of selectcd rates for commercial discounts, credits, and loans. Source: IMF and World Bank data; Bank of Spain. Bank Rs rructaring in Spain, 1977-5 89 in stock market values, putting heavy pressure on col- in light of the transition to freer markets. These lateral values and on banks holding large portfolios of shortcomings were particularly unfortunate given the corporate stock. growth of large industrial conglomerates (in the case Until the mid-1970s Spain's banking sector was of Runiasa, within a nonbank holding company highly profitable, benefiting from a protccted market structure) and the large number of related nonbank and a rapidly growing economy that averaged 7 per- financial companies created to sustain the ownership cent annual growth during 1961-74. Interest rates structure of new or acquired banks. The complex of on loans and deposits were regulated, market entry related companies together with creative accounting was severely restricted (and foreign banks were entire- techniques obscured the real consolidated financial ly excluded), and chartering of new branches was position of banks and their customers from bank heavily constrained until 1977. Moreover, domestic examiners, as bank managers struggled to survive in banks were forced to allocate substantial funds to pri- the new competitive environment. ority investments; as late as 1985, 48 percent of Under these economic, financial, and regulatory banks' resources were still regulated. Commercial and conditions, liberalization had a grave impact on the industrial banks were functionally separated. viability of many banks because of the timing and Under pressure from the Bank of Spain, important manner of bank competition and expansion it gener- reforms occurred in 1969, 1974, 1977, and 1981 that ated. Specialization gave way to multipurpose bank- aimed to open the system to greater competition. In ing. Although the banking sector became more 1969 interest rates were freed for foreign deposits, sophisticated and innovative, many new entrants interbank operations, loans of more than three years, lacked professional experience, adequate financial and domestic liabilities with maturities of more than backing, and in some cases the ethical standards two years. Reforms in 1974, 1977, and 1981 rational- expected of sound bankers. None of the banks creat- ized the rate structure until all rates were free except ed during the financial liberalization of the 1970s those on liabilities of less than a year (or less than six survived as independent institutions, and 90 percent months for deposits over $6,500 equivalent). This led of the banks involved in the banking crisis were creat- to competitive innovation as new financial instru- ed between 1973 and 1978. ments were introduced. After the 1971 and 1974 There were three types of unstable banks: self- reforms it was easier to open new banks and new financing (those with an artificial structure that had branches. In 1977 foreign banks were admitted. concentrated risk primarily in finance for the acquisi- Although they were restricted to having three offices tion of their own assets); medium-size banks that had nationwide, foreign banks brought new instruments extended an imprudent concentration of risk to their to the market. Bank specialization was abandoned in controlling groups to finance speculative investments, 1977, and gradual reduction of forced investment often in related companies; and larger institutions began. At the same time the creation of a money mar- hurt by the impact of persistent economic crisis on ket led the Bank of Spain to rely more heavily on open their enterprise borrowers (Alvarez Rendueles 1983). market operations-rather than reserve requirements Industrial banks, for example, had greater investment and rediscount facilities-to facilitate and broaden freedom than savings or commercial banks and built monetary control. The money market expanded their asset portfolios by raising high-cost funds.' beyond a simple interbank market to include non- They prospered during the period of rapid economic banking institutions, many of them private. growth but were unable to disengage quickly from Banking regulation, however, remained outmod- their narrow base of lending to highly indebted ed, focusing on compliance with legal requirements industries when prosperity ended. and technical ratios-not uncommon where banking The era also saw expansion and growth through systems have been repressed. Moreover, until the the formation of large industrial bank groups. mid-1980s most regulatory and disciplinary power Existing banks often were purchased with only a small rested with the Ministry of Finance rather than with down payment; the balance was covered by debt and the Bank of Spain. In addition, the 1946 Banking in some cases banks were bought with credits extend- Law (which remains the principal banking legisla- ed by the acquired bank itself. The relatively inexpen- tion) and the penal code did not provide an appro- sive deposit base was used for speculative investment priate basis for action against wrongdoers. in banks whose purchase prices were often highly Information disclosure was also inadequate, especially inflated. In the hands of the new owners banks 90 BANK RESTRUCTURING directed credit to related companies and to sharehold- cent by 1984, clearly exacerbating the liquidity prob- er companies and individuals to have access to funds lems of enterprises. Nonperforming enterprise loans to repay their debts, including those incurred to pur- accumulated, followed by distress borrowing. chase the bank in the first place. Complex networks of Enterprises consumed their capital, leading to wide- related financial entities developed through intricate spread business failures that had a direct impact on lending schemes and cross-ownerships, sometimes bank solvency. In a now-familiar vicious circle, banks with other banks. Lending limits to single borrowers continued to raise interest rates in order to increase could be evaded easily. margins. They had to compensate for falling income Such growth could be sustained only with further and the high level of forced investment, and to gener- speculation and higher borrowing. Nonbank compa- ate surpluses for increasing their loan loss provisions. nies were formed to sustain the quoted prices of Predictably, higher financial costs caused even worse banking companies in the hope that stock market problems for enterprises. appreciation would vindicate the valuation of the original bank acquisition. The sharp decline of the Response to the Crisis Spanish stock market during 1974-80 led banks and their related or controlling groups to turn to real The initial response to the growing crisis was a simple estate investment. Speculation in real estate, initially deposit insurance scheme-the Deposit Guarantee fueled by growing tourism, was expected to produce Fund, introduced in November 1977. The fund's earnings sufficient to cover the interest owed to banks principal objective was to insure small deposits and on the loans that they had granted to related entities, avoid the threat of massive deposit withdrawals that as well as other loans whose investment returns dete- could spark a systemwide panic. The scheme provid- riorated in the depressed economic climate. When ed deposit insurance of 500,000 pesetas (nearly the oversupply retarded property prices the banks $10,000) per depositor. Administered as an account were forced ro renew loans and capitalize interest within the Bank of Spain, the scheme was funded arrears to the borrowers to ensure their own survival. equally by the central bank and by participating com- Banks that were connected to industrial concerns mercial banks, who paid a one-time premium through cross-holdings were dragged into further bad amounting to I peseta per 1,000 pesetas of deposit credits. "Good" bankers became "bad" bankers to liabilities. Because the fund could only be used where cover their own mistakes (de Juan 1987). banks had been closed and were going to be liquidat- This unsound competitive expansion occurred in ed, it was of little real assistance in resolving problem a market environment in which many retail deposit banks. A mechanism was needed for facilitating rates were still regulated, so that competition restructuring and resale rather than closure. occurred mainly over market share-through aggres- Moreover, as the scale of bank problems mounted the sive branching-and, where possible, by bidding up inadequacy of the insurance guarantees became clear. interest rates on wholesale deposits. The number of The limit was raised first to 750,000 pesetas per bank offices more than tripled between 1973 and depositor in 1980, then to 1.5 million pesetas in 1983, while the number of people employed in bank- 1981. ing rose from 150,000 in 1975 to more than To improve flexibility in handling individual 180,000 by 1980. Financial intermediation costs rose bank crises, the Banking Corporation was established as a result. Net interest margins, for example, rose in March 1978 to supplement the deposit insurance from 3.45 percent in 1974 to 4.45 percent in 1978 function of the Deposit Guarantee Fund. Equally and had only fallen to 4.33 percent by 1980 (Cuervo funded by private banks and the Bank of Spain, the 1988). Operating expenses, equivalent to 2.55 per- corporation was empowered to take over and tem- cent of average domestic liabilities in 1974, had risen porarily administer insolvent banks (sometimes to 3.53 percent by 1979. through ownership) but not to recapitalize them. Higher margins coincided with higher real inter- Once a bank confronted a crisis, the Bank of Spain est rates at a time when falling sales margins and the would restrict and apply conditions to its liquidity indebted capital structure of industry led to liquidity support. Under pressure, controlling shareholders and debt-servicing difficulties. Although banks' real either had to recapitalize the bank or pass control to lending rates were negative or flat until 1980, they the Banking Corporation at a nominal price. The had reached a positive real rate of more than 5 per- corporation then took over administration of the Bank Restructurwing in Spain, 1977-85 91 bank, began to rationalize operations, and searched ized, mixed (public-private), autonomous institution for a buyer. The corporation's main purpose was to with a clearly delineated mandate throughout this restore banks' viability. Even though bank operations consolidation and contraction. In the Rumasa case, and management improved under this system, the however, the fund was used only for those purposes corporation faced a serious dilemma because it could for which it had a comparative advantage, namely not inject new capital in the institutions, most of provisional administration and disposition. The capi- which had lost their entire capital at least once. tal deficiencies of the Rumasa group were handled Forcing banks to the brink of closure without the through a comprehensive nationalization and repriva- means to restore solvency undercut the corporation's tization effort. In the end the fund participated ability to sell them off again. directly or indirectly in the rescue of twenty-nine The Bank of Spain began to tighten supervision banks during 1978-85 (table 5.2) and in the admin- and control during this period. It devoted more istration and final sale of twenty banks previously attention to loan and investment analysis and to asset owned by the Rumasa group. quality. In addition, the bank implemented an early An eight-member board of directors oversaw the warning system that, among other things, forecast find's operations. The board consisted of four private income and expenditures for each bank and detected bankers appointed by the Ministry of Finance and deviations. There were important improvements in four representatives from the Bank of Spain. The vice information disclosure as well. governor of the Bank of Spain chaired the board, vot- ing only in the case of a tie. Among its duties and The Deposit Guarantee Fund responsibilities, the board could request external audits of member banks, including their holding and In 1980 the Deposit Guarantee Fund was restruc- related companies. This structure-combined with tured and its powers substantially increased. A royal the close relationship of the fund and the Bank of decree redefined the fund's objective by empowering Spain-mitigated the likelihood of bureaucratization it to "carry out all the operations deemed necessary to and politicization and fostered a public perception of reinforce the solvency and operation of the banks, in fairness (annex 5. 1). The fund coordinated its opera- defense of the interests of the depositors and of the tions closely with the central bank, particularly in the fund itself." The fund became a separate public entity early and final stages of intervention. The fund also operating under private law. Its expanded powers worked closely with the bank in designing the finan- included not only administration of the enlarged cial packages that were the basis of negotiations for deposit insurance scheme but also the ability to pur- the resale of banks to other banking institutions. It chase any bank assets at book value, make deposits at worked speedily and efficiently, selling most of the any rate or term, and provide various guarantees to banks it acquired within one year, and frequently acquiring institutions. It was also authorized to within six months. assume bank losses in order to restore solvency and Rescuing the banks. Once the Bank of Spain facilitate sale within a maximum term of one year. requested that the fund intervene, the fund followed a Moreover, the fund could inject new capital in order number of steps in dealing with each problem bank. to restore equilibrium to insolvent institutions.2 In its Before that request was made, the Bank of Spain pro- new form the fund supplanted the Banking vided emergency liquidity support to the problem Corporation, which had not had the means to restore bank. In some cases the Bank of Spain determined solvency to banks, making it difficult to find buyers that a bank was viable if remedial actions were taken for the banks it took over. and agreed to a plan of action that might include an The fund pursued its principal objectives of pro- injection of capital from shareholders and short-term tecting depositors and restructuring banks for central bank liquidity support. In other cases the cen- prompt sale even where there were undesirable trade- tral bank identified and assessed the extent of insol- offs, such as further concentration in the Spanish vency and demanded recapitalization by existing banking industry. By mid-1985 the number of banks shareholders before inviting fund intervention.3 The had fallen from I 10 to 90, for example, and 85 per- Bank of Spain also permitted fund appraisers to col- cent of the total deposits of the system were held by laborate in the assessment so that preliminary esti- eight Spanish banking groups (de Juan 1985). The mates of the steps necessary to resolve the bank's crisis fund maintained its identity as a functionally special- could begin before the fund took complete control. 92 BANK RESTRUCTURING Table 5.2 Interventions by the Deposit Guarantee Fund, 1978-85 Bank Date of initial intervention Date ofiale Buyer Navarra January 1978 Closed for liquidation (May 1978) Cantabrico February 1978 July 1980 Banco Exterior de Espana Meridional April 1978 July 1981 Banco de Vizcaya Valladolid December 1978 March 1981 Barclays Bank Granada January 1979 December 1980 Banco Central Credito Comercial January 1979 Januasy 1980 Banco de Vizcaya Asturias February 1980 November 1980 Banca March Lopez Quesada April 1980 June 1981 Banque Nationale de Paris Promocion de Negocios April 1980 November 1981 Banco de Bilbao Catalan de Desarrollo May 1980 May 1980 Banco Espanol de Credito Industrial del Mediterrano July 1980 January 1981 Banca Catalana Occidental July 1981 July 1982 Banco de Vizcaya Comercial Occidental July 1981 July 1982 Banco de Vizcaya Descuento November 1981 July 1983 BCCI Holding (Luxembourg) Pirineos December 1981 - Closed for liquidation (December 1981) Union March 1982 April 1982 Banco Hispano Americano Prestamo y Ahorro March 1982 April 1982 Banco de Vizcaya Mas Sarda March 1982 April 1982 Banco de Bilbao Lzvante October 1982 July 1983 Citibank/Banco Zaragozano Catalana November 1982 May 1983 Group offourteen banks Industrial de Cataluna November 1982 May 1983 Group of fourteen banks Industrial del Mediterraneo November 1982 May 1983 Group offourteen banks Barcelona November 1982 May 1983 Group offourteen banks Gerona November 1982 May 1983 Banca March Alicante November 1982 May 1983 Banco Exterior de Espana Credito e Inversiones November 1982 May 1983 Banco Central Simeon December 1983 Banco Exterior de Espana Urquijo-Union July 1985 Banco Hispano Americano Finanzas September 1985 The Chase Manhattan Bank - Not applicable. a. The fund did not arrange these rescues, but it did collaborate in the financial reorganization. Sourre: Deposit Guarantee Fund; Cuervo 1988. Shareholders unwilling to inject new capital faced objectives: amortizing potential losses in the bank the prospect of no new liquidity support from the and penalizing shareholders by diluting or writing off Bank of Spain and possible suspension of their bank's their participation (de Juan 1985). In some cases the charter. Under such pressure they had no choice but fund obtained control only after applying the accor- to relinquish control of their banks to the fund. An dion to the holding company. exceptional amendment to the General Corporate The new capital requirements were determined Law lowered the quorum of shareholders required to not only on the basis of losses suffered through write- approve necessary capital increases. This enabled the offs, but also by taking into account the amount of fund to act quickly and decisively once the Bank of deposit liabilities the bank had under normal operat- Spain requested its intervention in an insolvent bank. ing circumstances. Thus the fund effectively replen- Transfer of control occurred at a shareholders ished the original equity and pre-panic level of assembly convened no later than seven days after the deposits. If further funds were needed, the fund could Bank of Spain requested it. The Bank of Spain purchase assets, offer guarantees or counterguarantees informed shareholders of the extent of losses and the on behalf of the restructured bank, grant long-term impact of writeoffs on the capital and reserves of the loans at subsidized rates, or permit temporary regula- institution. If the bank had no net worth-that is, all tory forbearance. The fund also immediately replaced the capital was lost-and shareholders would not all top executives with trusted professionals. The new recapitalize it, the fund would pay a nominal 1 peseta management introduced administrative reforms, per share (more if the capital had not been complete- improved operational efficiency, and enhanced the ly eroded). Once the fund had gained full control marketability of the resuscitated institution. In any through this first phase of the "accordion" operation event, all action taken by the fund was predicated on (simultaneous reduction and increase in capital), it the assumption that the revived banks would continue could subscribe the new capital needed to restore the to be viable after the fund's support ceased and the bank's viability. The accordion operation served two bank's liabilities to the fund were eliminated. Bank Restructuring in Spain, 1977-85 93 The fund subscribed substantial new capital in to make counterproposals that were submitted to the those banks with solvency problems (annex 5.2). fund's board of directors. Once the board selected an Only three banks showed no change in their original offer, it had to communicate the decision to the capital. In nineteen cases more than 50 percent of Ministry of Finance, which had fifteen days to exer- original capital was written off, and in thirteen of cise its option to buy the shares in cases where the those cases almost all initial capital was lost. Among national interest was affected. This two-stage process the nineteen, only four received capital increases from provided a check and balance to the restructuring the fund that were less than the amount of their ini- efforts. tial capital endowment. In two cases the fund The special case of Rumasa. Rumasa, a Spanish replaced capital fully to maintain the size of the bank, holding company comprising more than 300 sub- and in the remaining thirteen cases the banks sidiaries, owned 20 small and medium-size banks and received more than the amount of their original capi- 100 productive industrial or commercial enterprises. talization from the fund. The capital increases sub- The rest of its offshoots were mainly investment scribed by the fund, however, represented only a frac- trusts and their instrumentalities, created to obscure tion of the resources used to restructure insolvent both risk concentrations and ultimate ownership of banks, the greatest cost being the purchase of nonper- various companies. The conglomerate was involved forming loans. Because the fund normally acquired a primarily in the beverage, hotel, commercial, con- substantial portion of the nonperforming assets of struction and property, and financial services indus- insolvent banks, it was in charge of selling all the tries. Four major companies comprised 51 percent of assets transferred to it, usually during a period the group, while sixteen companies accounted for 71 beyond that of its provisional administration of the percent of turnover. All the companies were highly rescued bank. leveraged and, according to official estimates, the When a suitable buyer for an insolvent bank holding company had a substantial negative net could be found immediately, the fund tried to sell the worth. Operating losses were increasing. The sub- bank right away in order to avoid loss of confidence group of banks was the source of finance for Rumasa's in the institution and further erosion of the deposit development. base. In other cases a comprehensive external audit by Starting in the early 1960s, when the banking sec- a recognized accounting firm was completed and tor was opened to limited competition from local and potential buyers identified. The Bank of Spain and regional banks, Rumasa began to acquire local banks. the fund together prepared the basis for the private It continued to build its banking group over the next offering. The prospectus included at the very least a twenty years. It pursued an aggressive acquisition summary of how the fund acquired the bank, action strategy in the early 1 970s involving inflated stock taken to restore its solvency, a clear definition of the market valuations for the target banks. Rumasa also fund's objectives in selling the bank, and a statement bought banks that were facing difficulties. By 1983 of maximum value and the specific conditions under seventeen of the Rumasa banks had extended more which problem assets could be carved out by the than 73 percent of their total credit to the private sec- fund. This last element was the main adjustment fac- tor, concentrated with Rumasa and its related compa- ror in ensuing negotiations. The basis of the offer nies. At the time of the expropriation in 1983 the detailed the financial support offered by the fund and banks showed a negative capital of about 21 billion the Bank of Spain, including loans, subsidized inter- pesetas ($146.4 million equivalent). est rates, and terms; exemptions to technical coeffi- Because of the magnitude of the Rumasa problem cients (regulatory forbearance); and any other infor- and the close relationships between its banks and mation deemed relevant, such as guarantees against nonfinancial companies, this case was handled differ- "hidden liabilities." It also required that the purchaser ently from the others. If the fund had intervened to renounce future claims or legal actions against the rescue the banks, most of the productive enterprises fund resulting from differences that might arise would have failed because of their excessive indebted- between expected and realized returns on assets, and ness. In addition, it was not considered an appropri- it stated the availability of additional information and ate disposition of fund resources (half of which came the final date for receipt of bids. from the banking system) to apply them to the rescue The prospectus was circulated to a select group of of nonfinancial companies. Moreover, the employ- domestic and foreign institutions, who were invited ment impact of such large-scale enterprise failures- 94 BANK RESTRUcruRING Rumasa employed 50,000 people, 11,000 of them in would be amortized over twelve years and repaid as its banks-presented a major public policy challenge. the government's twelve-year debt issue matured. In Consequently, in early 1983 the government decided effect, the government would repay the twelve to nationalize (expropriate) the holding company and Spanish banks; they in turn would repay the deposits its related businesses. of the former Rumasa banks; and those institutions Administration of the holding company was would repay their outstanding credits from the Bank entrusted to the Department of National Heritage, a of Spain (de Juan 1985). government department, which followed a two- pronged strategy. First, it took effective control of all Sharing the costs the companies, some of which were discovered weeks after the holding company had been expropriated. In all cases shareholders suffered the first losses by los- Second, it developed a plan to sell the companies as ing their equity. Except for the three small liquidated soon as possible. The Rumasa group was separated banks, depositors and other creditors did not suffer into subgroups for this purpose: two large productive direcdy from the crisis.4 The balance of the losses were companies, small and medium-size productive com- absorbed jointly by the largely private banking system, panies, and banking institutions. An external audit of the government, and the Bank of Spain. In the case of all companies began immediately, and a conservative the banks controlled by the Deposit Guarantee Fund, management strategy was followed as the sale pro- the losses were shared in equal parts by the Bank of gram progressed. First Boston, an international Spain and the banking system, reducing the system's investment bank, was entrusted with the sale of large technical reserves. Contributions from the banking companies and acted as an adviser to the government community to the fund amounted to 1.2 per 1,000 of regarding any offer to purchase the small and medi- liabilities in domestic and foreign currency per year, um-size companies. Management and sale of the equivalent to 10 percent of annual profits. By banks rested with the fund, which in this case acted December 1984 these contributions totaled about solely as manager. Productive companies were sold $0.5 billion, too little to absorb all the losses. rapidly, sometimes to foreign investors, because the Additional borrowing from the Bank of Spain was government assumed enough of the companies' bad necessary, which totaled about $2.9 billion. The Bank debts to make the enterprises viable. The largest of Spain loans were long term with a 7.25 percent Rumasa bank was sold to a foreign consortium, and annual interest rate. According to the fund's estimates, two others were sold to domestic investors. The and assuming that no additional banks require its remaining seventeen banks were purchased by and intervention in the period considered, its reserves will divided among the seven largest Spanish banks. The be sufficient to repay the Bank of Spain loan com- banking system played an important part in this sal- pletely by the end of the 1990s. vage operation, not only by acquiring the Rumasa In the Rumasa case, costs to the government banks but also by bearing some of the financial losses. included the annual interest on its twelve-year debt Twelve banks, for instance, purchased low-interest issue, less whatever it recovered from the Treasury (9.5 percent) twelve-year government bonds and loan to Rumasa. In addition, the government had to accepted 13.5 percent interest-bearing deposits from underwrite the financing cost of the Bank of Spain Rumasa banks. This debt instrument provided the loans to the Rumasa banks in excess of what the bank main source of funds for the government's long-term would have lent to cover ordinary deposit with- (twelve-year) interest-free loan, amounting to drawals. Costs to the banking community included 400,000 million pesetas ($2.5 billion equivalent), to the difference (4 percent) between the interest paid to the Rumasa holding company. That loan was chan- the Rumasa banks on their deposits and the interest nelled to the Rumasa banks to repay the loans they earned from the public debt issue held by banks, had made to other Rumasa entities. In addition, the which would decrease over time. Bank of Spain extended credits of 400,000 million In addition, the public shared in the losses pesetas, at 8 percent interest, to the Rumasa banks through higher lending rates and lower deposit rates before they were sold. These banks therefore enjoyed imposed by the banks to cover the cost of the insur- a 5.5 percent margin between the cost of funds from ance premiums to the fund. Taxpayers also assumed the Bank of Spain and the income from their a loss from the larger financial burden of new public deposits with the twelve Spanish banks. This amount debt issues arising from the crisis and from the Bank Restructuring in Spain, 1977-85 95 higher inflationary tax resulting from the central amounted to 1.52 percent of total assets in 1990, a bank's credits to the fund. slight decrease from the previous year but above the 1.4 percent reported in 1988 (Financial Times, 23 The big-bank problem May 1991). Another challenge to the banking system loomed, Conclusion however. In the late 1980s some of Spain's largest banks also were confronted with the need to clean up The institutional reforms adopted in Spain to deal their loan losses and deal with other problems. Given with the banking crisis evolved rapidly as financial the Bank of Spain's gradualist approach, it was only conditions deteriorated. Although the authorities ini- after the crisis in the other banks had been tackled tially underestimated the extent and depth of the that attention turned to the big banks. banking crisis and provided only limited deposit Although some banks were in a worse position guarantees, they soon developed a fair, flexible, and than others, most of their problems stemmed from pragmatic approach to resolving individual bank five principal difficulties: crises. Moreover, they were realistic in assessing the * Loan losses from their extensive relations with appropriate limits of the Deposit Guarantee Fund's industry. capabilities by separating the problems of the Rumasa * Losses from their controlling interests in banks group from the scope of fund intervention. that the authorities had persuaded them to buy Substantial improvements in the legal and because of overlapping directorships or close rela- accounting frameworks and in bank regulation and tions, but in which the Deposit Guarantee Fund supervision have flowed from the crisis as well. Law had not intervened. 13 (1985) reduced the amount of compulsory invest- • Excessive overheads. ment by banks and addressed matters of capital ade- * Loans to developing countries, which were affect- quacy and financial disclosure. Royal Decree 1144 ed by the new rules on provisioning for country (1988) regulates the creation of new banks and the risk. installation of foreign banks. Law 26 (1988) enhances • Lapses in pension fund accruals to cover retire- official intervention and enforcement powers, while ment plan contingencies. The Bank of Spain Law 19 (1988) compels external audits of all financial forced the banks to constitute the necessary institutions supervised by the Bank of Spain. In addi- reserves. tion, an outpouring of central bank circulars have It is difficult to quantify the massive provisions refined and clarified supervision and regulation. made in the late 1980s by several of the biggest The Spanish crisis was manageable, limited to 20 banks, but it is clear that profitability was squeezed as percent of banking system deposits and 21 percent of a result. In at least two cases banks were forced to sus- total net worth. The largest private commercial pend dividend payments. Between 1986 and 1988 banks, together with the Bankers' Association, were three large banks constituted provisions and charged able to assist substantially in resolving the fate of off amounts (totaling $6 billion) that were nearly nearly all the smaller failed banks, albeit under con- double that of the three best-performing big banks. siderable state pressure. Competition in the home In addition to tapping their normal operating profits, market from foreign banks also provided an incentive some banks had to liquidate and sell off nonbanking for Spain's private commercial banks to acquire recap- assets in order to finance the provisioning effort. italized Spanish banks, sometimes even by assuming Despite these problems Spanish banks remain losses in the short run. highly profitable because of Spain's continued eco- Indeed, the banking crisis of 1977-85 spurred a nomic prosperity and because of their access to rela- fundamental restructuring in the Spanish banking tively cheap deposits and their ability to lend at rela- industry, resulting in fewer but larger institutions. tively high rates. In 1991, for example, two of the This is reflected in the fact that in 1990 the largest larger banks were among the five most profitable seven Spanish banks accounted for 61 percent of banks in the world. In 1989 the lending margin for total bank assets, 47 percent of all deposits, and 63 Spanish banks was equivalent to 4.4 percent of total percent of all bank loans. The presence of more for- bank assets, the highest of any European Community eign banks added a competitive impetus to innova- member except Greece or Portugal. Net profits tion as well, although they generally did not fare well 96 BANK RESTRUCTURING against the domestic competition in the years follow- antees to any depositor directly related to the finan- ing the crisis. cial difficulties of a bank. Furthermore, the board can Mergers with and acquisitions of the smaller approve the purchase of assets by the fund from a banks by the larger solvent ones did not lead, howev- bank facing financing difficulties (so long as such er, to needed rationalization in the domestic banking purchases avoid a further involvement of the fund) system. Even the spate of bank mergers in 1988 and without precluding further requests to the bank man- 1989-a movement endorsed by the government- agement to take additional remedial actions. Also, the was followed by a period of relative self-satisfaction board requests advances from the Bank of Spain to within the Spanish banking community. Interest rates the fund when needed. remained high, resulting in higher retail margins than The fund is organized in three departments: legal, in almost all other European countries. The resulting administration and control, and asset management. high profits compensated for enduring problems: The Legal Department can initiate legal or criminal namely, the ingrained inefficiency of high staff levels action against previous administrators of banks under and too many branches and imposition of tight mon- intervention or temporary administration by the etary policies, including high reserve requirements fund. The fund always tries to minimize its losses by and continued compulsory investments. In addition, maximizing recovery of assets purchased at book value the pressure to provision against past losses has been from an insolvent bank and by repossessing any guar- compounded by the need to improve capital levels to antees received as collateral for nonperforming loans. internationally accepted standards. Consequently, the The Legal Department supports these efforts and prospect of greater competition within a unified defends the interests of the fund and of the banks European market promises far greater rationalization, under its control or liquidation. Finally, the depart- more consolidation, and broader diversification for ment is legal counsel for the other departments. Spanish banks in the 1990s. The Department of Administration and Control is responsible for the internal administration of the Annex 5.1 Organization of the Deposit Guarantee fund and of those companies temporarily owned by Fund it as a result of asset purchases. It controls and recov- ers claims acquired by the fund, maintains timely The board of directors of the Deposit Guarantee records, and services obligations of fixed assets (for Fund has many functions, including informing and example, property taxes). In addition, the department advising the Bank of Spain on all matters regarding coordinates the sale of assets with other departments, the operation of the fund, determining the best ways including reprivatizing banks. It also prepares the of achieving its objectives, and preparing and approv- annual budget and financial plan of the fund, includ- ing the fund's financial statements. The board ing requests for advances from the Bank of Spain. informs the central bank which banks face financial The Asset Management Department is responsi- difficulties possibly requiring fund intervention. It ble for the effective administration and sale of all also determines the form of payment of annual pre- assets received by the fund but not directly related to miums paid into the fund by the banks, sets the banking activities (for example, industrial firms). It requirements for admitting new banks to the fund, evaluates their financial viability, minimizes further and informs the public of any change in member- financial commitments by the fund, and tries to sell ship. It requests external audits of member banks, the best assets quickly. It tries to improve less attrac- decides the frequency and extent of those audits, and tive assets while searching for possible buyers. It also also may request external audits of related companies. appraises, or subcontracts appraisal of, the fixed assets In addition, it can suspend payment of deposit guar- in the fund portfolio. Bank Restnicturing in Spain, 1977-85 97 Annex 5.2 Deposit Guarantee Fund control and subscription of new values (millions of pesetas) Writeoff against Capital Initial capital Capital Final incirasel Bank Means of acquisition capital value incrrase capital initial capital Cantabrico I peseta per share 764 267 795 1,292 1.04 Meridional I peseta per share 1,125 1,012 1,102 1,125 0.90 Valladolid I peseta per share, plus 700 2,200 1,100 4,400 5,500 2.00 million pesetas to cancel debt of previous owners Granada I pesera per share 2,247 786 3,054 4,515 1.36 Crediro Comercial Controlled through its 776 0 0 776 0.00 holding bank. Asturias - 936 468 936 1,404 1.00 Lopez Quesada - 1,621 810 3,100 3,911 1.91 Promocion de Negocios - 1,444 722 1,444 2,166 1.00 Catalan de Desarrollo - 2,625 2,625 3,000 3,000 1.14 Industrial Mediterraneo - 2,100 1,050 2,500 3,550 1.19 Occidental I pesera per share 4,630 4,621 7,000 7,009 1.51 Comercial Occidental Controlled through its 1,625 1,621 500 504 0.31 holding banks Descuento I peseta per share 2,250 2,245 2,500 2,505 1.11 Union 50 percent of the nominal price 7,723 0 0 7,723 0.00 of each share Presramo yAhorro Controlled through its 2,256 2,233 1,500 1,523 0.66 holding banks Mas Sarda 2,283 1,529 3,000 3,754 1.31 Levante I peseta per share 1,992 1,990 5,500 5,502 2.76 Catalana 5,754 5,748 15,344 15,350 2.67 Industrial de Cataluna Controlled through its 4,258 4,254 4,500 4,504 1.06 holding banks Industrial del Mediterraneo Controlled through its 3,550 3,543 3,500 3,507 1.00 holding banks Barcelona Controlled through its 568 567 1,137 1,138 2.00 holding banks Gerona Controlled through its 397 0 0 397 0.00 holding banks Alicante Controlled through its 1,594 1,592 2,500 2,502 1.57 holding banks Credito e Inversiones I peseta per share, controlled 1,595 1,592 3,500 3,503 2.19 through holding banks -Not applicable. Source: Deposit Guarantee Fund; Larrain and Montes-Negret 1986; author's calculations. 98 BANK RESTRUCTURPNG Notes 4. In the cases of liquidation, large creditors shared in the losses to the extent that the portion of deposits above 1.5 mil- This chapter is based principally on Larrain and Montes-Negret lion pesetas (which was uninsured) could be recovered only from (1986) and Cuervo (1988). the net proceeds of liquidation. 1. Industrial banks were permitted to lend up to 10 percent of their total resources (own funds and borrowed ones) to a sin- References gle borrower. Given that capital adequacy was set at a 10 percent gearing ratio, in practice total exposure to a single customer Alvarez Rendueles, Jose Ram6n. 1983. "El Tratamiento de las could reach up to 110 percent of capital. The limits on lending Crisis Bancarias en Espafia." In Crisis Bancarias: Soluciones to a single borrower were far less liberal for commercial banks, Comparadas, seminario desarrollado en la Universidad with maximum exposure reaching as much as 33.75 percent of Internacional Menendez Pelayo, en colaboraci6n con la capital. Asociaci6n Espafiola de Banca Privada, Julio, Santander. 2. One limitaiion involving the deposit insurance system is Cuervo, Alvaro. 1988. La Crisis Bancaria en Espafia 1977-1985. the asymmetry resulting when a depositor has a claim in a Barcelona: Editorial Ariel. bank that is liquidated by the fund. In contrast to cases where deJuan, Aristobulo. 1985. 'Dealing with Problem Banks: The Case the fund resells banks, in liquidation depositors receive only of Spain." Paper presented at the seminar on central banking, the portion of their deposits covered by the guarantee-100 International Monetary Fund, July 1-12, Washington, D.C. percent up to 1.5 million pesetas. Any remaining balance is . 1987. "From Good Bankers to Bad Bankers: Ineffective included among outstanding claims against the net proceeds of Supervision and Management Deterioration as Major the liquidation. In other cases, however, all deposits are Elements in Banking Crises." Economic Development insured independently of their size, since the fund insures the institute Working Paper. World Bank, Washington, D.C. solvency and continued operation of rescued banks. This dis- Financial Times. 1991. "Spain." May 23. tinction was of little consequence during the banking crisis, Larrain, Mauricio, and Fernando Montes-Negret. 1986. "The however, because only three small banks were liquidated. Spanish Deposit Guarantee Fund." World Bank, Financial 3. The Bank of Spain's assessment included the quality of a Policy and Systems Division, Washington, D.C. bank's loan portfolio, concentration of loans and credits to relat- Lopez-Claros, Augusto. 1989. "Growth-Oriented Adjustment: ed parties, and estimates of losses and of new capital needs. In Spain in the 1980s." Finance and Development. 26 (March): most cases, however, a realistic assessment could not be made 38-41. until the fund had taken control of the bank because cosmetic Minsky, Hyman P. 1982. "The Potential for Financial Crises." accounting methods were used by management to obscure the Working Paper 46. Washington University, Department of magnitude of the bank's problems. Economics, St. Louis. CHAPTER 6 Structural Weaknesses and Colombia's Banking Crisis, 1982-88 Fernando Montes-Negret Signs of an impending crisis in Colombia's banking tion-including outright nationalization of some system first appeared in 1981 when the country's banks-and assistance to nonbank financial interme- terms of trade began to deteriorate, resulting in diaries and enterprises. Because Colombia had a large loan losses. Seven banks-accounting for strong fiscal and external position, the authorities about a quarter of the total assets of the Colombian were able to resolve bank failures and introduce a system-failed in the ensuing crisis of 1982-88. macroeconomic stabilization program more success- Superficially, their failure was part of a classic fully than many other countries that experienced boom-bust economic cycle that began in the mid- banking crises. Although adjustment policies initially 1970s. The business cycle explanation, however, is exacerbated the loan loss and cash flow problems of inadequate. Until the crisis Colombia had enjoyed the banking system, in the end improved public relatively long-term economic stability. Real growth finances supported the restructuring effort and decelerated sharply from a peak of 8.5 percent in restored public confidence. 1978 to 0.9 percent in 1982, though it was never negative during the 1980s. Furthermore, foreign- Macroeconomic Background controlled banks in Colombia withstood the dis- tress far better than their domestic competitors, Major macroeconomic developments and policy largely because they were managed more conserva- decisions affected the financial system's performance tively and had a larger capital base. And most finan- over three stages: the coffee boom of 1975-80, the cial intermediaries, such as development finance domestic and international recession of 1981-85, companies and savings and loan corporations, also and the recovery of 1986-88. During the first phase were generally resilient. (Most of the failures in the a large and unexpected improvement in the terms of nonbanking sector were concentrated among trade triggered sustained growth. Real GNP growth finance companies, where substantial fraud and more than doubled, from around 2.5 percent in the other abuses occurred.) early 1970s to an annual average of 5.4 percent dur- Rather, the Colombian crisis was the result of ing 1976-80, peaking at 8.5 percent in 1978 (table structural weakness in the banking system, high cor- 6.1). International coffee prices rose from $0.65 a porate leverage, inadequate bank supervision, and pound in 1975 to $2.35 in 1977, boosting export fundamental policy misjudgments, which exacerbat- revenues from $635 million to $2 billion by the end ed the effects of the business cycle. Two distinctive of the decade. Net international reserves rose tenfold features of the Colombian crisis are the linkages of in five years, reaching $5.6 billion in 1981. banks with industrial conglomerates, or groups (gru- The increase in prosperity, however, was accom- pos), and the large losses incurred by the Colombian panied by rising inflation, which reached a peak of banks' offshore affiliates and subsidiaries. 33 percent in 1978. Authorities fought inflation The restructuring of the Colombian banking sys- with tight monetary policy and a crawling peg tem required government-financed recapitaliza- exchange rate mechanism, which resulted in a 99 100 BANK RESrRUCTURING cumulative total real appreciation of 27 percent dized credit, tax biases favoring debt accumulation, between 1974 and 1982. Monetary policy com- and legal disincentives to retain earnings. The heavy prised two major measures: the introduction in concentration of ownership in the corporate sector 1977 of a marginal minimum legal reserve require- reinforced the drive toward debt leveraging. New ment of 100 percent on demand deposits, with equity issues were limited so that existing sharehold- higher requirements on other deposits, and a central ers could maintain their controlling interest, control- bank, dollar-denominated cash certificate issued to ling the large groups with high debt and low equity. exporters at favorable interest rates to neutralize any Where groups also controlled banks, these banks inflation pressures from export proceeds. Both mea- provided a captive source of financing for groups sures affected banks' profitability. The legal reserve seeking to expand. Moreover, restrictive monetary requirement reduced the money multiplier, lowered policies and an overvalued exchange rate encouraged profits, weakened capitalization, and made money foreign borrowing, since foreign loans were available scarce and expensive. The cash certificates encour- at lower interest rates. aged disintermediation because banks were still sub- The banking system responded by rapidly ject to interest rate controls. Overvaluation of the expanding credit. Overall growth in the banking sys- exchange rate encouraged imports and reduced tem was 39 percent in 1977, but private commercial domestic competitiveness, with widespread conse- banks expanded credit by 48 percent. Forced invest- quences for all economic sectors. ments and limits on lending to individual borrowers This mix of policies was imposed on a highly encouraged banks to book loans abroad.' Foreign leveraged and interconnected banking and corporate subsidiaries of Colombian banks handled the foreign system that was augmented by largely unregulated borrowing of Colombian corporations, frequently offshore activities. The consequences were disas- directing the lending to related parties. The proceeds trous. The economic boom encouraged a sharp rise of some loans were illegally repatriated to Colombia, in corporate borrowing to finance investment and where they financed takeovers or highly leveraged speculative activities. Corporate debt soared to more projects that were unlikely to generate foreign than 70 percent of total liabilities in the first half of exchange in the short run, if at all. High-risk curren- the 1980s, compared with 24 percent in 1950 and cy mismatches appeared in the balance sheets of about 50 percent at the end of the 1970s. Such high banks' related companies and caused the quality of debt ratios were the result of readily available subsi- bank portfolios to deteriorate. Table 6.1 Macroeconomic indicators, 1975-88 (percent unless otherwise specified) Real Curmrnt fiscal Terms Nominal Real Change in Change in exchange rated Realgrowth accountl budgetl of tradea Inflation interest interest industrial agricultural (December Year in GNP GNP GNP (1980=100) rate b rate' rate output output 1986-100) 1975 2.3 -0.8 -0.2 79.2 23.1 25.2 1.7 1.2 5.8 87.5 1976 4.7 1.4 1.0 117.6 20.2 28.1 6.6 4.4 3.1 83.5 1977 4.2 2.3 0.6 154.4 33.1 26.7 -4.8 1.4 3.3 75.1 1978 8.5 1.4 0.7 118.2 17.8 28.8 9.4 10.0 8.1 74.9 1979 5.4 1.8 -1.0 113.2 24.6 33.4 7.0 6.1 4.8 71.5 1980 4.1 -0.5 -1.9 100.0 26.6 35.3 6.9 1.2 2.2 73.1 1981 2.3 -5.2 -3.0 90.5 27.5 37.3 7.7 -2.6 3.2 71.4 1982 0.9 -7.5 -4.8 96.4 24.6 38.0 10.8 -1.4 -1.9 66.2 1983 1.6 -7.2 -4.3 94.5 19.8 33.7 11.6 1.1 2.8 64.4 1984 3.4 -3.3 -4.4 99.0 16.1 34.8 16.1 6.0 1.8 69.9 1985 3.1 -5.3 -3.3 98.3 25.0 35.2 8.1 3.0 1.6 80.0 1986 5.8 1.1 -1.6 99.9 18.9 31.0 11.0 5.9 3.4 95.0 1987 5.4 0.9 -0.7 70.3 23.3 31.1 6.3 6.2 6.4 97.3 1988 3.7 -1.6 -1.3 68.1 28.1 33.9 4.5 2.2 2.6 97.4 - Nor available. a. Ratio of the index of average export prices to the index of average import prices. b. Consumer price index. c. For 1975-79, rare on 120-day Certificados de Abono Tributario (CATs). For 1980-88, rate on 90-day certificates of deposit. d. Colombian pesos to U.S. dollar. Sourme: Montes-Negret 1990; IMF, various years. Stuctural Weaknesses and Colombias Banking Crisis, 1982-88 101 Financial Sector Structure and Policies ing countries in the late 1970s. This lending allowed banks to expand their offshore and international The Colombian financial system is a paradox: con- activities-particularly in Panama, where supervision centration prevails amid competition and the public was minimal. There were several reasons for this sector dominates amid private sector oligopoly. Of international expansion. First, offshore subsidiaries the twenty-six commercial banks, nine publicly allowed banks to circumnvent the functional separa- owned ones accounted for 63 percent of the banking tion of commercial and investment banking because system's total assets at the end of 1987. (Four of the their Panamanian subsidiaries could participate in nine were nationalized after 1982.) Private bank syndicated lending and other international money ownership is concentrated in a few powerful con- market operations. Second, the small size of the glomerates, with the five largest financial groups con- banks' capital base ($524 million at the end of 1981) trolling half the combined equity of commercial restricted their capacity to lend, particularly to large banks, finance companies, housing banks, and trade projects and on an unsecured basis to individual finance companies. clients. Thus even though Colombian law restricted During the 1950s and 1960s the Colombian capital-liability ratios to 1:10, Panamanian affiliates financial sector was severely repressed. The govern- and subsidiaries sometimes registered capital-liability ment controlled interest rates, determined credit allo- ratios of 1:20 or more. Third, the export boom and cation, and subsidized directed credits through the the associated surplus in foreign exchange encouraged rediscount mechanism of the central bank. These capital outflow and overseas expansion. Overseas policies impeded efficient resource allocation and investments amounted to 80 percent of the capital domestic mobilization of savings and led to the and reserves of the domestic Colombian banking growth of a large curb market. system at the end of 1984. By the end of 1986 Financial sector liberalization in the 1970s Colombian banks had three branches, twenty-three increased competition and domestic savings and subsidiaries, and eighteen affiliates abroad. resulted in the creation of savings and loan corpora- Colombian supervisory authorities did not fully tions. These corporations eroded the market share of appreciate the additional risks incurred by this expan- the banks from 57 percent in 1976 to 46 percent in sion overseas, nor did they recognize the magnitude of 1988. The monetary reform in 1974 further opened loan concentration to related parties or the extent of the financial sector in order to improve monetary exchange control violations. In most cases the loans control and rein in the growing curb market. funding the overseas subsidiaries and affiliates were Commercial banks' access to the central bank was obtained through international borrowing secured by restricted primarily to lender-of-last-resort facilities, guarantees furnished by the parent banks. These cred- although access to other sources of funds was it supplements were registered mostly as foreign loans enhanced. The minimum legal reserve requirement in Colombia or were brought in illegally to finance was lowered, and banks were allowed to offer attrac- takeovers or domestic speculation.2 The offshore enti- tive interest rates. Market-determined interest rates ties also overextended themselves in syndicated lend- were gradually adopted to sustain a positive real ing to regional sovereign borrowers, assuming large interest rate structure. Limits on the growth of com- liquidity risks. As a result these offshoots compro- mercial bank loans also were abolished. mised the solvency of their parent banks. The introduction in 1975 of the "Colombianiza- The risks were not wholly overseas, however. A tion" policy, however, retarded domestic competition. study of the banks' past-due loans as of 1987 suggests Foreign investment in banks was forbidden, and exist- that such loans were highly concentrated by industrial ing foreign banks had to convert to mixed ownership, sector and by company. Almost a third of past-due with at least 51 percent of capital held by Colombians. loans were owed by eleven manufacturing companies; As a result concentration increased in the large domes- two of them were state owned. One company tic conglomerates. The lack of competition hampered accounted for a third of the past-due loans owed by innovation and technological change, and banks' mar- these companies. The manufacturing sector's debt-ser- ket share fell in a number of financial services. vicing difficulties reveal not only flawed financial A significant development during this period was structures, but also inefficient production and man- the internationalization of Colombian banks, a result agement. These past-due loans were highly concen- of the rise in international bank lending to develop- trated in the banking sector: 76 percent (by value) of 102 BANK RESTRUCTURING the loans were held by six banks, five of which eventu- Table 6.2 International reserves and domestic ally required government assistance of some kind. liquidity, 1980-88 Finally, a major flaw in the Colombian banking (percent unless otherwise specified) system was the quality of its capital base. In 1980 Net local banks were allowed to include as part of their internationes , , , . ,- , , , ,, ~~~~~~~Extern2al dbt rejerves Change in capital the difference between the market and histor- (millions of (millions of Change Change monetary ical value of their fixed assets. By 1984 more than Year U.S. dollars) U.S. dollars) in MI in A2 base half of total bank equity was accounted for by the 1980 6,941 5,356 - - - revaluation component, thus eroding the capacity of 1981 8,716 5,630 1982 10,306 4,891 25.4 24.6 17.7 the banks to absorb bad asset and other losses. 1983 11,412 3,078 24.7 29.8 13.5 1984 12,039 1,796 23.4 24.5 18.3 Crisis and Resolution 1985 14,241 2,068 28.2 34.4 25.9 1986 15,364 3,479 22.8 28.8 28.2 1987 17,008 3,451 32.9 28.0 31.5 The boom ended in 1981 with the onset of the 1988 17,001 3,811 25.8 23.4 26.9 international recession, with coffee revenues and -Not available. nontraditional exports falling sharply. The effects of SGurre: Montcs-Negret 1990; World Bank dara. the overvaluation of the peso and the decline in exports initially were cushioned by an increase in oil Inflation moderated from 27 percent in 1981 to 16 exports and by the high levels of accumulated exter- percent in 1984, thanks to the stabilization program nal reserves. Imports continued at unsustainably adopted and the continued high unemployment (13 high levels, however, despite restrictive policies intro- percent in 1984). But sticky interest rates, the crowd- duced in 1982. At first the authorities chose to post- ing-out effect of public sector borrowing, and higher pone fundamental adjustment, hoping the economy risk premiums in anticipation of a peso devaluation would grow out of its difficulties through expansion- resulted in high real rates of interest on time deposits ary fiscal and monetary policies. This strategy (between 10 percent and 12 percent) and loans (11.0 swelled the deficit of the overall nonfinancial public percent to 14.5 percent). This compounded the loan sector to 6.7 percent of GDP in 1984. quality problems of banks and their overindebted cor- These deficits were financed by external borrow- porate customers-the number of which rose sharply ing, central bank loans to the government (until with higher real lending rates (table 6.3). 1983), and capital gains taxes on foreign currency The first indication of serious banking troubles and gold holdings. But the net services deficit in the came in June 1982 when the authorities intervened in balance of payments deteriorated markedly after Banco Nacional, which accounted for 2.3 percent of 1980-sliding from a deficit of $74 million to almost total bank assets, and subsequendy liquidated it. Small $1.7 billion in 1983, and $2 billion thereafter-as a depositors were reimbursed. Other creditors received result of record-high international interest rates and soft loans and, along with the central bank, were to be higher interest payments on Colombids rapidly grow- repaid through asset recovery and sales. Because the ing external debt.3 Net capital inflows also dropped, bank had engaged in highly speculative ventures and from about $2.2 billion in 1982 to $0.9 billion in had concentrated loans in companies of Grupo 1984, because of the debt crisis, loss of confidence in Colombia, the controlling conglomerate, its problems Colombian banking operations, and the corporate were attributed to the abuses of a new dass of bankers sector's efforts to repay foreign currency debts. that had appeared following the 1974 liberalization. Government spending, regional relief funds, exten- Although the episode was treated as an isolated event, sive rediscount facilities for distressed private manu- rumors persisted of problems elsewhere. Indeed, in facturing enterprises, and substantially lower legal October 1982 the government was forced to declare a reserve requirements on bank deposits generated twenty-four-hour economic emergency and nationalize excess liquidity, which accelerated external reserve the private Banco del Estado (3.5 percent of total losses. By 1984 the central bank's international banking assets) after it had exhausted its liquidity lines reserves had dropped to 32 percent of their 1981 from the central bank. Although relatively small, level (table 6.2). The current account turned from a Banco del Estado was an important regional bank slight surplus in 1980 to a deficit of 10.8 percent of whose restoration was considered important to public GDP in 1983. confidence. This bank also had concentrated loans to Strctural Weaknesses and Colombias Banking Crisis, 1982-88 103 Table 6.3 Banking system indicators (billions of pesos unless otherwise specified) Past-due loans in Net Substandard Total frerign currencyl Return Total revaluation Loan Substandard loans/total Total past-due totalpast-due on total Year equity of bank assets provisions loans equity (perrent) bank credit loans leansb (percent) capitalc 1975 9.8 0.85 0.69 3.55 29.18 59.75 1,550 32.58 - 1976 12.4 1.05 0.31 5.87 40.76 76.66 3,410 56.74 - 1977 16.4 1.40 1.00 6.48 33.43 96.12 2,900 35.10 - 1978 22.1 3.59 1.10 7.90 30.74 117.46 3,600 30.19 - 1979 29.4 7.07 1.51 10-12 29.27 152.11 4,400 37.25 - 1980 41.5 11.41 2.11 16.80 35.40 233.89 6,220 27.56 - 1981 54.0 19.15 3.05 27.77 45.78 319.19 9.440 24.01 13.8 1982 70.7 28.34 7.24 57.07 70.48 389.92 21,680 21.80 5.6 1983 84.7 37.46 12.15 83.25 83.94 514.08 40,060 19.06 5.5 1984 86.8 42.95 23.43 148.51 144.10 585.94 95,490 35.17 6.3 1985 34.1 56.26 90.70 192.80 298.41 768.80 140,100 46.27 -6.3 1986 163.9 63.49 103.49 216.40 68.91 956.60 156,800 40.62 4.0 1987 196.2 74.70 116.69 224.82 55.13 1,217.66 158,800 41.81 10.0 1988 285.7 120.26 63.06 185.64 42.91 1,583.49 101,430 9.47 7.6 - Nor available. a. Includes paid capital, legal reserves, contingent reserves, revaluations, exchange rate adjustments, callable capital, and bonds forcefully converted into equity ± accumulated profits and losses. b. Commercial banks only. c. Average of six-month figures. Sourre: Montes-Negret 1990. related parties and ignored legal and prudential regula- normal liquidity support from the central bank had tions, and losses proved greater than originally estimat- been exhausted or normal procedures for intervention ed. This rescue set a precedent for the extension of and liquidation might jeopardize public confidence. In public deposit insurance coverage to the entire banking addition, the government was empowered to appoint system-a precedent that had important fiscal conse- new boards of directors and management, assume and quences as the crisis unfolded. capitalize the bank's obligations with the central bank, purchase the shares of previous shareholders, and Improving institutional responses define the mechanism to determine the price of shares and allow the future sale of nationalized institutions. The problems of Banco Nacional and Banco del Finally, the decree introduced financial sector reform Estado marked only the first phase of Colombia's measures, including empowering the government to banking crisis and demonstrated that Colombia was broaden ownership of financial institutions by setting not prepared to deal with extensive bank distress. The limits on individual and group participation and man- decree nationalizing Banco del Estado, however, dating gradual divestiture of stakes above these ceilings. marked a turning point. The decree introduced a code A number of other initiatives were introduced in of conduct for bank managers, prohibiting lending to 1983 and 1984 to strengthen the financial and cor- shareholders or related parties that would endanger the porate sectors. An enterprise capitalization fund was liquidity or solvency of the banks. It also prohibited created to provide preferential credit for purchasing the use of funds to finance speculative takeovers and new shares or convertible bonds in highly leveraged cross-ownerships in excess of legal limits and imposed but viable companies. A comprehensive and volun- sanctions for assisting in tax evasion or for supplying tary mechanism was established to restructure private misleading information about the financial condition corporate debt. An innovative capitalization fund was of institutions. The decree also applied sanctions for created in which banks could invest up to 7 percent- violating prudential regulations and for engaging in age points of their legally mandated reserves in cen- unorthodox business practices, and made management tral bank bonds to rediscount loans at preferential subject to criminal prosecution and severe penalties. interest rates for the purpose of selling new bank The decree also allowed the government to nation- shares. Preferential credits were provided to buyers of alize any insolvent financial institution without com- financial institutions' repossessed assets. Lower pensating those responsible for the failure so long as reserve requirements and higher interest rates on 104 BANK RESTRUCTURING forced investments boosted the profitability of finan- the depletion of international assets. Nominal peso cial intermediaries (Vargas and others 1988). devaluation reached 43 percent against the dollar Though helpful, these initiatives did not avert fur- between March and December 1985, amounting to a ther bank failures. In 1983 Colombia's second largest real devaluation of 28.5 percent in 1985 and 38.0 per- commercial bank, Banco de Bogota, sustained serious cent between 1985 and 1986. Prices increased only loan losses from its Panamanian operations, sovereign moderately, restrained by excess industrial capacity, a debt exposure, and loan concentration. It also suffered good harvest, subsidized foodstuffs from Venezuela, from losses incurred by the controlling Grupo Bolivar, lower international commodity prices, falling interna- which was engaged in an expensive defense against an tional interest rates, and continued tight wage and unfriendly takeover by Grupo Sarmiento-Angulo. monetary policies. Reluctant to nationalize the bank because of its size, the Although this reversal of fortune provided authori- Monetary Board allowed competing financial groups to ties with the resources to undertake bank restructur- endorse the bank's shares to a trust fund administered ings, monetary and exchange rate policies also by a state-owned bank. The authorities wanted to avoid increased pressures on the banking system. As observed further concentration in the banking sysrem and hoped in other banking crises, real interest rates on bank to obtain additional capital, which would resolve the loans continued to be high, reflecting distress borrow- ownership struggle. The central bank provided a cash ing by enterprises, higher risks, and tight liquidity. advance as partial payment of the estimated value of the Moreover, the depreciation of the peso created large shares being put into trust. The advance would be can- foreign exchange losses for banks and enterprises with celed once the trust fund sold the shares to the public, net foreign debt, increasing not only their liabilities but in the meantime the two competing groups but also their debt servicing burden. At its peak in received a subsidized credit. The government appointed 1985, past-due debt in foreign currency amounted to a new board and reconfirmed the existing management. 46 percent of total past-due loans. The scheme did not work as planned, however, because the shares in trust for sale to the public were too expen- Banco de Colombia sive given the bank's inability to pay dividends. At this point Colombian authorities faced several Amid its economic stabilization efforts, the govern- challenges. They had to undertake an economic stabi- ment was forced to rescue Colombia's oldest and lization program. They needed to obtain additional largest commercial bank, Banco de Colombia (18 per- external credits to roll over the external public debt. cent of total bank assets), which was embroiled in the And they had to replenish the banks' lost capital, fraudulent investment practices of its controlling preferably without further nationalizing and without Grupo Grancolombiano. This was the most complex bailing out owners and managers as they had in the bailout, involving 168 financial and nonfinancial case of Banco de Bogota. companies-real and on paper. Public savings invest- ed in two investment funds managed by the group Resumedgrowth had been used to finance costly takeovers of large manufacturing companies in the food processing, Coffee prices rebounded in the second half of 1985 cement, and steel sectors, all sensitive to the business after a well-implemented adjustment plan introduced cycle. The fund managers manipulated the publicly in late 1984 began to restore economic growth (Garay quoted price of the underlying companies in 1981 by and Carrasquilla 1987). A drastic reduction in the selling shares to group-related companies for less than public sector deficit was followed by a rapid correction the funds had paid for the shares in 1980. The result- of the overvalued peso through an accelerated crawling ing run on the funds led Banco de Colombia to peg. The government budget moved from a deficit of extend large credits to group-related companies so 4.2 percent of GDP in 1985 to a surplus of 0.2 per- that they could repay or purchase additional shares cent in 1986. Government revenues increased by from the funds. When the related companies could almost 50 percent in 1985 because of improved collec- not repay the bank, Colombian authorities intervened tion, a broader tax base, and new taxes (notably an to stem a liquidity crisis. Grupo Grancolombiano was import surcharge). An improved current account sur- forced to absorb the losses in the investment funds plus of 1.3 percent of GDP in 1986 relieved and compensate defrauded investors. In 1983 the Colombia's reliance on external financing and curbed Superintendency of Banks replaced the bank's man- Structural Weaknesses and Colombial Banking C isi, 1982-88 105 agement, and subsequently the government forced the fiduciary responsibility and fraud. But other failures controlling shareholders to establish a trust fund shared some of the same characteristics. The share of pledging the shares of the group's holding company, substandard bank portfolios in the total portfolio Cingra, to back the loans made by Banco de rose from an average of 7 percent during 1975-80 to Colombia to group-related companies. Banco de 24 percent in 1984. Loans in foreign currency tradi- Colombia was then able to reverse some of its loan tionally accounted for a large share of all past-due loss provisions and ended 1983 with a profit. loans, but their share and value jumped in 1984 to This intervention, however, merely postponed roughly ten times the level in 1981, and nearly dou- Banco de Colombia's day of reckoning. External audi- bled again in 1985. In 1985 officials found that the tors appointed in 1983 found massive irregularities past-due foreign exchange loans of insolvent banks stemming from illegal and unaccounted foreign exceeded $450 million (almost half of all past-due exchange transactions, a large concentration of delin- loans), putting at risk Colombia's dwindling net quent loans that had been rolled over to related, international reserves and threatening the success of unprofitable, or nonviable enterprises, and lack of the urgent new international syndications. For exam- provisions against expected losses from unsecured, ple, two of the Panamanian offshoots had loan expo- nonperforming loans (particularly to textile, automo- sures exceeding $1 billion (Caballero 1988). tive, and cotton-producing companies, all suffering from the economic downturn). By September 1985, The guaranteefind 42 percent of the bank's total loans were nonperform- ing (equivalent to about 2.5 percent of GDP). Banco It was against this background that Congress, at the de Colombia's small net capital produced a gearing end of 1985, finally legislated a guarantee fund, which ratio of 38:1, nearly four times the legal limit of 10:1, provided a means for resolving institutions short of and its net financial margin was negative. Moreover, full-scale nationalization or liquidation. One mecha- the different financial companies of the Grupo nism was "officialization," with private management Grancolombiano held one-third of all the external but de facto official ownership. Under officialization debt of Colombia's financial system. shareholders were able to retain a minority stake while The government initially assumed a substantial the fund took majority control. The government's amount of overdue dollar-denominated and domestic deposit guarantee was proportional to its capital stake. public sector debts, ordered the sale of group enter- Once the capital stake reached 51 percent the funds to prises to repay the bank, and attempted to reschedule officialize the bank came from either the fund or the Banco de Colombias foreign debt. The central bank central bank, but there was no budgetary outlay. The provided liquidity support, but attempts to recapital- fund also helped rescue financial institutions by sub- ize the bank failed when Banco de Colombia's accu- scribing additional equity, extending loans, purchas- mulated losses reached more than twice its equity by ing real or financial assets, arranging repurchase agree- the fourth quarter of 1985. The bank was national- ments, assuming domestic and external liabilities, ized in 1986 after losses had amounted to $356 mil- offering guarantees, insuring deposits, and administer- lion, of which 69 percent was attributable to its ing viable institutions. The fund was prohibited from Panamanian operations. The government used the assuming losses when purchasing bank assets and did Spanish "accordion" technique to shrink capital to one not receive a capital endowment. It was placed under centavo per share and then capitalized the 3 billion the oversight of the Superintendency of Banks, with a pesos ($15 million) of outstanding loans from the view to eventual self-financing through insurance pre- central bank. In addition, it subscribed 60 billion miums. In practice the fund relied on fees, insurance pesos ($300 million) of callable capital, and the new premiums, bond subscriptions by banks and develop- guarantee fund supplied additional soft credits. ment finance companies, and loans from the National Because of continued mounting losses resulting from Coffee Fund, guaranteed by the central bank. rapid peso devaluation, the callable capital did not have a positive financial impact. In 1988 the central Further insolvencies bank issued dollar-denominated bonds to hedge the bank's rescheduled foreign exchange liabilities. Three small banks were officialized toward the end By far the most costly of the resolutions, Banco of the crisis in 1986. Banco de los Trabajadores, de Colombia highlighted the worst cases of abuse of with 0.4 percent of total banking system assets, 106 BANK RESTRucrURING suffered recurrent liquidity problems between 1983 returned to surplus, aided by higher coffee prices and 1985. The guarantee fund intervened because and petroleum and coal exports. The fiscal position more than 40 percent of the bank's loan portfolio went from a deficit of 3.5 percent of GDP in 1985 was nonperforming, its financial margin was insuf- to a surplus of 0.6 percent in 1986. With confidence ficient to cover operating costs, and controlling and private investment restored, growth recovered to shareholders failed to raise sufficient new capital. 5.8 percent in 1986 and 5.3 percent in 1987. These Banco Tequendama, accounting for 0.7 percent of environmental factors, aided by bank restructuring, banking system assets, was more problematic further reduced the stock of nonperforming loans because it was a Colombian-Venezuelan joint ven- and restored bank profitability. By 1988 substandard ture. More than half its business was denominated loans as a percentage of total loans fell to 12 percent, in foreign currency and was concentrated in compared with a peak of 25 percent in 1984. The financing trade between Colombia and Venezuela. ratio of substandard bank loans to capital and Credit lines shrank as a result of the 1982 debt cri- reserves dropped to just under 43 percent in 1988, sis and Venezuela's balance of payments difficulties, compared with 144 percent in 1984. Between 1982 exacerbating the inherent instability of the bank, and 1987 the government provided more than 150 which relied on costly and unstable deposits from billion pesos-about $750 million at the 1986 Colombian agencies. Normal liquidity support and exchange rate, or 2 percent of GNP-to recapitalize emergency interim subsidies failed to stabilize the financial institutions. bank. Nonperforming assets were roughly 65 per- On the regulatory front, the Superintendency cent of total assets, and by the end of 1985 the of Banks revised rules covering loans to related bank had lost almost three times it capital. Its lim- parties, set limits on loans to single borrowers, and ited scope of business and the unknown extent of improved bank capital and reserves by transferring problems in its partially owned Venezuelan sub- substandard or nonperforming loans to the guar- sidiary and its office in Curacao made the restruc- antee fund. Raising the loan loss provisioning turing difficult. Finally, Banco del Comercio also requirements also improved capital ratios. had serious management problems and could not Institutional and legal mechanisms were estab- compete adequately. It engaged extensively in lished to facilitate future bank restructurings and intrabank deposits and guarantees in foreign cur- assistance to distressed financial intermediaries and rency with foreign subsidiaries in Panama and nonfinancial enterprises, and to protect the Montserrat. Between 1982 and 1985 nonperform- deposits ofsmall savers. ing assets increased from 7 percent of total assets to The Colombian case was not a generalized 20 percent. When a recapitalization effort failed, bankin,, crisis. It involved only seven of twenty-six the guarantee fund pared down the balance sheet, banks and affected at most one-quarter of bank purchased bad assets (16 billion pesos) with a bond assets, compared with 60 percent of bank assets in yielding 24 percent a year, and subscribed new cap- the Chilean case, for example. The situation unrav- ital of more than 6 billion pesos. eled over a relatively long period (1982-87), and In addition, between 1982 and 1984 the the equity base and bank profitability had already Superintendency of Banks intervened in a number weakened before then. Although the downturn of of consumer finance companies, and in 1987 it the business cycle might have precipitated the bank intervened in or liquidated five development failures, the high concentration of banks' loan port- finance companies and eight finance companies. folios was the most important reason for the insol- Most of these were part of financial conglomerates vencies. Contributing factors inciude weak supervi- built around insolvent commercial banks. Other sion before the crisis, along with inadequate legal finance companies also received financial support and institutional arrangements and excessive gov- from the guarantee fund, although they were not ernment regulation in the form of low-yielding, officialized. large forced investments and relatively high legal reserve requirements. Fraud, bad credit decisions, Conclusion undercapitalization of banks and their corporate customers, the failure of overseas affiliates and sub- The 1984 economic adjustment program was quite sidiaries, and increasing operational and financial successful. By 1986 the current account had costs also played a part. Structural Weaknesses and Colombia' Banking Crisi, 1982-88 107 Notes Carrizosa, Mauricio. 1986. Hacia la Recuperacion del Mercado de Capitales en Colombia. Bogota: Bolsa de Bogoti. This chapter is an abbreviated version of Montes-Negret (1990). Colombia. 1986. "La Nacionalizaci6n del Banco de Colombia y 1. Forced investments amounting to 16.5 percent of bank el Grupo Grancolombiano: Antecedentes y Justificaci6n." loans were channeled to agriculture and represented a substantial Economia Colombiana, Revista de la Contraloria General de subsidy to the state. Banks held securities of the central bank, la Republica. but these yielded only 8 percent at a time whien nominal rates Garay, Luis, and Alberto Carrasquilla. 1987. 'Dinamica del ranged from 28 to 38 percent. In 1984 the yield was raised to 15 Desajuste y Proceso de Saneamiento Economico en percent, but the additional 7 percent return had to be used for Colombia en la Decada de los Ochenta." Ensayos Sobre loan loss provisioning. By booking loans abroad, banks could Politica Economica 11. Banco de la Republica, Bogota. reduce the base of computation for forced investments. Foreign IMF (International Monetary Fund). Various years. currency lending was finaUy exduded from the base of computa- International Financial Statistics. Washington, D.C. tion in December 1987. JaramiUlo, J. C. 1977. "Sector Externo." LIV Informe Anual del 2. In addition, Colombian banks were authorized to pur- Gerente a la Junta Directiva. Segunda Parte, Banco de la chase trade-related, legally registered loans of their overseas sub- Republica, Bogoti. sidiaries. These purchases were funded with credits provided by . 1982. "La Liberaci6n del Mercado Financiero." Ensayos the same international bank that had originally extended the Sabre Politica Economica I (March). Banco de la Republica, loan to the Panamanian offshoot. This enhanced the influence Bogoti. of international banks in negotiating syndicated loans to Lee, Martha. 1983. "Comentarios a la Reforma del Sistema Colombia. Financiero." Ensayos Sobre Politica Economica 4(December). 3. The external debt doubled from $6.8 billion in 1980 to Banco de la Republica, Bogoti. $12.3 billion in 1984. Montes-Negret, Fernando. 1990. "An Overview of Colombia's Banking Crisis: 1982-88." Background paper prepared for World Development Report 1989. World Bank, Washington, References D.C. Montes-Negret, Fernando, and Alberto Carrasquilla. 1986. Arango, M. A. 1988. "El Sector Industrial y las Deudas de "Sensibilidad de la Tasa de Interes Activa de los Bancos a Dudoso Recaudo del Sector Fianciero." DCI, Banco de la Cambios en los Parmetros de Politica y Estructura." Ensayos Republica, Bogoti. Sobre Politica Economica I0(December). Banco de la Banco de la Republica. 1986. Fondo de Garantias de Instituciones Republica, Bogoti. Financieras, Antecedentes, y Normas Basicas. Bogoti. Ortega, Francisco J., and Rudolf Hommes. 1984. 'Estado y - 1988a. "La Recuperacion del Sistema Financiero." Notas Evoluci6n de la Capitalizaci6n de Bancos y Corporaciones Editoriales, Revista del Banco de la Republica 58(693). Financieras." Banca y Finanzas 1 86(December). - 1988b. "Orienracion del Sistema Financiero y sus Vargas, H., M. Lee, F Montes-Negret, and R. Steiner. 1988. "La Resultados." Notas Editoriales, Revista del Banco de la Evolucion del Sistema Financiero en los Ultimos Afios." Republica 727(May). Revista del Banco de la Republica 731. Bogoti. Caballero, Carlos. 1988. 'La Experiencia de Tres Bancos World Bank. 1989. World Development Report 1989: Financial Colombianos en Panama." Coyuntura Economica 18(l). Systems and Development. New York: Oxford University Press. CHAPTER 7 Malaysia's Bank Restructuring, 1985-88 Andrew Sheng Sharp deflation in the Malaysian economy in twelve merchant banks, and seven discount houses. 1985-86 precipitated large losses among a number of There are also several specialized financial intermedi- financial institutions, most notably four (of thirty- aries, including development banks, building soci- nine) commercial banks, four (of forty-seven) finance eties, provident funds, and insurance companies. companies, thirty-two (of thirty-five) deposit-taking Foreign banks account for a quarter of total bank cooperatives, and thirty-three illegal deposit-taking loans and one-fifth of total bank deposits. At the end institutions. Losses among the unregulated deposit- of 1987 assets of the financial system totaled M$203 taking institutions were relatively small; those sus- billion, equivalent to 269 percent of gross national tained by the banks and finance companies-which product (GNP) (table 7.1). were supervised by the central bank-and the The degree of monetization, measured as the ratio deposit-taking cooperatives were far greater. Between of broad money to GNP, was 0.94 in 1986, compar- 1985 and 1987 those forty institutions lost roughly ing favorably with that of the Republic of Korea 3.1 billion ringgit (M$), equivalent to 4.7 percent of (0.94), the United States (0.82), the United gross domestic product (GDP) in 1986. Losses were Kingdom (0.77), and Australia (0.73). Malaysia also distributed more heavily among the lightly supervised has active foreign exchange and capital markets and a deposit-taking cooperatives, whereas bank losses liberal exchange control regime that allows free con- amounted to only 2.4 percent of total deposits in vertibility of the ringgit with all major foreign curren- 1986. This suggests that deficiencies in oversight and cies. The Kuala Lumpur foreign exchange market has delay in recognizing the magnitude of the portfolio an average monthly turnover of M$20 billion, and problems of nonbank financial intermediaries were the Kuala Lumpur stock exchange-which reported a chiefly to blame for this short-lived crisis. market capitalization of M$97.1 billion ($37.8 bil- Restructuring, while costly, was aided by sound struc- lion) in the first half of 1988-is one of the most tural adjustment in the Malaysian economy, specifi- active stock exchanges in the developing world. In cally fiscal retrenchment and currency depteciation. 1986 equities amounted to 24 percent of total finan- Economic recovery brought a return to enterprise cial assets, about the same proportion as in Canada profitability, a high level of national savings, and ade- but far greater than in Korea (4 percent) or Taiwan, quate foreign exchange reserves to cushion the system China (9 percent). against international economic shocks. Licensed deposit-taking institutions, which accounted for 70 percent of total financial system Structure of the Financial System assets at the end of 1987, are under the direct super- vision of Bank Negara, the central bank. The central Malaysia has a relatively broad and deep financial sys- bank has two bank supervision departments, one for rem consisting of thirty-nine commercial banks off-site monitoring and the other for on-site exami- (twenty-two domestic, sixteen foreign, and one nations. Banks are constrained from acquiring shares Islamic), forty-seven licensed finance companies, without the central bank's permission and from 109 110 BANK RESTRUcruRING Table 7.1 Assets of the Malaysian financial system (billions of ringgit) Annual change Assets at Share of total assets Type of institution 1986 1987 end-1987 (percent) Banking sysrem 11.3 11.0 140.6 69.3 Monetary institutions 9.0 9.6 110.0 54.2 Central bank 3.8 3.8 24.2 11.9 Commercial bank? 5.2 5.8 85.8 42.3 Nonmonetary institutions 2.3 1.4 30.6 15.1 Finance companies 1.8 1.6 21.3 10.5 Merchant banks 0.1 0.0 6.3 3.1 Discount houses 0.4 -0.2 3.0 1.5 Nonbank financial intermediaries 5.4 6.4 62.4 30.7 Provident, pension, and insurance funds 4.6 4.8 42.0 20.6 Employees Provident Fund 3.8 3.8 32.3 15.9 Other provident funds 0.4 0.3 3.5 1.7 Life insurance funds 0.4 0.6 4.7 2.3 General insurance funds 0.0 0.1 1.5 0.7 Development finance institutions' 0.3 0.2 4.6 2.3 Savings institutions' -1.2 0.4 7.2 3.6 Other financial intermediariesd 1.7 1.0 8.6 4.2 Total 16.7 17.4 203.0 100.0 a. Includes the Islamic bank. b. Includes all development banks. c. Includes the National Savings Bank and deposit-taking cooperatives. d. Includes budding societies and capital market funds. Source: Bank Negara Malaysia. making property investments except for operating more than doubled, from 336 to 770 between 1970 needs. The central bank also supervises the forty- and 1986, and employment in banking quadrupled, seven licensed finance companies, traditional exceeding 40,000 people at the end of 1985. The providers of consumer and housing finance. Other larger domestic banks also began to expand abroad, financial institutions are supervised by various agen- boasting forty-five foreign branches at the end of cies. For example, all deposit-taking cooperatives are 1986 and foreign assets of M$11.2 billion (15 per- under the authority of the Ministry of National and cent of all assets). Rural Development, while some development banks Nonbank financial intermediaries accounted for are supervised by the Ministry of Public Enterprises 31 percent of the assets of the Malaysian financial and others by the Ministry of Finance. Although the system at the end of 1987. The bulk of these assets system is comprehensive, it applied varying degrees was held by provident and pension funds and the of supervisory stringency to different types of depos- National Savings Bank, whose assets were primarily itory institutions and failed to rein in a handful of in government bonds or deposits with the banking illegal deposit-taking institutions, whose number system. During the 1970s, however, deposit-taking increased rapidly during the economic prosperity of institutions and some deposit-taking cooperatives the 1970s and early 1980s. emerged as quasi finance companies, collecting As the economy prospered during the early 1970s deposits to finance a range of business activities while the banking system expanded rapidly. Incomes rose remaining outside the supervisory authority of the with favorable commodity prices and the advent of central bank. The unlicensed deposit-taking institu- substantial oil revenues, and national savings tions acted as pawnbrokers or credit and leasing com- increased sharply, contributing to the rapid growth of panies and took deposits from the public illegally to savings in the banking system. Asset growth also fund their lending and investment activities. accelerated, from an average annual rate of 19.1 per- Cooperatives, by contrast, operated under a legal cent in the first half of the 1970s to 24.4 percent in loophole permitting them to take deposits from their the second half. Growth stayed at double-digit rates members (who simply paid a small fee). The until 1984, falling sharply to 7 percent by 1987. The Department of Cooperative Development, which was value of loans grew by an average of 23 percent a year responsible for supervising 3,000 cooperatives, was during 1975-84. The number of bank branches not equipped to monitor the thirty-five deposit- Malaysias Bank Restructmring, 1985-88 111 taking cooperatives, whose complex operations GDP growth peaking at 9.3 percent in 1979, but expanded rapidly during the 1970s. These coopera- inflation iose to an average of 6 percent, peaking at tives had diversified out of traditional consumer 11.1 percent in mid-1981 (table 7.2). finance and had acquired more than I million mem- During 1979-81 Malaysia witnessed its sharpest bers, 600 branches, and total deposits exceeding M$4 deflation since the post-Korean War recession of billion. There was no lender-of-last-resort facility for 1952-53. In 1981, in light of its strong financial posi- these cooperatives, which frequently did not observe tion and to counteract the impact of the severe inter- the mandated 25 percent liquid deposit ratio. national recession that had begun il 1980, the gov- ernment embarked on a countercyclical policy intend- Macroeconomic Background ed to build up infrastructure and the industrial base. The policy relied primarily on external borrowing. Malaysia is a small, open, middle-income, oil-export- Federal government development expenditure rose ing country. The Malaysian economy historically has from an average of M$3.0-4.0 billion a year during been sensitive to changes in the terms of trade (figure 1977-79 to a peak of M$11.5 billion in 1982, result- 7.1); its external trade to GNP ratio was 130 percent ing in a deterioration in the federal fiscal deficit to in 1987. Well-endowed with natural resources, it is 18.7 percent of GNP (figure 7.2). The terms of trade the largest exporter of rubber, palm oil, tin, and trop- deteriorated by 10.2 percent as commodity prices fell ical hardwoods; an important producer of cocoa and in 1981-82, and the current account went from near pepper; and a net exporter of crude oil and natural balance in 1980 to a deficit of M$8.4 billion, or 6.0 gas. Manufacturing accounted for 23 percent of percent of GNP, in 1982. The net result was a tripling GDP and manufactured exports for 45 percent of in the foreign debt from M$10 billion in 1980 gross exports in 1987. In 1988 GNP reached $31.6 (matched by equal foreign exchange reserves) to billion ($1,869 per capita). M$31.8 billion at the end of 1983 (49 percent of The government, considered a model of cautious GNP). The debt service ratio worsened at the same economic management, has a stated policy of main- time, as international bank lending declined in taining a current account surplus, strong external 1982-83, making a countercyclical policy financed by reserves, and low inflation. In the 1960s Malaysia external borrowing unsustainable. enjoyed steady real GDP expansion averaging 5.2 The government undertook a stringent structural percent a year, while inflation remained at less than I adjustment program in 1983 to reduce the fiscal and percent a year. During the 1970s the economy grew balance of payments deficits and to control the by an average 8 percent a year in real terms, with growth of external debt, especially the expenditure and debt incurred by the nonfinancial public enter- Figure 7.1 Terms of trade and current account prises. Temporary recovery in commodity prices dur- deficit, 1970-88 ing 1983-84 and the resulting revival of GDP (percent change) growth to 7.8 percent in 1984 contributed to the 30 - success of the adjustment effort. The overall fiscal 25 - Terms of srade deficit fell to 9.5 percent of GNP in 1984. 20 - Beginning in the second half of 1985, however, 15 _ Z \ I\ global commodity prices dropped sharply, and they lo - 8 Z \ A continued to fall in 1986. Crude oil prices fell 62 percent from their peak, and palm oil prices fell 57 5 ~ | \ \ A X \ percent, to below the cost of production. Terms of 0 A trade worsened by 4.5 percent in 1985 and by 15.5 -5 percent in 1986. Total export income fell by 2.6 per- -10 cent in 1985 and 5.9 percent in 1986. Real GDP -15 Current amoun growth became negative (-1.0 percent) in 1985 for -20- deficit/GNP the first time since 1975, while nominal GNP -25 1 1 I I I I declined by 2.9 percent in 1985 and 7.9 percent in -25 "~' hA°N95NE95) 9ftE9SiNq eqSq0.seqNq)¢sE5& #1986. Cutbacks in public spending exacerbated the decline in GNP, which for the first time declined for Source: IMF, various years. two consecutive years in current terms. As revenue 112 BANK RESTRUCTURING Table 7.2 Macroeconomic indicators, 1970-89 (percent unless otherwise specified) Terms of Exchange rate Real growth Current tradea Inflation Government Deposit (ringg,its Year in GDP account/GNP (1980=100) rateb balance/GNP interest rate U.S. dollar) 1970 6.0 0.1 77.0 1.8 -3.7 - 3.1 1971 5.8 -0.8 70.8 1.6 -7.7 - 2.9 1972 9.4 -1.7 59.4 3.2 -9.0 - 2.8 1973 11.7 0.6 72.3 10.6 -5.5 - 2.5 1974 8.3 -2.4 81.6 17.3 -6.0 - 2.3 1975 0.8 -2.2 68.8 4.5 -8.4 - 2.6 1976 11.6 2.1 76.9 2.6 -7.1 5.5 2.5 1977 7.8 1.3 84.7 4.8 -8.5 5.2 2.4 1978 6.7 0.3 81.3 4.9 -8.0 5.1 2.2 1979 9.3 2.1 99.0 3.7 -8.3 5.5 2.2 1980 7.4 -0.6 100.0 6.7 -13.8 6 2 2.2 1981 6.9 -4.5 86.4 9.7 -19.8 9.7 2.2 1982 5.9 -6.0 80.6 5.8 -18.7 9.8 2.3 1983 6.3 -5.3 79.9 3.7 -14.0 8.0 2.3 1984 7.8 -2.3 87.0 3.9 -9.5 9.5 2.4 1985 -1.0 -0.9 79.2 0.3 -7.9 8.8 2.4 1986 1.2 -0.2 64.8 0.7 -11.2 7.2 2.6 1987 5.3 3.5 77.4 0.9 -8.2 3.0 2.5 1988 8.8 2.1 77.8 2.0 -4.5 3.5 2.7 1989 8.8 -0.2 77.6 2.8 -5.1 4.6 2.7 - Not availablc. a. Ratio of the index of average export prices to the index of average import prices. b. Consumer price index. Source: IMF, various years; World Bank data. declined, the overall fiscal deficit increased again in about an adjustment in the balance of payments. The 1986, to 11.2 percent of GNP, after the modest ringgit depreciated by 16.7 percent against its com- recovery in 1985. Again, the government cut spend- posite from September 1985 (Plaza Accord) to the ing, bringing development expenditure down to end of 1986. Despite the rise the depreciation caused M$3.2 billion in 1987, the lowest level since 1979. in import costs, inflation fell sharply to 0.7 percent in The government also decided to allow the 1986, compared with 9.7 percent in 1981. exchange rate to depreciate freely in order to bring On the real side of the economy, major shifts occurred in the direction of domestic investment dur- ing the halcyon years of the second half of the 1970s Figure 7.2 Fiscal deficit and nominal GNP and the early 1980s-when real GDP grew at an aver- (percent change) age rate of 8.6 percent, the exchange rate was strong, and foreign investment flowed in. Exchange rate 25 ~\ appreciation brought a shift toward the nontradable 20 n GNP A sector. Strong corporate profits and speculation drove 15 \_r \ the Kuala Lumpur stock exchange composite index from 100 in 1977 to a peak of 427 in early 1984. 10 - Bank lending to property acquisition and devel- 5 _ opment was stimulated by the influx of foreign work- ers allied to inward investment in the petroleum o \ / industry and other areas of the economy. Demand -5 - for office and residential rental properties caused -o -1 Fiscal drfci/GNP prices to triple between 1975 and 1983. In 1986 the number of completed property developments in -15 ~\ /Kuala Lumpur alone was five times the number in -20 L 1983. In 1983, at the height of the property boom, q.b qM qy q:9 e° eX 9b9prime commercial property in Kuala Lumpur was Sourre: IMF, various years. valued at M$450 a square foot, with a monthly rental Malayjias Bank Restructuring, 1985-88 113 of M$3.20 a square foot, but by 1986 overbuilding cent in 1986. For the corporate sector, excluding and forced selling drove values down to M$180 and deposits from the oil sector, growth in deposits was rental prices to M$0.90 a square foot. By the end of -19.2 percent at that time. Faced with threats to their 1986 total new banking system loans to real estate survival, many cash-strapped enterprises engaged in amounted to 55 percent of total new loans, com- distress borrowing, paying exceptionally high rates for pared with 32 percent in 1980. The share of property funds in order to meet temporary liquidity shortfalls. loans in total bank loans outstanding rose from 26 By 1985 the real rate of borrowing had risen to an percent at the end of 1980 to 36 percent at the end unprecedented 11.8 percent, compared with 2.0 to of 1986. 3.0 percent in 1981 (figure 7.4). The growth of bank The 1985 slowdown brought a sharp contraction loans moderated to 6 percent at the end of 1986, in cash flow for many Malaysian companies. Incomes compared with 37 percent in 1980. declined dramatically in the face of falling commodi- The combination of tight liquidity, low inflation, ty prices, weak export sales, and poor domestic and a dramatic decline in share and property prices demand. Companies fell behind in debt repayments, brought into focus the financial overcommitments of while their committed outflows, particularly in the many entrepreneurs. Those who had built up their property sector, continued to drain resources. Share gross assets through speculation in shares and proper- prices dropped 60 percent from their 1984 peak, and ty, financed through excessive gearing, were caught in total market capitalization fell 44 percent, to M$46.7 a triple squeeze: they faced a sharp decline in income billion in 1986. Inflation had fallen to almost zero, flows, a collapse in asset values, and a rise in the cost and the growth rate of narrow money (Ml) became of debt service. As the slowdown in the economy negative by the end of 1984. Broad money (M3) gathered momentum during 1985, tightening liquid- growth also slowed sharply to 3.6 percent by 1987, ity combined with bad economic news to cause ner- the slowest rate in fifteen years (figure 7.3). vousness among depositors. As liquidity began to tighten in the early 1980s with the decline in commodity prices, the banking The Impact of Recession on the Financial System system's loans-to-deposits ratio rose from an average of 75 percent in the 1970s to 90 percent in 1983. The dangers of rapid expansion into new areas of Private sector expenditure initially was slow to adjust growth without fully understanding the implications to the declining income, resulting in a drawdown of first became evident in 1982, when Bumiputra savings. Annual growth of deposits in the banking Malaysia Finance (BMF) nearly failed. BMF, the system fell sharply, from 20 percent in 1984 to 4 per- Hong Kong finance company subsidiary of Malaysia's Figure 7.3 Money supply and real GDP growth, Figure 7.4 Real interest rate and real GDP 1970-88 growth, 1980-88 (percent change) (percent) Real interest raze 40- 12 Ml I 35- I10 30- 9 M3 ~~~~~~~~~~~~~~8 25 7 20- 6 2 5 Rel D Irwd So0rcr: Bank Negara Malaysia. Soarce: Bank Negara Malaysia. Source: Bank Negara Malaysia. Source: Bank Neg2ra Malaysia. 114 BANK RESTRUCTURING largest domestic bank, had extended M$2.4 billion in Figure 7.5 Banking system pretax profits, loans to several Hong Kong corporations and indi- 1979-88 viduals. When borrowers defaulted on the loans, the (millions of ringgit) parent bank absorbed the losses, and Malaysia's 800 national oil corporation, Petronas, became the largest 700- shareholder by injecting fresh capital into the bank. 600- Petronas bought M$1.26 million in loans from the 500- bank; the remaining M$1.0 billion in nonperforming 400- loans was written off in 1983, constituting one of the 300 - largest losses in the history of Malaysian banking. But 200- the warning that the BMP episode signaled of events 100- to follow went unheeded, especially when economic 0- performance rebounded during 1983-85. -100 -200 Banks -300 -400 By mid-1985 financial problems had reemerged. In 14`9 10" s9t s9 9% s9 1$ s9 I* o9¢ July the failure of Overseas Trust Bank in Hong Source: Bank Negara Malaysia. Kong prompted rumors about problems in a large domestic Malaysian bank, sparking a run on its aged only 1.0 to 1.5 percent of total loans. Foreclosed branches. Although the panic was quelled, this was property could easily be sold at values higher than the the first such incident in Malaysia since the mid- outstanding loans that financed them. However, with 1960s. In September Setia Timor Credit and Leasing, M$37.3 billion wiped off stock market capitalization a small leasing company engaged in illegal deposit- and property prices falling under pressure of distress taking, failed. The first failure among the deposit- sales, the banks faced rising levels of nonperforming raking institutions, it sparked isolated runs on several loans and bad and doubtful debts in 1985-86. licensed finance companies. The panic withdrawals To identify the impact of the recession on the were halted quickly. But in December 1985 Pan- banking system, the central bank in 1985 introduced Electric, a large, publicly listed company in uniform guidelines for financial institutions on the Singapore, collapsed, causing widespread panic that treatment of nonaccrued income, or interest-in-sus- led to an unprecedented three-day closure of the pense. At the beginning of the crisis in 1984 bad debt Kuala Lumpur and Singapore stock exchanges. A run provisions and interest-in-suspense were only M$2 then occurred on a medium-size finance company billion, or 3.5 percent of the total loans of the bank- associated with businessmen-later arrested-who ing system (table 7.3). At 2.3 percent, bad debt pro- also had interests in Pan-Electric. Bank Negara pro- visions were not out of line with international stan- vided market liquidity, the run subsided, and an dards, such as those for the major British clearing experienced professional was appointed to manage banks. By the end of 1988, however, total interest-in- the finance company. Nevertheless, sporadic bank suspense and bad debt provisions amounted to runs throughout 1986 culminated in the failure of M$11.7 billion, or 14.5 percent of the gross loans of the twenty-four deposit-taking cooperatives in July the banking system. These provisions covered and August. At that point there was danger of sys- approximately 47 percent of total nonperforming temic financial failure, a possibility that persisted for loans, a relatively high level compared with the aver- the remainder of 1986. age provisions of 25 to 40 percent for the interna- The traumatic events of 1985-86 were as much a tional banks exposed to highly indebted countries. shock to bank management as to their borrowers. The fastest-rising component was interest-in- Throughout the previous two decades, the Malaysian suspense, indicating private sector difficulties in ser- banking system had enjoyed rising profits, with pre- vicing debt during the recession. Between 1984 and tax profits peaking in 1984 (figure 7.5). During the 1988 interest-in-suspense grew at an average annual period of uninterrupted growth and rising property rate of 73 percent, from M$664 million to M$5.9 (and hence security) values, bad loans were negligible. billion, while bad debt provisions rose at a slower rate As late as 1983 specific and bad debt provisions aver- of 42 percent, from M$1.3 billion to M$5.7 billion. Malaysial Bank Restructsring, 1985-88 115 Table 7.3 Outstanding interest-in-suspense and and 1985. By 1987 losses had risen to M$252 mil- bad debt provisions, 1984-88 lion. The central bank assumed control of four com- (millions of ringgit) panies unable to inject new capital to cover their loss- Interest-in- Bad debt Total asashare of es. Many of the others had strong shareholders (one- Year suspense provisions total loans (percent) third were subsidiaries of banks) and were able to 1984 664 1,346 3.5 cover their capital deficiencies. 1985 1,500 2,493 5.6 1986 2,844 4,056 97 NonbankJinancial intermediaries 1987 4,242 5,188 12.9 1988 5,932 5,722 14.5 Source: Bank Negara Malaysia, various years. Setia Timor's inability in September 1985 to meet its deposit withdrawals was the first sign of serious These large provisions ultimately cut bank profits financial distress among deposit-taking institutions. across the board. Even the best-managed foreign Press reports stating that directors and staff of a few banks showed substantially lower profits and several deposit-taking institutions had absconded with funds reported large losses, which they covered through highlighted the problem of illegal deposit-taking fresh injections of capital. Four medium-size domes- institutions, which attracted deposits by paying tic banks were the most severely affected, mainly by interest rates of 24 to 36 percent a year compared heavy losses from their involvement in the property with the 8 to 10 percent offered in the formal bank- sector. All four banks were relative newcomers: one, ing sector. During 1985-87 the central bank investi- whose license had been issued in 1979, was majority gated thirty-three failed, illegal deposit-taking insti- state owned; two had been restructured within the tutions involving 8,000 depositors and total deposits previous ten years from branches of foreign banks; of M$49 million. and the fourth was a joint-venture bank. The four Nervousness among depositors spread to the banks collectively had expanded their loan base deposit-taking cooperatives as well. In early 1986 aggressively between 1980 and 1985, increasing their Kosatu, with M$156 million in deposits from 53,000 market share from 7.2 percent to 8.8 percent by the members in sixty-seven branches, suspended pay- end of 1985. But mismanagement, fraud, over- ments. The thirty-five deposit-taking cooperatives, stretched managerial resources, and poor internal which had started in the 1960s as a slow-growing, procedures and controls had led to lax control over grassroots network, diversified and grew rapidly dur- costs and a rapid rise in nonperforming loans. The ing the 1970s. They moved into property and share central bank stepped in to change the management investment, often in companies connected with board and the boards of directors, and to arrange for capital members or staff. Expansion was spearheaded by the injections, including direct infusions from the central largest among them, the Cooperative Central Bank, bank in three cases. and largely unrestrained due to the absence of reserve requirements and lack of enforcement of the 25 per- Licensedfinance companies cent liquid deposit requirement. Audit reports were sometimes as much as two years late. The thirty-two Of the forty-seven licensed finance companies, tradi- deposit-taking cooperatives that ultimately were inves- tional providers of consumer and housing finance, tigated by Bank Negara had 106 related companies those that aggressively lent to finance real estate ranging from newspapers to cosmetics distributors. development were particularly vulnerabie. Finance Almost all the cooperatives were affected by the companies lagged behind commercial banks in build- recession, including the Cooperative Central Bank, ing professional expertise in commercial credit opera- with 363,749 members and total deposits of M$1.5 tions and were slower to restrict loans to the property billion. Its 1987 audited accounts showed accumulat- sector. Problems also occurred disproportionately ed losses exceeding M$726 million and a capital defi- among new entrants. Consequently, the industry was ciency of M$652 million. The bank was placed in hit rather severely by the downward spiral in property receivership, and the government made available a prices and economic performance. By 1984 eleven standby facility of M$323 million to meet liquidity finance companies had reported losses totaling M$17 needs. By the end of 1987 thirty-two deposit-taking million (of industry profits of M$227 million). Eight cooperatives, accounting for M$3. 1 billion in of these firms had started operating between 1979 deposits (an estimated 77 percent of total cooperative 116 BANK RESTRUCTURING deposits), were under investigation by or under the Economic and monetary measures supervision of Bank Negara. The insurance industry also suffered from the In 1985-86 Bank Negara introduced a series of poli- recession. At the end of fiscal 1986 fourteen of sixty- cy measures designed to address the macroeconomic one insurance companies did not comply with the impact of the recession on the banking system. Of minimum solvency requirement of 20 percent of their key macroeconomic importance was the move to net premium income. Inspection by the Office of the allow the currency to find its own foreign exchange Director-General of Insurance revealed that solvency level, with only some intervention to stabilize spo- problems had been compounded by the underprovi- radic speculation. Other measures included extensive sioning of outstanding claims reserves, bad debts from reform of the bank's export credit refinancing scheme poor agency collection, and declining investment val- to promote exports, a reduction in the liquidity and ues for quoted securities and landed properties. As statutory reserve ratios of the commercial banks to part of the effort to coordinate supervision of the lower their effective cost of funds, and the creation of financial system, the duties and responsibilities of the a M$1 billion investment fund to shift bank lending director-general of insurance were transferred from out of real estate to the productive (tradables) sectors the treasury to the central bank in April 1988. of agriculture, manufacturing, and tourism. The bank also introduced a more flexible interest rate The Response regime by freeing deposit rates and encouraged the establishment of a secondary mortgage market to The bank restructuring exercise was approached from securitize long-term housing loans. two fronts: monetary and regulatory. The objectives, While distressed bank borrowers were forced to however, were sometimes contradictory. Tighter pay high real interest rates, the banking system was monetary policy raised interest rates but also generally unwilling to reduce lending rates even when increased the level of nonperforming loans in the general liquidity improved and deposit rates began to banking system. On the other hand, institutional fall. Between 1985 and 1988 the gross interest margin adjustments at the microeconomic level could not be of the banks actually rose from 4.2 percent to 5.2 per- achieved without macroeconomic changes, both in cent (table 7.4). The banks maintained high lending terms of monetary policy and regulatory guidelines. rates because they either were not collecting interest A main objective of monetary policy was to ease the from their nonperforming loans or were forced to contractionary effects of fiscal retrenchment and the recover lost profits due to higher provisions for bad deterioration in the balance of payments. However, debts. Thus depositors and good borrowers were subsi- Bank Negara had considerable difficulty reducing the dizing the banks' losses from nonperforming loans. In abnormally high real rates of interest, even when the short term this approach benefited banks and the deposit rates fell in 1987. The bank had to make banking system, since the banks could recover to prof- major adjustments to meet both objectives, including itability sooner. But high real lending rates deterred some monetary and prudential deregulation to pro- new investment to aid the economic recovery effort, mote greater efficiency in the banking industry. while low deposit rates encouraged disintermediation. Table 7.4 Commercial bank loan margins, 1985-88 (annual percentages) 1985 1986 1987 1988 Indicator Ratce Share Rate Share Rate Share Rate Share Average lending rate 15.2 100.0 15.0 100.0 12.7 100.0 10.9 100.0 -costsoffunds 11.0 72.4 10.1 67.3 7.6 59.8 5.7 52.3 = gross margin 4.2 27.6 4.9 32.7 5.1 40.2 5.2 47.7 -interest-in-suspense 1.4 9.2 2.1 14.0 2.3 18.1 2.5 22.9 =net margin 2.8 18.4 2.8 18.7 2.8 22.1 2.7 24.8 -overhead 3.5 23.0 3.2 21.4 3.2 25.2 3.2 29.4 -bad debt provisions 1.9 12.5 2.3 15.3 1.8 14.3 1.0 9.2 = net loan margin -2.6 -17.1 -2.7 -18.0 -2.2 -17.4 -1.5 -13.8 + nonloan income 2.1 13.8 2.2 14.7 2.7 21.3 2.8 25.7 = net yield -0.5 -3.3 -0.5 -3.3 0.5 3.9 1.3 11.9 Source: Bank Negara Malaysia, various years. Malaysiai Bank Restructuring, 1985-88 117 To ensure that the benefits of lower deposit rates individuals, including family holding companies, in would pass through to the borrowers, Bank Negara the equity of a financial institution were limited to 10 persuaded the banks to reduce their operating costs percent, while any company or cooperative was limit- and interest rate margins. ed to a 20 percent ownership stake. Penalties were introduced to prevent abuses of authority in bank Institutional remedies lending. Bank credit to single customers was restrict- ed to 30 percent of shareholders' funds, and lending The central bank was only the lender of last resort to to the directors and staff of banks and finance com- the banking system; it was prohibited by law from panies was prohibited. providing emergency liquidity to institutions not The central bank obtained powers to lend against under its supervision. But there was a danger that the the shares of, and purchase equity in, ailing financial liquidity problems of failing unsupervised financial institutions so that it could inject additional equity institutions would spread to the monetary sector, that quickly in the event of insolvency or illiquidity. Bank is, to licensed banks and finance companies. The Negara also introduced guidelines on suspension of threat was exacerbated by the fact that sharp fluctua- interest on nonperforming loans and on provisions tions in commodity, share, and property prices had 'or bad and doubtful debts to ensure that the finan- subjected parts of the financial system to exceptional cial community followed sound, consistent, and pru- stresses that many institutions might not be able to dent lending policies and to standardize the account- absorb. In addition, the uneven quality of assets, ing treatment of income from those sources. Audit management, capital cushion, and supervision within and examination committees were established to the financial system, including the formal sector, reinforce boards of directors' oversight of bank man- meant than an uncoordinated supervisory mecha- agements' handling of day-to-day operations. A cen- nism might not be able to cope with the impact of tral credit bureau was established to monitor and the recession. The system needed a legal framework improve consolidated credit information on bank that delineated supervisory powers and provided the and finance company customers. Statistical reporting administrative capacity to deal with future financial to the central bank was improved and computerized, problems. including regular reporting on the size of nonper- Deposit insurance. Bank Negara evaluated a forming loans, exposures to share and property scheme to cover small deposits that would keep small financing, loan margins, and bank productivity. In depositors at bay whenever news of instability sur- addition, on-site bank examinations were increased rounded a deposit-taking institution. A major disad- and the bank examination staff strengthened. vantage of the idea was that the burden of financing Legal changes. Regulations were promulgated in the deposit insurance fund would fall unevenly on July 1986 to remedy legislative deficiencies that the institutions, since the strong ones would be pay- obstructed pursuit of illegal deposit-taking activities. ing proportionately as much as the weak ones but These regulations also improved the central bank's drawing little or no cash support. In addition, any ability to act quickly in an emergency. Bank Negara deposit insurance scheme runs the risk of encourag- was empowered to investigate the affairs of any ing moral hazard among bank managers who, know- deposit-taker, including the power to enter and ing that their liabilities are protected, might take search any office or place of business, to interrogate undue lending risks. In the end no deposit insurance on oath or affirmation, to detain people pending the scheme was established. completion of a search or interrogation, and to com- Prudential measures. In 1985-86 Bank Negara pel the production and retention of accounts, books, made a number of substantial changes designed to and documents. In addition, the law authorized the strengthen the structure of the banking system and central bank to order property to be frozen and to the regulatory powers to prevent and control damage restrict a person's departure from Malaysia, including arising from the recession. These included key impounding their passports. Bank Negara also gained changes to the banking laws and regulations. A mini- the power to require institutions to cease taking new mum capital adequacy requirement was introduced, deposits, to refund existing deposits, and to assume effectively raising the average capital-asset ratio of control and carry on the business of the deposit-taker Malaysian banks from 7.4 percent at the end of 1984 or appoint someone to do so. Furthermore, the law to 8.1 percent at the end of 1987. The holdings of empowered the central bank to apply to the high 118 BANK RESTRUCTURING court to appoint a receiver to manage the affairs and to acknowledge the enormity of the problems facing property of a deposit-taker or to petition the high the banks. Once Bank Negara took control in 1986, it court to wind up an illegal deposit-taker. To protect appointed new directors and chief executives to the public against abuse of these wide powers, the law undertake a thorough review. Rights issues yielded established an advisory panel comprising representa- M$159 million in new equity from existing share- tives of the private sector and the Treasury, along holders, against total losses in the three banks of with the attorney general and the chairman of the M$1,203 million between 1985 and 1986 (table 7.5). Association of Banks. The central bank took up the shortfall in rights by The government also changed the cooperative injecting M$672 million, and the balance was met laws to restrict cooperatives to taking deposits for through subordinated loans of M$401 million. Shares housing and eduction only-not savings or fixed subscribed by Bank Negara are held in trust for dis- deposits-from their members. The thirty-two posal later under a buyback scheme that allows partic- deposit-taking cooperatives were placed under Bank ipating shareholders to repurchase their unsubscribed Negara's supervision until all soft loans to acquiring shares at par plus holding costs. institutions were repaid. To prevent lawsuits from Deposit-taking cooperatives. Following the freeze creditors or depositors from jeopardizing the rescue on deposit-taking and investigations of the twenty- effort, the high court appointed receivers from four ailing deposit-taking cooperatives in July-August accounting firms to manage the cooperatives' assets. 1986, pressure mounted for a quick rescue plan. The The high court also determined the priority of pay- situation was highly charged because more than ment to unsecured creditors of such obligations as 522,000 depositors (in a country of 16.5 million salaries and wages, legal fees, and essential services. people) and approximately M$1.5 billion in deposits were involved. To assess public opinion on the issue, Rescue and restructuring the government appointed an action committee on cooperatives, chaired by Bank Negara, with represen- In light of prevailing economic circumstances and tatives from the government and the private sector. given the budgetary constraint imposed on govern. Depositors believed that their deposits should be ment by high fiscal deficits, it was decided that Bank guaranteed in full by the government. They objected Negara should shoulder most of the burden of inves- strongly to the conversion of their deposits into equi- tigation, diagnosis, evaluation, and restructuring ty and demanded prompt legal action against the entailed in the broad rescue effort. staff of the cooperatives responsible for mismanage- Banks. The central bank generally followed three ment, fraud, and criminal breach of trust.' steps in dealing with each problem bank. First, it The committee determined that the cooperatives' required the ailing institution to recognize all losses capital deficiency was at least M$680 million (table and interest-in-suspense immediately, rather than 7.6). It recommended that the government invite a stretching the losses out over time. Second, the bank's number of strong banks or finance companies to management was replaced. Bank Negara revamped assume the deposit liabilities and attendant assets of the board of directors, appointed a reputable profes- each deposit-taking institution whose net asset to sional to serve as chief executive, and then allowed the deposit ratio was close to one.2 To assist the rescue, new board to take necessary steps to stem losses and the committee recommended that Bank Negara restore profitability. Third, shareholders were required to inject as much new capital as possible through Table 7.5 Indicators for three ailing commercial rights issues, while the central bank filled any remain- banks, 1984-87 ing gaps to meet the minimum capital adequacy (millions of ringgit) requirements. In the case of especially wc-dk institu- Indhcator 1984 1985 1986 1987a tions, the central bank tightened requirements for reg- Total deposits 4,309 4,568 3,945 3,513 ular reviews, discussions, and follow-up inspections. Total loans 3,898 4,326 4,466 3,367 For three ailing banks, United Asia Bank, Perwina Total pretax losses 39 591 612 2 Habib Bank, and Sabah Bank, in which Bank Negara Change in capitalb 38 -5 162 1,130 subsequently injected capital, inspections for fiscal Capital adequacy ratio 4.3 3.0 2.8 5.2 1985 revealed some common weaknesses. In all three a. Projected b. Paid-up capital and subordinated loans. cases bank management and shareholders had refused Source: Bank Negara Malaysia, Annual Report 1987 Malaysias Bank Restruauring, 1985-88 119 provide soft loans to meet the liquidity needs of the Table 7.6 Losses of twenty-four failed deposit- acquiring banks and finance companies. For coopera- taking cooperatives, 1986 tives with large losses, the committee advised that (millions of ringgit) depositors be offered a combination of cash and equi- Losses as a ty or convertible bonds. share of The government white paper on the deposit-tak- Book E&timated book value Type of assavalue losses (perrenr) ing cooperatives adopted most of the committee's recommendation (Malaysia 19a,b. TheFixed assets 150.5 4.4 2.9 recommendations (Malaysia 1986a,b). The report Housingprojects 185.5 30.6 16.5 revealed that the twenty-four ailing cooperatives had Investment in shares 263.6 93.6 35.5 invested nearly half their total assets in connected Loans 948.0 532.8 56.2 Other assets 214.3 21.7 10.1 lending or in subsidiaries and related companies and Total 1,761.9 683.1 38.8 shares. One-fifth of total assets was tied up in land, Source. Malaysia 1986b. property, and housing; another fifth was in loans to cooperative members. Only 9 percent was in cash from their receivers. Depositors were to receive their and liquid assets at the time of the freeze, mainly equity in the form of ordinary M$ 1 shares of because of panic deposit withdrawals. So that the KUMB, which were to be floated publicly as soon as public did not have to bear the full brunt of the large they became eligible for public listing, so that share- losses arising from bad management, depositors were holders might reap capital gains. to receive a share of the net asset backing of their Stage three involved the largest cooperative, cooperative proportionate to their deposits. To facili- Koperatif Serbaguna Malaysia (KSM), with deposits tate the reconstruction process, all rescue schemes of M$549 million and 166,000 depositors. It had an involving debt-equity conversions for depositors asset backing of about 50 sen per ringgit of deposits. would follow legal due process and be approved by The government accepted a proposal by Magnum the courts. Corporation Berhad (MCB), a large publicly listed The final rescue package was carried out in three company, and its licensed finance company sub- stages. Stage one involved the unfreezing of the sidiary, Magnum Finance Berhad, to take over the deposits of eleven cooperatives with relatively small net assets and deposit liabilities of KSM, which had capital deficiencies. They had total deposits of indirect interests in MCB. Depositors of KSM were M$191 million from 85,000 members, with net asset to be repaid in full, half in cash (to be paid in stages backing per ringgit of deposits of close to 1 ringgit. between 1987 and 1989) and half in irredeemable, In January 1987 the eleven received soft loans from convertible, unsecured loan stock of MCB. The loan the central bank and reached agreement with their stock would be non-interest bearing for the first two appointed banks or finance companies to take over years and would be convertible into ordinary MCB their assets and liabilities. Depositors were to be shares at a predetermined rate during 1989-91. repaid in full-but without interest-over periods of To enable the cooperative rescue scheme to work, up to five years. the central bank provided the rescuing financial insti- Stage two of the rescue plan involved the twelve tutions with M$720 million in soft loans at 1 percent cooperatives with moderate to heavy losses (average a year, plus M$280 million in commercial loans at 4 asset backing of 39 sen for every ringgit of deposits). percent a year for a term of ten years. The rescue also The scheme provided for a 1:1 return to all deposi- incurred M$23.4 million in professional fees for the tors through a combined cash and equity arrange- investigations and receivership. ment. At least 50 percent would be in cash paid over a period up to December 1989, while the balance Recovery would be converted into equity in a licensed financial institution that would absorb all the assets and liabili- The prompt recapitalization of the three banks and ties of the twelve cooperatives. Following passage of announcement of the cooperative rescue package in an amendment to the 1958 central banking law, 1987 and early 1988 reestablished public confidence Bank Negara purchased a small, ailing finance com- in the financial system. A generalized panic was pany, which it renamed Kewangan Usahasama avoided at a time when the economy was in the Makmur Berhad (KUMB). KUMB acquired the net trough of its deep recession. Commodity prices assets and deposit liabilities of the twelve institutions began to rise in the fourth quarter of 1986, leading to 120 BANK RESTRUcrURING an immediate injection of liquidity into the banking facturing, trade, agriculture, and property, evaluate system from growing export income. The economy the viability of eligible enterprises and recommend rebounded strongly in 1987: GDP grew 5.3 percent assistance under the fund. in real terms, terms of trade improved 19.4 percent, the current account of the balance of payments Lessons showed a M$6.4 billion surplus, and gross external reserves rose to M$19.4 billion (7.4 months of One key lesson of the Malaysian experience is that retained imports). prompt, decisive action must be taken to address In 1988 the economy continued to rebound vig- problem areas in ailing financial institutions. The orously. Although the federal budget remained in extent of damage must be determined fairly and deficit despite cutbacks in development spending, the accurately and the problems of capital adequacy and financing deficit was reduced sharply, to 5.6 percent competent management must be addressed. Failure of GNP, allowing the government to repay more than to admit the size of the problems or delays in resolv- M$4 billion of its external debt. Real GDP grew at ing them almost always compound the costs. The 8.8 percent and export proceeds rose 21.3 percent. As problems in Malaysia's ailing financial institutions confidence revived, market capitalization in the Kuala generally were worse than they first appeared. Lumpur stock exchange rose by 45 percent to a peak Management was reluctant to recognize and report M$101.9 million in August 1988, from its post-crash accurately to the regulatory authorities the extent of trough after October 1987. Similarly, property prices balance sheet and operating deficiencies. began to firm, with strong sales in low-cost housing Equity capital also proved to be a vital part of the in urban centers and revival in commercial property rescue package. Nonperforming loans can be sup- values from the low of M$180 a square foot in 1986 ported only by non-interest bearing capital. Lending to M$279 in early 1989. The reflation of asset prices high-interest loans to tide insolvent banks over a liq- and enterprise cash flow improved bank profitability. uidity crisis merely pours good money after bad. Preliminary, unaudited data revealed pretax profits of Without adequate capital and an appropriate gearing M$794 million in 1988, compared with M$125 mil- ratio, even the strongest management would be hard- lion in 1987. Bad debt provisions fell by 28.1 per- pressed to restore profitability. cent. Most of the recovery was in the writeback of Supervisory authorities and bank managminent provisions for nonperforming loans, amounting to must develop efficient and accurate reporting systems M$828 million in 1988, or nearly double the so that the true financial position, including risk amount the previous year. By 1989 the national exposure, is clear. Monitoring compliance with tradi- external debt had fallen to M$42 bil!ion, or 44 per- tionai balance sheet ratios is not enough. It is more cent of GNP, compared with 76 percent in 1986. important to monitor the profit and loss account fre- Clearly, the structural adjustment program was yield- quently-monthly or even daily for certain opera- ing results. tions-in order to pinpoint structural weaknesses. Devoting more resources to regular inspections and Enterprise restructuring accurate monitoring is in the long run cheaper than rescue operations. Light supervision wirhout lender- As the banking environment improved, the focus of of-last resort facilities for the deposit-taking coopera- loan rehabilitation switched from banks to enter- tives resulted in large losses that were only partly prises. In 1988 the central bank established a attributable to the effects of economywide deflation M$500 million enterprise rehabilitation fund aimed in asset prices. at reducing the overhang of stalled projects and Financial supervision must evolve in line with nonperforming loans. The fund, financed by Bank competition, technology, internationalizarion, and Negara and managed by a development bank, pro- the erosion of lines of business demarcation in bank- vides seed capital to recession-hit bumiputera ing. Supervisory authorities must monitor not only (indigenous) enterprises that are found to be funda- their traditional monetary institutions but also inter- mentally viable. The projects are cofinanced with relationships with capital markets, rural deposit insti- the existing lenders, who provide additional work- tutions, development banks, and other financial ing capital for the turnaround enterprises. Specialist intermediaries. The financial safety net may have to turnaround groups, composed of leaders in manu- be strengthened and widened. Credit information Malaysias Bank Restr'ucuring, 1985-88 121 needs to be pooled to provide an effective early adequate foreign exchange reserves to cushion warning system. Indeed, substantial economies of against systemic shocks, financial restructuring scale may be gained by coordinating and pooling would have been much more protracted and supervisory resources. difficult. The political will and financial discipline to Banking laws and regulations also need to address the twin deficits in the balance of payments change as technology evolves because the transmis- and public sector were vital ingredients of the recov- sion of shocks accelerates through the payments ery program. mechanism, while the scope for insider trading and financial fraud and theft increases rapidly. Banking Notes laws must give scope to greater competition in financial products while tightening prudential regu- 1. Twenty-two directors of eight deposit-taking cooperatives lation against abuses of the system. At the same were charged in court. Four were found guilty and received jail sentences. time laws relating to bankruptcy (market exit) need 2. Among the options studied was a plan, referred to as to be reformed to expedite and reduce the costs of the 25:25:50 solution, that involved the payment of up to 25 financial restructuring. In 1989 a new Banking and percent of deposits in cash immediately, a further 25 percent Financial Institutions Bill addressing these issues in rwo-year deposits at a maximum of 6 percent annual inter- was passed by Parliament. It included a provision est, and conversion of the balance into equity. This framework authorizing the central bank to act quickly in future was rejected after it was leaked to the press. The option to place the twenty-four cooperatives into liquidation was reject- ed because forced selling would only increase the level of loss- cial system. And with greater internationalization, es. Exchanging deposits into unit trust holdings (to facilitate regulatory supervision has to be coordinated not depositor liquidity) was rejected because not all depositors only across sectors but even across borders. would agree, and most of the cooperatives' assets would not Although there remain pockets of unresolved qualify as trustee assets under existing law. Converting problemis, especially among the smaller finance deposits into shares of a public company would have trans- ferred losses to the public company, which then would have companies, the Malaysian banking sector on the been unable to obtain a public listing under equity quotation whole has regained health. To be sure, the rescue guidelines. Merging several deposit-taking cooperatives into operations were costly: M$1.3 billion in equity one apex cooperative bank would have transferred the losses and at least M$1.3 billion in assistance loans. Over to a single institution, and merging different managements time, with the sustained recovery in property val- and staff would have compounded the difficulties of rehabili- tation. Full government guarantee of deposits implied a gov- ues and turnaround operations, the equity compo- ernment cash injection of M$1.5 billion and losses of nearly nent and a large part of the loans may well be M$700 million, unacceptable given prevailing fiscal and recovered. financial stringency. In the final analysis, after capital and proper control procedures are put in place, the key to References sound and healthy financial institutions remains good management. Where large sums of money are Bank Negara Malaysia. 1987. "Press Statement on Purchase of involved, a code of ethics, improved internal con- Shares in Problem Banks." Kuala Lumpur. trols and procedures,.and peDaities against breaches . 1988. "Press Statement by the Governor on Details of trols and procedures, and penalties against breaches the Rescue Package of the Remaining Thirteen Deposit- of the law are insufficient to deter insider trading, Taking Cooperatives." Kuala Lumpur. conflicts of interest, and lending to related parties. . Various years. Annual Report. Kuala Lumpur. Fraud and abuse must be detected quickly and halt- Hussein, Tan Sii Dato' Jaffar. 1987. "The Management of ed early so that better procedures and competent, the Banking System-Thoughts on Bank Regulation professional, trustworthy management can be and Supervision." Address in conjunction with the Tenth Anniversary of IBBM. Bank Negara Malaysia, installed. November, Kuala Lumpur. Finally, the success of bank restructuring owes International Monetary Fund (IMF). Various years. much to Malaysia's macroeconomic adjustment. The International Financial Statistics. Washington, D.C. beneficial effects of fiscal retrenchment and deprecia- Malaysia. 1 986a. "Investigation Report on Twenty-Four Deposit- tion of the currency brought broadly based econom- Taking Cooperatives." Attachment to Command Paper 50. w hr --. I1986b. "Report on the Deposit-Taking Cooperatives." IC recovery, which assisted the bank restructuring CmadPpr5.KaaLmir ic ecoery Command Paper 50. Kuala Lumpur. process. Without a high level of national savings, Ministry of Finance. 1987. 25th Annual Report of the Director recovery of underlying enterprise profitability, and General of Insurance. Kuala Lumpur. 122 BANK RESTRUCTURING Sheng, Andrew. 1987a. 'Banking and Monetary Policies-A World Bank. 1988. Malaysia: Matching Risks and Rewards in a Framework for Economic Recovery." Bank Ncgara Mixed Economy. World Bank, Asia Region, Country Malaysia, Kuala Lumpur. Operations Department, Washington, D.C. . 1987b. 'Capital Adequacy and Banking." Bank Negara Yusof, Zainal Aznam, and orhers. 1994. 'Financial Reform in Malaysia, Kuala Lumpur. Malaysia." In Gerard Caprio, Izak Atiyas, and James . 1988. 'Regulation of the Banking System." Lecture pre- Hanson, eds., Financial Reform: Theory and Experience. sented to Bank Negara Malaysia staff, June, Kuala Lumpur. New York: Cambridge University Press. CHAPTER 8 Ghana's Financial Restructuring, 1983-91 Andrew Sheng and Archibald A. Tannor Financial sector restructuring in Ghana was part of a Saharan Africa, with considerable assistance from the comprehensive economic recovery program intro- World Bank, the International Monetary Fund, and duced in 1983 to reverse years of economic decline. various donors. The prograrm cost an estimated $300 Poor economic management and prolonged drought million, or 6 percent of GDP (World Bank 1989). It had caused the real gross domestic product (GDP) to involved not only significant institution-building in fall by an average of 0.5 percent a year between 1970 the financial sector, but also major changes in devel- and 1982-30 percent overall in per capita terms- oping financial markets and indirect instruments of and export earnings to shrink by 50 percent. When monetary control. To combat financial imbalances, economic growth declined 6.9 percent in real terms banks were restructured through the innovative in 1982 and inflation accelerated to an annual rate of Nonperforming Assets Recovery Trust. Revaluation 123 percent the following year, the government real- losses in the central bank also were carved out. ized that profound reforms were needed (table 8.1). A second World Bank financial sector adjustment The economic recovery plan introduced fundamental credit was approved in 1992, the start of the second changes in macroeconomic policy, including liberal- phase of financial sector restructuring. The program's ization of the exchange and trade system, improved success ultimately will depend on continued fiscal and monetary discipline, and rehabilitation of improvements in external trade, economic perfor- the social and economic infrastructure. These mance, and external debt management, as well as sus- changes were necessary to slow inflation, reinvigorate tained development aid and a revival of the private the real sector, and restore the solvency and efficiency sector. Achievements in the structural adjustment of a banking system battered by years of severe finan- program have been impressive: real growth has been cial repression and disintermediation. sustained, fiscal discipline has been maintained, and A financial sector review begun in 1985 identified inflation has been brought down to single digits. several financial sector issues as potentially serious There also have been substantial improvements in the obstacles to sustained economic recovery and price sta- financial system, including enhanced supervisory and bility. Public sector credit accounted for 92 percent of regulatory capacity and the development of a more total bank credit at the end of 1982. Negative deposit market-oriented financial environment. rates lowered the ratio of M2 to GDP (a measure of financial deepening) from 28 percent in 1975 to 12 Structure of the Financial System percent in 1983 (table 8.2). Government intervention in the banking system resulted in an exceptionally high Ghana has both a formal and an informal financial currency to deposit ratio of 77 percent by 1981 (Sowa system. The formal system consists of 3 main com- 1989b). Savings in the banking system were clearly mercial banks, 7 secondary banks, a small cooperative insufficient to finance the growing levels of domestic bank, and 112 rural banks, all under the supervision investment needed to sustain economic growth. of the Bank of Ghana. Some 300 credit unions, regu- The financial sector restructuring begun in 1988 lated by the Department of Cooperatives, round out was one of the most comprehensive programs in Sub- the formal system. With one exception, all primary 123 124 BANK REsrRucTuRING Table 8.1 Macroeconomic indicators, 1970-91 (percent unless otherwise specified) Real Tenms of Interest rates Exchange rater growth Current trad Inflation Government - (cediui Year in GDP account/GDP (1980=100) rateb balanceGDP Lending Deposit U.S. dolar) 1970 9.7 -3.0 28.2 3.0 -2.2 - - 1.0 1971 5.2 -6.0 17.4 9.5 -3.5 - - 1.0 1972 -2.5 5.1 25.7 10.0 -5.7 - - 1.3 1973 2.9 5.1 45.3 17.7 -5.3 - - 1.1 1974 6.8 -5.9 54.3 18.1 -4.2 - - 1.1 1975 -12.4 0.6 37.9 29.8 -7.6 - - 1.1 1976 -3.5 -2.7 75.9 56.0 -11.3 - - 1.1 1977 2.2 -2.5 161.4 116.4 -9.5 - - 1.1 1978 8.5 -1.2 137.8 73.1 -9.0 19.0 11.5 1.7 1979 -2.5 3.0 146.7 54.4 -6.4 19.0 11.5 2.7 1980 0.5 0.7 100.0 50.1 -4.2 19.0 11.5 2.7 1981 -3.5 -9.9 70.3 116.5 -6.5 19.0 11.5 2.7 1982 -6.9 -2.7 53.2 22.3 -5.6 19.0 11.5 2.7 1983 -4.5 -4.3 75.7 122.9 -2.7 19.0 11.5 8.8 1984 8.6 -0.9 86.0 39.7 -1.8 21.1 15.0 35.9 1985 5.1 -2.9 74.3 10.3 -2.2 21.1 15.7 54.3 1986 5.2 -0.7 66.7 24.6 0.1 20.0 17.0 89.2 1987 4.7 -1.9 70.8 39.8 0.5 25.5 17.5 153.7 1988 6.3 -1.2 58.5 31.4 0.4 25.5 16.5 202.3 1989 5.1 -1.4 47.7 25.2 0.7 25.6 16.5 270.0 1990 3.3 -3.7 53.1 37.2 0.2 - - 326.3 1991 5.3 -4.3 43.9 18.0 1.5 - 21.3 367.8 - Not available. a. Ratio of the index of average export prices to the index of average import prices. b. Consurmer price index. c. Average par rate. Seurce. IMF, various years; World Bank data. and secondary banks are wholly or partly state Ghana stock exchange began trading the shares of owned. The three main banks account for 57 percent twenty-five to thirty companies in 1990. Two mer- of the total assets of the banking system. The largest, chant banks were established in the early 1990s. The Ghana Commercial Bank, is state owned and insurance and social security sector is dominated by accounts for just under half of total branches and the Social Security and National Insurance Trust and one-third of total banking system assets. The other about twenty small insurance companies. two, Bardays and Standard Chartered, have 40 per- The formal banking system is urban-based. The cent and 25 percent state ownership, respectively. Of greater Accra area accounts for two-thirds of total the secondary banks, three are government-owned bank deposits, and the Northern and Upper East and development banks that shifted toward commercial West regions account for about 10 percent each. banking activities in 1986. A money market is Adding to this urban bias is the fact that commercial emerging around two discount houses, and the banks prefer short-term lending to the commercial Table 8.2 Financial sector indicators, 1975 and 1980-83 (percent) Indicator 1975 1980 1981 1982 1983 M2/GDP 27.6 20.0 17.2 18.5 11.9 Currency outside banks/GDP 9.2 8.0 8.2 7.8 5.5 Currency outside banks/M2 35.0 39.4 43.6 39.2 41.3 Currency/deposits 53.8 65.0 77.3 64.5 70.4 Deposit interest rate Three-month fixed 7.6 12.1 18.3 8.5 11.0 Savings deposits 7.5 12.0 18.0 8.0 11.5 Inflation rate 29.8 50.2 116.5 22.3 122.8 Not: Monetary figures are average monthly figures. Source: World Bank 1989; IMF, various years. Ghana's Fgnancial Restn uring, 1983-91 125 and trade sectors, depriving the agriculture and man- Information) Decree, which opened private deposit ufacturing sectors of long-term financing. accounts to government scrutiny. Other policies The rural banking system was established in introduced in the 1970s and early 1980s also did not 1976 in an attempt to mobilize rural savings. The help. Demonetization of 50-cedi notes, for example, Bank of Ghana contributed to the initial capital of made over-the-counter bank transactions inconve- about 120 rural banks, as did some members of the nient-and time-consuming when cash was unavwil- rural communities. These institutions account for able. To thwart black market activity, the government less than 3 percent of the total deposits of the bank- froze all deposit accounts in excess of 50,000 cedis in ing system (nearly 7 percent of total savings). Rural 1984 and investigated them for tax evasion, fraud, or banks operate the "Akuafo" check system for paying corruption. These actions discouraged the use of cocoa farmers. Although the rural banks serve farm- bank deposits for long-term savings and encouraged ers, their loan structure also is mostly short term, greater use of currency and informal credit channels. and they have accumulated a serious backlog of non- Because of the negative deposit rates, demand performing assets since 1983. deposits accounted for more than 64 percent of total There also are about 300 credit unions. Because deposits in fiscal 1985. Time deposits, which had they place their deposits in the formal banking sys- accounted for 14 percent of total deposits in the tem, they are an important link between the formal banking system in 1976, had fallen to well below 10 and informal systems. Credit unions are inherently percent over the next decade. Such short-term conservative in their lending policies, and their deposit liabilities curtailed long-term lending. transaction costs are low because most staff are vol- The severe disintermediation was compounded untary. Nevertheless, credit unions also suffered a by credit controls and allocation policies. The cen- serious erosion of loan quality, resulting in a loan tral bank imposed quarterly credit expansion guide- collection rate of only about 25 percent in 1987. lines consistent with its projections of economic The rural banks, credit unions, and group lending growth and inflation. In addition, sector-specific schemes through the Agricultural Development credit restrictions were imposed on each bank Bank and others provide credit to about 40 percent according to its volume of outstanding loans and the of small farmers. priorities of the annual development plan. At least The informal system includes a variety of finan- 20 percent of lending had to go to agriculture, and cial arrangements: rural moneylenders, local money that sector also enjoyed preferential lending rates dealerships (susu), parallel foreign exchange markets, that encouraged demand for (but discouraged banks rotating savings and credit associations, privately from supplying) agricultural loans. The lending organized groups that share a common bond, and quota for agriculture served as a cZi!ing on lending trade credit among producers, wholesalers, and retail- to other sectors. Banks kept surplus funds in liquid ers. Informal credit markets expanded rapidly during assets that far exceeded statutory reserve require- the period of financial repression, when private sector ments. Despite government policies, however, com- borrowers were crowded out of the formal system mercial banks generally slowed their lending to agri- and there were substantial differences between the culture, reduced their lending to mining, and official exchange rate and the black market rate. increased their financing of manufacturing and trade (table 8.3). After 1986 most lending rates were Financial repression and disintermediation unified and in late 1987 they were completely liber- alized, but access to credit did not improve substan- Highly distorted interest rates over the past two tially. Lending rates moved very little and remained decades created severe financial repression and disin- negative in real terms. termediation in the banking system. Except for 1985, By the end of 1987 the banking system was suf- when inflation was brought down to 10 percent, real fering from several major sources of losses. First, high deposit and lending rates were severely negative. The liquidity and reserve requirements ensured that all banking system was unable to mobilize deposits to banks held substantial amounts of low-interest bear- any great extent because of a lack of public trust and ing assets. At the end of 1987 cash, demand deposits, the poor quality of customer services. and special deposits with the Bank of Ghana totaled Suspicion of the financial system stems from the 32 percent of bank assets. Lending to the public sec- 1979 Banking and Financial Institutions (Request for tor amounted to an additional 24 percent of total 126 BANK RESTRUCTURING Table 8.3 Loans and advances by sector, 1980-90 (millions of current cedis) Year Agriculture Mining Manufacturing Constructon Other Total 1980 412 61 483 282 1,055 2,2Zi3 1981 670 151 629 400 1,462 3,3 1 1982 1,102 243 635 511 1,211 3,7(2 1983 2,013 423 1,369 795 1,714 6,312 1984 3,779 498 3,288 1,274 3,157 11,996 1985 5,208 960 6,046 2,018 7,169 21,401 1986 7,476 1,437 10,880 3,899 14,963 38,656 1987 10,651 2,631 15,472 6,267 18,465 53,486 1988 10,351 1,254 20,964 7,230 25,950 65,748 1989 11,477 1,928 26,847 8,868 30,186 79,306 1990 12,645 1,074 27,099 11,012 27,570 79,401 Source.:ArYetey, Duggleby, and Hettige 1992. gross assets. Second, banks had a net foreign Financial Distress exchange liability of 7.9 billion cedis ($39.7 million), exposing them to foreign exchange losses as the cedi Two features of the Ghanaian financial sector restruc- depreciated. Third, loan losses continued to mount turing are common to centrally planned, highly in the banks' credit portfolios as formerly protected indebted economies. First, because the financial enterprises adjusted to the liberalized trade regime. system was mostly state owned and invested predom- Finally, interest deregulation increased the banks' cost inantly in the public sector, no sudden liquidity crisis of funds without allowing them to pass on the full or bank runs occurred. The main challenges were to costs to their earnings base (because of the high level restore public confidence in the banking system and required of low-earning and liquid assets). to revive intermediation (deposit mobilization and credit allocation) in order to support general eco- Regulation and supervision nomic growth. The need for bank restructuring was not forced on the authorities by events, but was dri- S8hortcomings in banking legislation and supervisory ven by the recognition that continuing losses in the capacity also contributed to the buildup of bank banking system and inappropriate credit allocation losses. Since supervision previously had concentrated were a drag on the economy. The time had come to on compliance with credit allocations and ceilings- deal decisively with the inherited damage to the not on the quality of bank assets-the Bank of enterprise and banking sectors. Ghana had insufficient capacity in bank supervision Second, the burden of financing the fiscal deficit and on-site examination. The Banking Act of 1970 and the quasi-fiscal deficits arising from the deprecia- was grossly outdated, permitting high concentrations tion of the cedi and revaluation of the external debt of portfolio risk, inadequate levels of capital and fell largely on the banking system. There was a danger reserves, and overstatement of profits. There were no that continued credit to loss-making public enterpris- limits on unsecured lending. The lack of uniform es and monetization of the quasi-fiscal deficits would accounting standards meant that interest accruals on fuel inflationary pressures that, in turn, would retard nonperforming loans continued to be treated as efforts at structural adjustment and stabilization. income and that there was no effective method of In 1987 international auditors were commis- assessing or classifying nonperforming loans. sioned to assess nine of the eleven main banks. The Consequently, transfers to reserves and provisioning audits revealed weak management, capital inadequa- for potential losses were insufficient. Since no mini- cy and technical insolvency, high operational costs mum capital adequacy ratios were in place, lending and overstated profits, weak accounting and manage- risk and asset quality went unsupervised. ment information systems, and low-quality loan Furthermore, development banks that engaged in portfolios with inadequate loan loss provisions commercial banking activities were not subject to (Tannor 1990). The nonperforming loans of the provisioning standards. Finally, because the penalties banking system reflected accumulated enterprise for infractions of the banking law were not adjusted losses following years of high inflation, devaluation, for inflation, they did little to deter abuses. and macroeconomic instability. Ghanas Financial Restructuring, 1983-9! 127 The incentives system was severely distorted. The estimated thot distressed enterprises might account industrialization drive that began in the mid-1960s for as much as one-quarter of domestic output. The was supported by protectionist policies that encour- need for financial scctor restructuring was urgent, aged import substitution, raising value added in since the absence of efficient financial intermediation manufacturing to about 17 percent of GDP. was impeding overall economic recovery. Manufacturing capacity was relatively well developed and diverse compared with other West African coun- Damage Control tries. Still, domestic manufacturing, which accounted for 17 percent of employment and 20 percent of total Under the umbrelia of the first World Bank financial bank loans in 1980, was highly inefficient because of sector adjustment credit, Ghana introduced a compre- the prolonged trade protection. hensive financial sector adjustment program in 1988. Also, economic prosperity in Ghana is closely The program focused on restructuring financially dis- linked to the performance of cocoa and gold. The tressed banks, improving the regulatory framework, world's third-largest exporter of cocoa, Ghana derived strengthening bank supervision, improving resource roughly two-thirds of its export earnings from that mobilization and credit allocation, and expanding commodity alone as late as 1986. Between 1980 and mechanisms for rural finance. In 1992 a second finan- 1 982 cocoa prices fell by more than 50 percent. cial sector adjustment credit aimed at correcting the During the economic crisis of 1980-83 export pro- structural imbalances that encouraged disintermedia- ceeds fell 60 percent. The government responded to tion, strengthening the Bank of Ghana (reducing its the foreign exchange shortage by imposing import revaluation losses, revamping its organizational struc- restrictions on important factor inputs, which wors- ture), and enhancing the effectiveness of nonbank ened the productive capacity of domestic industries. financial institutions, such as insurance companies The depreciation of the cedi not only caused foreign and capital markets, in mobilizing savings. In addi- exchange losses for firms that borrowed foreign tion, the credit provided assistance for restructuring exchange to finance their investments, but also distressed but viable enterprises, continued the increased their production and debt-servicing costs. restructuring plans for banks, and provided training These firms faced significant competition from programs for bankers, accountants, and auditors to cheaper imports when trade was liberalized. Output develop an indigenous cadre of financial professionals. recovery also was inhibited by the collapse of the eco- nomic and social infrastructure. Inefficient but Imp roving resource mobilization and allocative efficiency potentially viable enterprises-apparel, for exam- ple-were adversely affected by sudden trade liberal- Several measures were designed to liberalize the finan- ization, while dairy producers were severely damaged cial sector and to deepen markets by creating an envi- by imports of subsidized products from the European ronment conducive to sound and efficient banking: Community (Weissman 1990). By 1985 the average Foreign exchange bureaus were introduced in capacity utilization of Ghanaian industry was only 25 April 1988 to continue the reform of the foreign percent, recovering slowly to 40 percent by 1989. exchange regime begun in 1986. This scheme A 1988 survey of distressed enterprises that had legalized the thriving parallel market for foreign stopped servicing their bank loans found the most exchange by licensing individuals, banks, or firms common problems to be: to buy and sell foreign exchange. By early 1990 * Foreign exchange losses. more than 180 such bureaus were in operation. * Limited access to new sources of capital (especial- * The monetary authorities gradually moved away ly working capital). from credit ceilings and credit allocation policies * Poor management. toward the use of indirect instruments of mone- * High gearing. tary control (buying and selling government secl- * Failure of banks to appraise loans properly or rities, charging reserve requirements). Even the monitor enterprise performance. mandatory minimum credit allocations for agri- Distress affected enterprises across the board: the culture were eliminated in late 1990. private sector accounted for a third of total nonper- * To develop the money market, the first discount forming loans, state-owned enterprises for a quarter, house was licensed in 1987 and a second in 1990. and various joint ventures for the rest. The survey Both houses accept short-term deposits from 128 BANK RESTRUCrURJNG banks and hold at least 70 percent of their assets banking system. It provided for the following in treasury bills, bankers acceptances, commercial improvements: bills, and certificates of deposits. The second dis- * New capital adequacy requirements were intro- count house was established to help develop a sec- duced for commercial and development banks. ondary market for commercial paper issued by The capital adequacy ratio, which had been public and private corporations. defined as a minimum of 5 percent of total * Two privately owned investment banks were estab- deposits, was redefined as a minimum of 6 per- lished to develop investment banking, corporate cent of the net risk-adjusted asset base (induding advisory services, and the securities market. off-balance sheet items). * The Ghana stock exchange was reorganized to * Risk exposures were regulated through credit lim- promote the growth of the capital market. its on groups and individuals and through limits Other measures to liberalize and improve finan- on loans and advances to bank directors and cial services included liberalizing interest rates in employees. Moreover, restrictions on direct expo- 1988 and giving banks the freedom to vary their sure to all forms of real estate were established. hours of business. Controls over bank charges also Foreign exchange exposure limits were also for- were eliminated, laying the groundwork for rational- mulated to safeguard the fledgling interbank for- ization of operations and a substantial reduction in eign exchange market. transaction costs. The central bank also issued 500- * New guidelines were imposed for classifying bad and 1,000-cedi notes and monitored their impact on and doubtful loans and for treating accrued interest. banking efficiency, with a view to introducing higher * Annual external audits and long-form audit denominations if necessary. Also, a credit clearing- reports were required of all banks. Year-end audits house was established in 1991 as an affi'ate of the had revealed an increase in provisions for bad and Ghanaian Bankers Association to assist banks in doubtful loans from about 7 billion cedis at the assessing the debt leverage and concentration of end of 1987 to roughly 38 billion cedis at the end prospective customers. of 1988. - The Bank of Ghana was granted powers to Regulation and supervision request information, impose fines for noncompli- ance with the law (including punishment for false A major component of the revamping of the financial and misleading reporting), levy liquid asset sector was the strengthening of the regulatory and requirements, impose lending policies, and supervisory framework. Banks adopted new pruden- demand that banks take remedial actions. tial reporting system and accounting standards, A major drive in the reform program was institu- including auditing standards for external audits based tional strengthening and training for commercial and on internationally accepted standards. In addition, an central bankers. The program included skills training intensive training program was established for bank for banking personnel at all levels to encourage a more supervisors, and a computerized off-site surveillance professional approach to bank managerent. A formal system was installed to enable supervisors to monitor training college for bankers was estabiished, and tech- bank performance and detect problems. On-site nical assistance was provided to the Ghana Institute of inspection capabilities were strengthened with the Bankers to develop standards and provide training assignment of two (external) examination specialists comparable to that offered by bankers institutes to the banking examination department and recruit- abroad. In addition, training and professional devel- ment of new bank examiners. By 1990 banks opment courses for accountants were offered through accounting for about 70 percent of total deposits had the Ghanaian Institute of Chartered Accountants. been examined. In August 1989 a new banking law replaced both Loss allocation the Banking Act of 1970 and the 1979 decree that had opened private deposit accounts to government Although various reform alternatives, such as liquida- scrutiny. The new law formed the basis of a compre- tion, merger or acquisition, and immediate privatiza- hensive regulatory and supervisory reform aimed at tion were considered, the phasing of banking sector improving transparency in banking operations and reforms was dictated by pragmatism rather than by fostering competition and supervision in the formal any particular model.' The immediate objectives Ghanas Financial Restutuwing, 1983-91 129 induded holding the line on further deterioration in 22.5 percent of GDP in 1987 to 8.4 percent in 1990. the financial condition of distressed banks, turning The improvement in the government's finances also struggling but viable banks around, and designing facilitated the absorption of the revaluation losses in and implementing a feasible restructuring framework the books of the central bank, which amounted to as for banking institutions that would contribute to much as another 15 percent of GDP their future viability. Who bore the brunt of the losses? One way to Managing the Restructuring Process look at that question is by examining the changes in private sector liquid wealth, defined as money and The bank restructuring program was guided by sever- quasi money less bank loans to the private sector al principles: (table 8.4). There was a sharp loss in the value of * A bank's operations must be premised on sound money and quasi money, equivalent to 22.6 percent banking principles. of GDP in 1983, as a result of the depreciation in the * Every bank must have profit maximization as its official exchange rate. The parallel market had objective, since meeting its social objectives and already depreciated significantly even before 1983. responsibilities as a corporate citizen depends on Thus banking system depositholders absorbed large its solvency and profitability. losses in real terms, while the major beneficiary was * Any public sector credit should be transparent the public sector. Some 90 percent of the credit from and funded through the state budget. the banking system went to finance fiscal deficits and * The board and management of each bank should nonfinancial public enterprises. be accountable and independent of outside inter- Between 1984 and 1990 interest rate controls ference, irrespective of bank ownership. that resulted in negative real deposit rates depressed Accordingly, the board and management must savings in the banking system. Little financial deep- have the requisite experience, skills, and knowl- ening occurred between 1983 and 1990 despite edge to fulfill these obligations. healthy real growth in the economy and improve- * There should be effective regulation and supervi- ment in the balance of payments. M3 as a share of sion to foster safe and sound banking. GDP rose only from 11.3 percent of GDP in 1983 * Public confidence must be sustained through a to 13.4 percent in 1990. Thus although depositors responsive legal environment that guarantees trans- did not lose in nominal terms-the carve-out was parency and due process and eliminates difficulties borne by the budget-they had already taken a associated with the foreclosure of collateral. Banks major loss in real terms in 1983. The cost to the should operate under the discipline of the market. budget for the carve-out, estimated at about 4 per- The implementation team for the restructuring cent of GDP in 1989, was comparatively small rela- program closely coordinated policy decisions among tive to depositors' real losses. It was perhaps the the Ministry of Finance, the central bank, and the shock of the large wealth loss that slowed the recov- banks concerned. The team had four components: ery in financial deepening in the banking system. * An oversight committee to decide on policy issues The restoration of the surplus in the budget gen- on behalf of the government. erated sufficient resources to address the problems in * A technical committee, which reported to the the banking system. The banking system's net credit oversight committee, to monitor and execute the to the central government declined from a peak of restructuring program. Table 8.4 Change in private wealth, 1983-90 (percentage of GDP, in U.S. dollars) Type of change 1983 1984 1985 1986 1987 1988 1989 1990 Money and quasi money -22.6 -0.7 2.2 -0.2 -3.4 1.4 2.2 -0.1 Private secor credit -2.3 0.3 0.9 0.5 -1.5 0.2 2.5 0.0 Net change -20.3 -1.1 1.3 -0.7 -1.9 1.2 -0.2 -0.1 Net foreign assets of the banking system -0.9 -2.2 -3.1 -1.2 -2.5 1.5 1.3 2.5 Memorandum items Claims on central government 15.9 14.6 14.6 16.6 22.5 15.7 9.9 8.4 Domestic credit 17.6 18.0 22.7 23.3 27.5 18.6 17.1 12.4 Sosure: IMF, various years. 130 BANK RESTRUCTURING * Consultants to develop the general framework of assets after loans were exchanged for bonds, as well the restructuring program and bank-specific plans. as responsibility for debt recovery. * Turnaround managers for each bank, supported Based on the audit results and consultant studies, by foreign experts or through twinning arrange- all nonperforming loans to the government and to ments with international banks, to manage the state-owned enterprises-amounting to 31.4 billion institution-building and restructuring process. cedis at the end of 1989-were transferred to the Decisions on whether to liquidate, merge, or Nonperforming Assets Recovery Trust. Under the recapitalize and restructure problem banks (three Nonperforming Assets Recovery Law of 1989, all development banks and six commercial banks) were rights and obligations of nonperforming bank assets made on the basis of external audits. A three-pronged could be transferred to the trust, which was autho- approach involving managerial restructuring, organi- rized to sue for recovery and administer proceeds zational change, and one-time financial restructuring from the debts. The trust was administered by a was devised for each of the banks. An initial estimate board consisting of representatives of the Ministry of of 63.2 billion cedis ($222 million) was projected for Finance and the central bank, the chief administrator, recapitalizing the nine banks-providing provision- an accountant and a lawyer from the private sector, ing, eliminating foreign exchange losses, and meeting and three other experts. new capital adequacy requirements. In the case of nonperforming public sector loans, Pending completion of the individual restructur- which were replaced by interest-bearing bonds in ing programs, urgent intermediate steps had to be 1990, the government retained responsibility for taken to arrest the erosion of capital. Damage control redeeming the loans at face value (principal plus measures introduced by the central bank in early interest) over time. Private sector nonperforming 1989 covered prudential guidelines on lending opera- loans were replaced by long-term Bank of Ghana tions, capital expenditures, operating costs, and other bonds n amounts equivalent to the face value of activities such as loan recoveries, treasury operations, accumulated provisions or were offset against Bank of and internal controls. Critically distressed institutions Ghana loans to the banks. faced more stringent measures. Another major step in the financial restructuring In January 1990 top management and boards of program was the government's assumption in 1990 of directors were reconstituted for banks that were being the revaluation losses of the Bank of Ghana, amount- restructured. All restructured banks had independent ing to 274 billion cedis at the end of 1989 and equiv- management teams with autonomous decisionmak- alent to 19 percent of GDP. These losses were subse- ing powers, supplemented by full-time foreign quently replaced by long-term government bonds at experts. Twinning arrangements with international an adjustable yield. With this important measure the financial institutions were negotiated for some banks, government recognized its quasi-fiscal deficits and whereas turnaround management teams were freed the central bank to control liquidity and money installed for others. Banks took steps to reduce staff, supply without undue concern about the overhang of close branches, reduce operating costs, and improve revaluation losses. efficiency. The results of the carve-out can be assessed fully Since the banks were predominantly state owned, only over time. Detailed profit and loss numbers are it was decided that the state would absorb the losses. unavailable, but the impact on the balance sheet of A carve-out, based broadly on the Spanish model the banks can be assessed. After adjustment and provi- (see chapter 5), was undertaken. Banks were catego- sioning for all nonperforming loans according to rized according to degree of distress or losses. international accounting standards, commercial banks Nonperforming loans and advances to the govern- (including the development banks) recorded a sharp ment and state-owned enterprises, including govern- decline in capital and reserves-from 13.4 billion ment guarantees, were first offset against government cedis prior to adjustment to -2.5 billion cedis at the loans, and then remaining balances were converted end of 1989. By the end of 1990 capital and reserves into bonds. Nonperforming private sector loans were were restored to 48 billion cedis, or 25 percent of total converted into bonds and transferred to the assets, reflecting mostly the effect of capitalization and Nonperforming Asset Recovery Trust, based on each partly the benefits of relatively high spreads. bank's degree of distress and other considerations. With a positive capital base, higher credit stan- The state assumed ownership of the nonperforming dards, and better management, the banks have Ghanas Financial Restructuring, 1983-91 131 enjoyed fairly good spreads of 7.0 to 7.5 percent of the creation, envisaged in late 1991, of a new, mostly total assets since 1990. Although insufficient to cover privately owned entity for providing venture capital the stock of inherited problem loans, the higher prof- and other financial, managerial, and technical ser- itability is clearly an incentive for management to vices to viable public and private enterprises. improve overall efficiency and lays the groundwork Public enterprises are being privatized, with for possible privatization. twenty-five scheduled in 1991 alone. Those remain- ing in public ownership will be given greater autono- Ruralfinance my in pricing, staffing, and procurement. Moreover, managers of the state-owned enterprises will be sub- The rural sector also warranted special attention. ject to performance-based incentive programs. Under a separate project with the World Bank, the government designed a program to improve interme- Conclusion diation in rural areas and to increase the flow of cred- it to small farmers. The project included a substantial Ghana has achieved a remarkable recovery in eco- line of credit to be administered by the banking sys- nomic and financial performance since 1983 Real tem; institution-building, including restructuring growth has averaged about 5 percent a year. two-thirds of rural banks; and strengthening the Inflation was brought down to 18 percent in 1991, financial management capabilities of credit unions. compared with more than 100 percent a decade ear- Other efforts included improving the capacities of lier. And there was a turnaround in the overall bal- the Association of Rural Banks and the Credit Union ance of payments despite weakening prices in cocoa Association to serve their members, strengthening and gold. External payment arrears were eliminated, rural credit appraisal, building stronger capacity with- and domestic trade and finance were significantly in the Bank of Ghana for examination of rural banks, liberalized. and establishing a unit for managing rural finance The country's achievements would have been programs, rural finance policy research, and program impossible without significant stabilization policies in monitoring. three areas: an exchange rate adjustment to achieve In the roughly 10 percent of cases where liquida- external balance, fiscal reforms that restored financial tion was deemed necessary, restructuring of rural discipline and monetary stability, and institutional banks included a payout of the net claims of deposi- restructuring and building to develop a more market- tors (a one-time deposit guarantee by the Bank of oriented economy. Ghana as liquidator); shareholders received nothing. Substantial foreign aid, both financial resources In most cases, however, rural banks with no prospects and technical assistance, played a part. External for recovery were merged with stronger units, and grants amounted to 0.5 to 2.2 percent of GDP a year limits on shareholdings were changed from absolute between 1985 and 1990. Annual net long-term offi- amounts to percentages of total equity (5 percent for cial capital inflows averaged $175 million over the individuals, 10 percent for companies). The emphasis same period. By contrast, net private capital flows was on improving loan collection and provisioning averaged only $14 million, although this figure rose for nonperforming debts, which were estimated at to about $50 million in 1990. Clearly, the long-run roughly 25 to 30 percent of the total. The govern- sustainability of growth and price stability will hinge ment also established a recapitalization fund, to be largely on the maintenance of fiscal and external bal- replenished as restructuring progressed, with banks ance, as well as on the reemergence of the private sec- contributing at least 10 percent of the requirements. tor as the engine of growth. The Ghanaian financial sector restructuring is Enterprise restructuring almost a textbook example of how strong macroeco- nomic stabilization, coupled with financial sector Financial restructuring gave the banks the institution- reforms, can restore growth and price stability. al and financial capacity to assist in restructuring Determination and political will were required to their client enterprises. The Nonperforming Asset enact major reforms. The restoration of fiscal disci- Recovery Trust, in addition, will attempt to restruc- pline was a key factor in assisting the financial sector ture the assets of the banking system that it is respon- restructuring. Major tax reforms, particularly the sible for recovering. The trust's role will be aided by shifting of the tax base from the heavy tax on cocoa 132 BANK RESTRUcrURING exports to taxes on domestic goods and services, chantment with the massively interventionist restored the revenue-GDP ratio to about 15 percent, policies of the past. Building the private sector's encouraged export production, and improved aggre- role in the economy will require changes in phi- gate demand management. In addition, controls on losophy and macroeconomic policies that guaran- expenditure, civil service employment, and subsidies, tee consensus and entrench market discipline. as well as a shift to infrastructure and social welfare * If holding-action measures are not implemented investment, brought a remarkable turnaround in the in a timely and efficient manner, the crisis may fiscal position, from a deficit of 2.7 percent of GDP get out of hand (Tannor 1990). in 1983 to a small surplus in 1986. It was the restora- tion of the primary surplus that allowed the govern- Note ment to program repayments to the banking system, which released resources to correct the damages of 1. This section is drawn substantially from Tannor 1990. the past and placed the banking system in a healthier, more sound position. References Ghanaian authorities were able to halt the disin- termediation process in the banking system through a Aryeetey, Ernest, Tamara Duggleby, and Mala Hettige. 1992. classic carve-out borne largely by the state. The carve- "The Financial Sector in Ghana." World Bank, Central and out became possible because of the improvement in West Africa Department, Washington, D.C. the fiscal account. Depositors did not lose in nominal IMF (International Monetary Fund). Various years. International Financial Statistics. Washington, D.C. terms, but they already had borne significant real loss- Sowa, Nii Kwaku. 1989a. "Financial Intermediation and es in earlier years because of financial repression and Economic Development." In Emmanuel Hansen and the inflation tax. Although the difficult tasks of insti- Kwame Ninsin, eds., The State and Development and Politics tutional strengthening and improved supervision have in Ghana. London: CODESRIA. begun, financial deepening in the banking system will - . 1989b. "Monetary Control in Ghana: 1957-1988." require private sector confidence in the banking sys- University ofGhana, Legon. Tannor, Archibald A. 1990. "Financial Sector Restructuring: tem. The potential for mobilizing additional domestic The Ghanaian Experience." Paper presented at the World savings through the formal banking system is not Bank, Federal Reserve Board, and Bank of Ghana seminar insignificant, given the thriving informal sector. for bank supervisors, February 5-18, Accra. A Ghanaian observer of the financial restructur- Weissman, Stephen R. 1990. "Structural Adjustment in Africa: ing has summed up the lessons of the Ghanaian Insights from the Experience of Ghana and Senegal." World Delopment 18(12): 1621-34. experience: World Bank. 1989. World Development Report 1989: Financial The desire for a higher profile for the private sec- Systems and Development. New York: Oxford University tor in economic development signifies a disen- Press. CHAPTER 9 Yugoslavia: Financial Restructuring in a Transition Economy, 1983-90 Andrew Sheng Yugoslavia presents an instructive case of the prob- budget surpluses, when in reality the quasi-fiscal lems of transition from a socialist to a market econ- deficits of the state were extremely large. For exam- omy.' A pioneer of transformation, Yugoslavia ple, the losses of self-managed enterprises (ultimate- introduced worker self-management of enterprises ly the burden of the state) averaged 6.6 percent of in 1971, which broke enterprises up into what were gross social product (GSP) during 1985-89, while called basic organizations of associated labor the quasi-fiscal losses of the National Bank of (Knight 1983). The lack of financial discipline on Yugoslavia were as high as 8.7 percent of GSP as the part of these self-managed enterprises and their early as 1985 (Lahiri 1991).2 In the Yugoslav case ownership of banks-conditions that mirrored the financial indiscipline, through losses in the enter- problems of connected lending in market-based prise sector, was ultimately reflected in the books of economies-had implications that took more than a the banking system (including the national bank) decade to unfold. In addition, constitutional and monetized and distributed throughout the changes in 1974 decentralized powers to individual economy by inflation. republics, provinces, and local governments. But the The restoration of financial stability required not financial and monetary consequences of local bod- only a credible macroeconomic stabilization program, ies' economic mismanagement fell on the central but also fundamental reforms in both the enterprise government, mostly the central bank (Bole and and financial sectors. Central to these reforms was the Gaspari 1991). private ownership and management of economic and These structural weaknesses were exacerbated by financial institutions. These efforts at reform-and an inward-looking investment strategy with high the difficulties of loss allocation in a federal structure domestic protection, funded by high external bor- of autonomous republics-crumbled in the face of rowing in the second half of the 1970s. To these political and regional fragmentation. strains were added domestic price distortions and inadequate adjustment to the oil shocks. These fac- Banking, Money, and Credit tors helped create the conditions of economic and financial fragility that erupted into hyperinflation The flawed structure of the Yugoslav financial sys- in 1989. tem lay at the heart of the economic crisis of 1989. The Yugoslav case is particularly informative The banking system was little more than a conduit regarding the pivotal role of banks in maintaining for funding enterprises and carrying out national financial discipline and facilitating the exit of loss- credit plans. Under decentralized social ownership making enterprises. In socialist economies property of the banking system, banks were seen as "service rights and obligations were not clearly defined organizations" for enterprises rather than as among the state, individuals, and economic entities. autonomous, profit-making entities. Thus banks Such imprecision created the illusion of fiscal disci- were owned by the enterprises that used their ser- pline in the form of apparent central government vices (Knight 1983). 133 134 BANK RESTRUcrURING Banking structure ings and contributions by the founders for capital and reserves. Yugoslavia had a loose federal banking system con- The basic banks were overseen by nine banking sisting of the National Bank of Yugoslavia and the groups, each of which included an associated bank national banks of the eight republics and to handle foreign exchange operations and large bor- autonomous regions. The system allocated credit rowing. Only one of the nine associated banks oper- from the center to regions, mainly through the ated nationwide, and there was considerable varia- national banks, in a manner intended to compensate tion among them in terms of their financial strength for regional disparities in economic endowments and and their relationships with their basic bank. productivity. The national banks functioned like Associated banks could not issue checking or savings regional branches of the central bank, but with accounts, but their managerial strength, technologi- greater autonomy. Their refinancing of commercial cal sophistication, large scale of operations, and spe- bank credits to priority sectors was funded by the cialized functions in foreign credit and exchange central bank. The central bank also offered discount- transactions allowed them to become the most pow- ing and emergency liquidity facilities, extended cred- erful institutions after the bank reforms in 1989. its directly to the nonbank sector, and provided inter- Some 200 or so internal banks were organized with- est-free credits to commercial banks related to the in enterprises to conduct payments, lending, and transfer of foreign currency deposits.3 credit transactions. They could accept deposits only Until the 1989 Law on Banks and Other from member enterprises (the basic organizations of Financial Organizations the three types of banks associated labor) or workers within the group, but were basic, associated, and internal. Basic and associ- unlike basic and associated banks they were not sub- ated banks, together with national banks, accounted ject to monetary regulation. As a result internal for more than 98 percent of the total assets of finan- banks expanded their activities to become major cial institutions (table 9.1). Nonbank financial insti- suppliers of finance to enterprises. tutions were small and comparatively insignificant financial actors.4 About 145 basic banks formed the Money and credit core of the commercial banking system. They were founded, owned, and directed almost exclusively by The banking system was little more than a conduit enterprises-mainly large ones-which also were for enterprise funding and federal and regional gov- their principal borrowers (although the actual bor- ernment policies. Monetary policy, bank credit rowers were the basic organizations of associated management, accounting practices, and regulatory labor that comprised the enterprises). Competition controls all failed to produce financial discipline. among banks was restricted since enterprises were In an effort to promote foreign exchange earn- allowed to keep only one bank account, which they ings and agriculture, monetary policies encouraged held with their affiliated bank. Banks also were foreign exchange borrowing and monetization of grossly undercapitalized, relying on retained earn- losses through selective and interest-free credits to banks from the central bank. Central bank refinanc- Table 9.1 Assets of financial institutions ing of credits to agriculture and exports at subsi- (share of total assets) dized interest rates accounted for 42 percent of its total credit to banks at the end of 1982. Interest- Basic and Other NationalE assciand fnciral free dinar credits to banks accounted for 48 percent. Nvational associrated filnancial Year banks banks institutions Total Banks lent to enterprises at highly subsidized rates 1980 19.4 76.2 4.4 100.0 while paying negative real rates on dinar deposits. 1981 19.9 76.0 4.1 100.0 Bank credit management was severely handi- 1982 21.7 74.4 3.9 100.0 capped by the close relationship with enterprises 1983 24.2 72.5 3.2 100.0 1984 25.5 71.8 2.8 100.0 and weak accounting standards. Until 1986 banks 1985 26.4 71.0 2.5 100.0 borrowed foreign currency to finance dinar lending, 1986 27.0 70.7 2.3 100.0 thereby accumulating exchange risk. Most of the 1987 22.7 76.0 1.3 100.0 1988 30.4 68.2 1.3 100.0 risk of devaluation on foreign currency deposit lia- 1989 29.1 70.2 0.6 100.0 bilities was transferred to the central bank under an Soure: National Bank of Yugoslavia. insurance scheme introduced in 1978; after 1985 Yugoslavia: Financial Restructuring in a Transition Economy, 1983-90 135 even interest rate risk was covered. The central bank Table 9.2 Net foreign exchange liabilities of the also started assuming a significant portion of out- banking system, 1973-89 standing foreign currency debt, especially that owed (share of gross social product) by underdeveloped republics and regions. With the Deposit dinar devaluation in 1983 and subsequent down- Year Bankingsystem Nationalbanks money banks ward adjustments, the net foreign exchange expo- 1973 1.4 -4.2 -2.8 sure of the banking system became a source of seri- 1975 -2.5 -2.5 ous instability. 1980 -16.5 -4.3 -12.2 OUS instablity. 1985 -31.6 -12.3 -19.2 Since banks were not independent of enterpris- 1989 -26.4 12.7 -39.1 es, they placed few limits on credit to nonviable Source:National Baik of Yugoslavi2. projects. Banks could not refuse rehabilitation loans to their founding member enterprises, nor could of total assets, more than 70 percent of which was they liquidate loss-making borrowers. In addition, estimated to have resulted from the foreign some enterprise losses were socialized. Losses were exchange insurance scheme. The rest arose from the borne by other, profitable enterprises through joint foreign exchange borrowings of the central bank reserve funds designed so that enterprises could help and from transfers to the central bank of foreign one another. exchange losses of selected enterprises. Poor accounting standards did not allow for The large foreign exchange exposure of the sound credit evaluation. The Soviet-based system banking system stemmed partly from banks' exter- did not use accrual accounting and understated the nal borrowing on behalf of their enterprise borrow- losses of enterprise borrowers. Until 1987 interest ers, but also from the rising level of foreign currency on nonperforming loans often was capitalized deposits. Yugoslavia initially benefited from the (interest in arrears was added onto principal), so large foreign currency remittances of nationals banks could not easily distinguish between good working abroad, which were deposited in the bank- income and income on nonperforming loans. ing system. But these deposits later became a chan- Outmoded accounting rules allowed enterprises to nel of capital flight as dinar depositholders attempt- recognize foreign exchange losses only when the ed to escape negative real deposit rates and fears of debts were paid; thus accounts grossly understated devaluation by switching to foreign currency foreign exchange losses. Some foreign debt was accounts. Between 1984 and 1986 total deposits as incurred by banks on behalf of their member a share of GSP fell from 78 percent to 64 percent enterprises. Between 1981 and 1984 some one- (table 9.3). The share of dinar deposits continued to quarter to one-third of banks' total long-term loans fall until 1988, when dinar deposits became less were in arrears. than half of the money supply, resulting in the near The supervisory framework also was inadequate. dollarization of the economy. The rising disinterme- The central bank and the national banks supervised diation was reversed only in 1989 as a result of the all basic and associated banks, focusing mainly on extremely high real interest rates paid during the compliance with monetary, credit, and foreign period of hyperinflation. exchange rules. The Social Accounting Service, an autonomous institution with branches nationwide, audited banks, enterprises, and sociolegal entities- 98tih9 but it did so using auditing and accounting stan- 198489 dards that did not accurately measure the solvency (share of gross social product) of banks according to international standards. Foreign currency As a result of these weaknesses the banking sys- Dinar Foreign deposits as a Dar cunrrncy Total sharr of/total tem incurred increasingly large net foreign exchange Year deposits deposits deposits deposits liabilities, reaching 32 percent of GSP in 1985 1984 49.1 28.5 77.6 36.7 (table 9.2). The national banks' stock of foreign 1985 44.4 27.1 71.5 37.9 exchange liabilities accounted for 70 percent of their 1986 40.5 23.3 63.8 36.6 1987 33.6 31.8 65.4 48.6 total liabilities in 1987, or 24 percent of GSP. With 1988 32.7 39.6 72.3 54.8 the continuing depreciation of the dinar, foreign 1989 49.3 68.4 117.6 58.1 exchange losses amounted to more than 60 percent Soure: National Bank of Yugoslavia. 136 BANK RESTRucrURING Economic Instability in the 1980s This structure of production was unsustainable in the long run because it tended to distort resource Yugoslavia's break in the 1950s from the Soviet allocation, but it did produce impressive results ini- model of centralized planning gave Yugoslav authori- tially. Between 1970 and 1979 real GDP growth ties an opportunity to experiment with economic averaged 6 percent a year (table 9.4), and investment and political decentralization and with worker con- exceeded 30 percent of GDP Enterprises expanded in trol and management of production. Planning was the favorable environment of the 1970s, when world devolved to republics, communes, and enterprises, trade was growing and foreign exchange resources and the market was used as a guide for resource allo- were readily available. But external resources were cut cation. By 1981, 61 percent of the work force was in off when the international debt crisis erupted in the socialized sector, which accounted for 85 percent 1982, and the economy had to deal simultaneously of gross domestic product (GDP). In the agriculture with its large external debt and the inefficient enter- sector, however, private ownership prevailed, prise structure. It was at this point that the flaws of accounting for 71 percent of production and 92 per- regionalism and self-managed enterprises began to cent of the jobs. surface. Growth took place in highly segmented Financial institutions, owned by their borrowers, domestic markets that reflected regional investment the regionally based enterprises, concentrated risks decisions that were duplicative, restrictive on interre- geographically and sectorally. Macroeconomic policy gional mobility of capital and labor, and missed out was designed to support enterprise growth without on economies of scale. adequate checks on leverage, efficiency, and viability. Four related factors contributed to the economic Monetary policy, for example, supported enterprises crisis that emerged in the 1980s: an overvalued through highly negative real lending rates. In addi- exchange rate, poor trade performance, low produc- tion, experiments with price reform-through state tivity growth, and overreliance on external and subsi- fixing of prices or interenterprise agreements under a dized domestic financing. Between the fourth quar- social compact-further distorted resource allocation. ters of 1972 and 1979 the dinar appreciated more Table 9.4. Macroeconomic indicators, 1970-91 (percent unless otherwise specified) Terns Nominal Nominal Exchange rate Real growth Current of trade Inflation Government deposit lending (dinarsl Year in GDP account/GDP (1980=100)' rater halance/GN1' interest rate interest rate U.S. dollar) 1970 5.6 - - 9.2 1.4 - - 0.0013 1971 8.1 - 0.0 16.0 0.5 - - 0.0017 1972 4.3 - 15.0 15.8 -0.4 - - 0.0017 1973 4.9 - -13.0 19.6 -0.8 - - 0.0016 1974 8.6 - --16.9 21.9 -1.3 - - 0.0017 1975 3.6 - 0.1 23.5 -1.2 - - 0.0018 1976 3.9 0.5 -2.7 11.2 -2.6 - - 0.0018 1977 8.0 -2.9 -6.6 14.7 -1.1 3.4 - 0.0018 1978 6.9 -2.3 4.0 14.1 -0.5 4.7 11.5 0.0019 1979 7.0 -5.4 0.0 20.7 -0.3 4.9 11.5 0.0019 1980 2.3 -3.8 -2.9 30.9 -1.1 5.9 11.5 0.0029 1981 1.4 -1.6 2.0 39.8 -0.1 7.4 12.0 0.0042 1982 0.5 -0.9 7.8 31.5 0.1 12.0 16.3 0.0062 1983 -1.0 -0.8 2.7 40.2 0.0 12.0 34.0 0.0126 1984 2.0 1.5 -1.7 54.7 0.0 30.8 44.5 0.0212 1985 0.5 2.1 0.0 72.3 -0.1 60.5 71.5 0.0313 1986 3.6 2.1 12.6 89.8 0.0 55.7 82.0 0.0457 1987 -1.0 1.9 -7.2 120.8 0.0 79.3 111.3 0.1244 1988 -2.0 4.2 13.8 194.1 0.1 279.2 455.2 0.5211 1989 -0.6 3.1 6.0 1,294.9 0.3 5,644.8 4,353.8 11.8160 1990 -7.6 -2.1 - 583.1 0.3 5,644.8 4,353.8 10.6574 1991 - -1.2 - 121.0 0.3 - - - - Not available. a. Ratio of the index of average export prices to the index of average import prices. b. Consumer price index. c. New dinars, end of year. Sourcre: IMF, various years; National Bank of Yugoslavia; World Bank data. Yugoslavia: Financial Restructuring in a Transition Economy 1983-90 137 than 15 percent in real terms against the currencies of program adopted by the Federal Assembly in 1983 the country's ten major trading partners, mainly attempted to address four institutional factors that because the authorities pegged the dinar against the had distorted resource allocation: inefficient invest- dollar even though domestic inflation was higher ment, wage and price distortions, the arbitrary for- than international rates (Edwards and Ng 1985). eign exchange allocation mechanism, and lack of Partly as a result of the overvaluation, as well as the financial discipline within banks and enterprises. worsening terms of trade, the deficit in the current Efforts to get these reforms entrenched continued account increased to $3.7 billion in 1979. In the through the rest of the 1980s. domestic sector labor productivity consistently fell The road to hyperinflation has been well docu- behind real personal incomes because self-managed mented elsewhere. Bole and Gaspari (1991) identify enterprises paid high nominal wages and encouraged three critical periods. Between 1982 and 1984 the overemployment and because of the poor quality of authorities tried to suppress demand and achieve investment projects in the 1970s. external balance by pushing exports, using devalua- The authorities accommodated domestic growth tion, export subsidies, and import and price controls. by granting liberal access to external borrowing and The authorities also tried to suppress inflation with making loans at negative real interest rates. Total direct price controls, but differences in pricing poli- external debt rose from about $2.0 billion in 1970 to cies across republics and provinces resulted in price $15.2 billion in 1979. To encourage remittances anomalies, and the criteria for setting prices were too from workers abroad, the authorities permitted for- vague to be applied successfully. Nevertheless, price eign exchange-denominated deposits, which rose controls were alternately relaxed and reimposed until from 16 percent of broad money in 1975 to 22 per- 1988, and the anticipation of their reimposition cent in 1979. These deposits were channeled by the caused firms to raise prices more than they would banks to the central bank (which absorbed the for- otherwise have done. Each time controls were reim- eign exchange risk) in exchange for dinars for onlend- posed the level of inflation jumped, reaching 50 per- ing to enterprises. Funds were easy to access because cent a year in 1984. enterprises also received favorable rediscount facili- During the second period, 1985-86, a relaxation ties-especially for exports, equipment, and agricul- of policies allowed real wages to rise, stimulating ture-from the central bank through their commer- demand and growth. Still, inflation accelerated to 90 cial banks at 6 percent a year, while inflation averaged percent a year by the end of 1986. By the third peri- 17 percent a year during the 1970s. Normal lending od, starting in 1987, indexation and changes in rates also averaged only 8 to 12 percent a year. accounting techniques had embedded inflation in Three shocks hit the Yugoslav economy in the business practices and household expectations. early 1980s: the sharp increase in oil prices, the inter- Inconsistencies in macroeconomic policies-with national recession, and the international debt crisis. substantial changes in interest rates, exchange rates, These developments raised the costs of debt servicing and wage policies-did nothing to slow inflation. By and reduced the availability of foreign funds. External 1989 the combination of the 23 percent devaluation resources dried up, with net external borrowing falling in June 1988, the lifting of wage and price controls, from $3.3 billion in 1980 to $30 million in 1983. the Tanzi effect fiscal gap resulting from the shrink- Since enterprises were committed to large invest- ing value of import taxes (caused by the lag between ments, the decline in exports and high debt servicing taxation and collection in a high-inflation environ- costs created large losses, estimated at 1.4 percent of ment), and an accommodating monetary policy cul- gross material product in 1981. But because enterpris- minated in hyperinflation. es were managed by workers, there was little incentive to liquidate loss-makers and retrench excess workers Enterprise losses to improve efficiency. Such losses were covered through interenterprise and bank borrowings and Enterprise losses, banking system losses, and policy pooled reserves at the commune and republican levels. mistakes all contributed to the sequence of events A series of economic stabilization efforts that finally ended in hyperinflation. The inefficient throughout the 1980s aimed initially at weathering structure of enterprises was the root cause of losses in the storm but subsequently focused on long-term the economy. In a market economy losses cannot structural adjustment. The economic stabilization continue unless they are financed. In a socialized 138 BANK RESTRUcrURING economy, with no clear definitions of ownership and into two pools: one for the central bank and the gov- obligations, all losses ultimately become a burden on ernment to service external debt, and the other for the state. The built-in instability of the socialist sys- enterprises to purchase imports and service enterprise tem lay in its flawed accounting and incentives sys- debt. With foreign exchange at a premium and rela- tem, whereby the state taxed enterprises regardless of tively easy access to domestic financing, firms their profitability, the banks continued to lend engaged in exports-even at a loss-in order to have regardless of solvency or inefficiency, and the losses of access to foreign exchange. the budget, banks, and enterprises continued to build up in the banking system as a monetary overhang. Banking system losses These quasi-fiscal losses were hidden because the cen- tral government maintained what appeared to be a The 50 percent devaluation in 1983 did not correct balanced fiscal account. distortions in internal resource allocation, but it did Incentives within the socialized management sys- result in a current account surplus between 1984 and tem were highly distorted. Worker management and 1988 and a net repayment of $3-$4 billion in out- the regional orientation of enterprises politicized standing external debt. On the other hand, growth decisionmaking. Low productivity reflected not only slowed to 0.7 percent a year between 1981 and 1985, the poor quality of investment decisions, but also the and the inflation rate accelerated to 80 percent a year selection of energy-intensive, inward-oriented, and by the end of 1985. import-intensive industries with long payback peri- The devaluation had an adverse impact on enter- ods. Because foreign exchange risks were borne by the prises and banks. Eighty percent of external debt was central bank, undervaluation of the cost of capital owed by enterprises, and half of that was guaranteed was endemic to the system. Moreover, the high rate by associated banks. By 1985 external liabilities were of enterprise taxation discouraged the use of retained so great that the central bank assumed $4$5 billion earnings to finance investment. of the external liabilities of enterprises and banks, Enterprise accounting did not address problems mostly in the underdeveloped regions. More than 80 of solvency or efficiency. Taxes were imposed on percent of the central bank's revenues were used to enterprises before wages were paid, so enterprises finance interest payments on foreign debt. The bank paid taxes even if they incurred losses (Rocha 1991). also assumed liability for the foreign currency Enterprise losses rose from 3 percent of GSP in 1985 deposits of the commercial banks, which were owed to 15 percent in 1989, and yet wage payments rose mostly to Yugoslav citizens (and were not part of the from 10 percent of income to 14 percent over the external debt). same period (table 9.5). Similarly, taxes on enterprises The central bank's purchase of foreign exchange remained at 7 to 9 percent of GSP despite enterprise to service foreign debt, payment of interest on for- losses. The losses were funded by the banking system, eign currency deposits, and continued lending to which was compelled by the connected ownership banks to refinance loss-making enterprises were structure to provide funds. highly expansionary in terms of money supply. An example of distorted incentives was the Whereas foreign borrowing had been used to finance rationing of foreign exchange by the central bank excess demand during the 1970s with little inflation- during the early 1980s. Export proceeds were divided ary consequence, excess demand during the 1980s Table 9.5 Enterprise losses and expenditures, 1985-90 (percent) Item 1985 1986 1987 1988 1989 1990 Losses as a share of total revenue 0.8 0.9 3.9 5.8 4.6 2.3 Wages as a share of total revenue 10.3 12.4 12.7 11.7 14.0 16.5 Materials as a share of total revenue 77.1 74.4 73.9 72.5 57.6 74.0 Taxesasashareoftotal revenue 1.6 2.4 2.5 2.3 3.6 2.4 Interest payments 5.6 6.2 9.2 14.0 23.8 4.0 Losses as a share of gross social product 2.8 3.0 6.6 5.7 15.0 Taxes as a share of gross social product 7.2 8.7 8.2 7.2 7.4 - Taxes as a share of public sector revenue 22.2 25.9 24.4 22.2 24.0 18.0 - Not available. Sounre:Lahiri 1991. Yugoslavia: Financial Restructuring in a Transition Economy, 1983-90 139 was accommodated by expansionary monetary poli- Table 9.6 Change in balance sheet item "other cies. This approach resulted in a vicious circle of assets" as a proxy for central bank losses, 1982-90 inflation and currency depreciation (figure 9.1). Had (percent) the uncovered capital losses of the central bank Other assets resulting from its large foreign exchange liabilities Currency as been borne by the enterprises and the budget Share of Share of ashareof total central gross social social instead, both would have been forced to cut spend- Year hank assets product Change product ing to maintain solvency-bringing aggregate 1982 35.5 15.4 n.a. 6.7 demand back to equilibrium with aggregate supply 1983 49.1 26.9 11.5 6.2 at a lower rate of inflation. 1984 54.7 31.6 4.7 5.2 Gaspari (1989) was the first to point out the 1985 60.9 33.6 2.0 4.9 1986 61.6 30.6 -3.0 5.3 enormity of the central bank's uncovered capital 1987 68.9 41.0 4.4 4.4 losses resulting from its large net foreign exchange 1988 65.0 46.3 5.3 3.9 liabilities. Capital losses rose from 3.3 percent of 1989 49.9 58.3 12.0 5.6 GSP in 1981 to 11.8 percent in 1986. Mates Not licbl (1991), in calculating the entire public sector deficit, -Not available. includes the quasi-fiscal deficit of the central bank Sore: National Bank of Yugoslavia; author calculaiions. accumulated in various forms: subsidized loans, loss- es assumed by the central bank, and claims on banks 1982 and 1989-about the level of average quasi-fis- which were written off. These quasi-fiscal deficits cal losses calculated by Mates. amounted to 5 percent of gross material product Under these circumstances control over monetary between 1980 and 1989, some 3.0 to 3.5 percent of supply was lost because foreign exchange transactions it financed by seigniorage. enlarged the money supply autonomously (Gaspari Although data on central bank losses have never 1989). Central banking authorities relied on direct been published, the item "other assets" in the central credit controls, rather than open market operations, bank's balance sheet contains both revaluation and to restrain credit growth. But defects in the design of other accumulated losses. Other assets grew from 36 credit ceilings actually accommodated increases in percent of total central bank assets in 1982 to 69 per- domestic spending because they did not take into cent by 1987, before declining somewhat as net for- account the revaluation of foreign currency deposits, eign liabilities fell (table 9.6). As a share of GSP, other whose dinar value increased with every exchange rate assets grew by an average of 6 percent a year between devaluation. Furthermore, commercial banks were Figure 9.1 The effects of expansionary policy SurplusIneaenNeunoed demand ...... tdc rcscptllse Source: Kessides and others 1989. 140 BANK RESTRucrURING able to circumvent the ceilings, which in any case did was liberalized, the dinar was devalued by almost 20 not apply to selected credits or to overdrafts. Even percent (even though Yugoslavia was running a cur- though bank lending halved in real terms during the rent account surplus), and steps were taken to liberal- 1980s, disintermediation provided parallel sources of ize domestic prices and foreign trade regimes. financing for enterprises through interenterprise cred- The program, however, was inconsistent and inef- it. Moreover, household consumption increased fective. Because the program did not address the because foreign currency deposits grew in dinar terms quasi-fiscal deficits in the central bank and enterpris- with every devaluation, creating the illusion of es, fiscal adjustments were not rigorous enough to wealth. These factors, together with continued nega- compensate for such losses. Money and credit growth tive interest rates, undermined stabilization even targets were poorly defined and permitted excessive though monetary policy officially was restrictive. expansion, even though real interest rates became positive and interenterprise lending to loss-making Hyperinfiation dzzvelops enterprises was reduced. Rising interest costs caused enterprises to default on their bank loans. Money Monetary policy was loosened in 1985-86 to avert growth targets were abandoned as the costs of financ- widespread enterprise bankruptcies. During this ing hidden losses rose. period of "programmed inflation," money was Several factors appear to have been critical to the pumped into the economy to push growth and development of hyperinflation in 1989. Because of improve the export sector. Real wages were allowed wage indexation the large real devaluation in 1988 to increase by an average of 8 percent a year in these quickly fed back into inflation with higher wages. two years, after a decline of 9 percent a year in the And enterprises quickly factored into their prices the previous two years. Although inflation continued to higher costs of imported inputs. In bringing interest rise, economic performance improved as internation- rates to positive real levels, banks significantly raised al interest rates and oil prices fell. Moreover, the the costs of production, with the interest component dinar was allowed to appreciate in real terms in of enterprise costs rising from 9 percent of revenue in 1986-87, which helped the inflation-adjusted public 1987 to 14 percent in 1988 and to nearly one-quar- sector deficit and slowed the rise in inflation. The ter in 1989 (see table 9.5). As before, the increase in increase in domestic demand from higher real wages losses was accommodated by bank credit policy, and government spending stimulated real growth of which kept credit flowing to loss-makers. 3.6 percent in 1986, the highest since 1979. Largely In addition, with positive real deposit rates and for political reasons, wage indexation was introduced the benefits of devaluation stimulating capital in 1987. inflow-on top of an increase in the surplus of the The revival of domestic demand in this second current account-the central bank's foreign phase was not sustainable. Domestic inflation contin- exchange reserves increased by $1.6 billion in 1988 ued to accelerate, and the trade and payments posi- and another $1.8 billion in 1989, thus injecting fur- tions worsened. Although in 1987 export perfor- ther liquidity into the system and accommodating mance improved and imports declined substantially, the rising prices. Worse, the reduction in import tax confidence in the dinar was shaken, resulting in a net collections in June 1989 aggravated the budget outflow of short-term capital of nearly $3 billion.5 In deficit and generated an enormous Tanzi effect, the absence of structural reforms in the enterprise which also required budgetary funding. Prices rose sector, and with negative real interest rates continuing by more than 240 percent in 1988, and exploded to until 1988, inflation accelerated to 120 percent in 1,200 percent in 1989. 1987 and real incomes fell again in 1987 and 1988. In May 1988 the government renewed its efforts Crisis in the Banking System to restrain domestic demand and inflation. The new program, an attempt to introduce comprehensive dis- Inflation had a highly distortionary effect on the cipline, aimed to contain inflation with progressive banking system. Disintermediation was evident- reductions in money and wage growth targets and narrow money (Ml) fell from 24 percent of GSP in with indexation of time deposits to reduce the sub- 1980 to 6 percent in 1989, and foreign currency sidy on bank credits to enterprises (Coricelli and deposits came to account for more than half of total Rocha 1991). The foreign exchange allocation system deposits (see table 9.3). Highly negative real deposit Yugoslavia: Financial Restructuring in a Transition Economy 1983-90 141 and lending rates eroded financial deepening. Starting in 1987 banks had to deduct current for- Between 1980 and 1986 money and quasi money fell eign exchange losses from their current income. This from 79 percent of GSP to 58 percent. There was move improved transparency but reduced bank prof- also a corresponding decline in domestic credit, from its. Provisioning for accumulated losses on doubtful 102 percent of GSP in 1980 to 56 percent in 1986, loans started in 1987 but did not reflect fully the as borrowers benefited from the erosion of their real degree of risk, leaving open the question of future debt. Enterprises responded to attempts to restrict drains on current income. Even after 1987, when bank credit by increasing their reliance on credit some enterprise debt was transferred to the central from other enterprises, in the form of supplier or bank, nearly 46 percent of bank loans (27 percent of trade credits and the accumulation of interenterprise assets) were still in foreign currency. In addition, con- arrears. The volume of interenterprise credit tingent liabilities (guarantees) represented more than increased from 26 percent of total credit in 1980 to half of banks' total assets, and a large portion of these 39 percent in 1987 as a result of the dramatic rise were foreign guarantees. (150 percent) in enterprise arrears. Banks attempted to recover their losses through As the currency depreciated and inflation rose, high interest rate spreads, which averaged 34 per- capital flight from dinars and dinar deposits became cent in real terms in 1987. A 1987 World Bank evident as domestic savers switched to foreign curren- study found that commercial banks would have had cy deposits. These deposits rose to more than a quar- to charge a spread of 78 percent over deposit rates ter of banking system liabilities in 1989, compared of 73 percent to recover the costs of (low-interest) with less than 15 percent in 1980, inflating the net public sector loans, low yields on reserves and liq- foreign liabilities of the banking system. Thus the uidity requirements, the costs of provisioning for banks were vulnerable to large losses whenever the nonperforming loans, and foreign exchange losses. dinar was devalued. As spreads reached 70 to 80 percent in 1988 and The impact on banks of enterprise losses, foreign 1989, and nominal lending rates hit more than 150 exchange exposure, and inefficient bank operations percent, the burden of nonperforming loans and could be detected-despite poor accounting stan- losses in the banking system began to fall on good dards-from the steady erosion of the capital base of borrowers. basic and associated banks (table 9.7). A large share These real rates on captive borrowers placed an of foreign exchange liabilities was transferred to the additional strain on debtors, encouraged distress bor- national banks, but some was retained on the books rowing, promoted disintermediation, and deterred of basic and associated banks. The continued depre- new investment. In addition, to stem the losses of the ciation of the dinar affected these banks' financial central bank, after October 1988 commercial banks position. With loan portfolios concentrated on a lim- could no longer transfer foreign exchange deposits to ited number of borrowers-usually affiliated enter- the central bank. Thus commercial banks became prises-some banks began to incur losses several fully responsible for the exchange risk on new times the value of their capital. deposits, though they were still prohibited from lend- ing in foreign currency (although they could contin- Table 9.7 Real interest rates of basic and associated ue to sell the currency to obtain funds for dinar lend- banks ing). Dinar lending rates, therefore, had to reflect the (percent) depreciation risks. When hyperinflation reached 1,200 percent in 1989, real lending rates rose as high Discount Deposit Lending Capital-asset ' Year rate rate rate ratio as 252 percent. 1980 -19.0 -19.1 -14.8 2.8 Such high real interest rates distorted resource 1981 -24.2 -23.2 -19.9 2.5 allocation as much as negative real rates had. 1982 -13.3 -14.8 -11.6 2.3 Insolvent enterprises had no qualms about borrow- 1983 -7.3 -20.1 -4.4 1.8 1984 -5.0 -15.5 -6.6 1.5 ing, since they had no intention of repaying. Worse, 1985 -6.5 -6.8 -0.4 1.4 they used interenterprise credit, so their insolvency 1986 -17.8 18.0 -4.1 1.6 threatened even healthy enterprises. The liquidity 1987 4.6 -18.8 -4.3 1.0 1988 60.5 28.9 88.0 1.2 crunch passed the burden of high real interest rates to 1989 518.5 328.8 252.0 8.1 viable enterprises, which were decapitalized, fueling Source: IMF, various years. further distress borrowing. 142 BANK RESrRUCTURING At the heart of the banking problems lay the fun- exchange reserves increased to about $9 billion. The damental need to restructure inefficient and unprof- financial sector, however, began to suffer from classic itable enterprises. Weak accounting frameworks, problems of deflation in the post-hyperinflation phase linked bank and enterprise ownership, distorted wage of recovery. In early 1990 annual real lending rates and price levels, inappropriate taxation, and bad were still high, at roughly 45 percent, while deposit investments had resulted in a scale of enterprise losses rates were 15 percent. Tight monetary policies actual- that was difficult, if not impossible, to measure accu- ly facilitated the disintermediation of funds to the rately. By the end of 1989 enterprise losses were esti- interenterprise credit market and increased distress mated at $10.4 billion (about 15 percent of GSP), borrowing because of systemic losses. Enterprise losses more than three-quarters of it in industrial sectors climbed, and with them nonperforming bank loans. (especially mining and energy). Tight monetary poli- The number of illiquid enterprises soared from less cy in early 1990 led to 350 enterprise failures by mid- than 1,000 at the end of 1989 to more than 3,900 by May, affecting 210,000 workers. About 7,000 enter- October 1990, involving 1.6 million workers. During prises were unable to pay back bank loans in the first the same period the number of bankruptcy proceed- half of the year, and 1.5 million workers were not ings increased from 200 to just under 1,000. paid in March and April so that enterprises could The real sector was not helped by rapidly rising meet their other financial obligations. domestic prices relative to international prices. With Precise estimates of banking system losses are not the pegged nominal exchange rate these inflation possible when domestic prices are severely distorted. disparities caused the dinar to appreciate by more Estimates for 1988 suggest that 35 to 40 percent of than 122 percent in real terms, pushing the current banking assets were nonperforming (Kos and Cvikl account deficit to $850 million. Enterprises lost 1991). Some $8-$10 billion was affected (social export markets. In the fourth quarter of 1990 the product was $55-60 billion, bank capital $3.0-$3.5 economy was hit by the higher oil prices caused by billion). A preliminary estimate for 1991 lifted the the Gulf war, as well as by lost export contracts and total to $12-15 billion (external audits were not worker remittances. Industrial output shrank 10 completed satisfactorily because of the political percent and inflation rose to nearly 600 percent in upheaval in 1991). 1990 as real wages kept rising and monetary policy was relaxed at the regional level. Stabilization and Restructuring Financial sector restructuring In December 1989 the Markovic administration announced a comprehensive stabilization program to Despite the considerable political and macroeconom- combat hyperinflation. The central government was ic difficulties, significant efforts were made to address limited in its financial control, however, because the the weaknesses of the banking system. A new central federal budget accounted for only 15 to 20 percent of bank law in 1989 (the Law on Banks and Other public expenditure and the center had little control Financial Organizations) strengthened the bank's over spending by the republics and provinces-espe- powers to restructure the banking system and develop cially at the enterprise level. Nevertheless, the burden open market operations. Rediscounts on selective of fiscal restraint fell on the federal government when credits were phased out, and any remaining subsidies the central bank's quasi-fiscal operations were incor- were to be funded from fiscal revenues. The burden porated in the state budget. Stabilization involved of foreign currency losses was transferred from the tight monetary and fiscal policies, a fixed nominal central bank to the federal government. The federal exchange rate pegged to the deutsche mark, a six- government no longer had any automatic access to month wage freeze, and a comprehensive program of central bank credit. Short-term credits from the cen- price and trade liberalization. Also included was a tral bank could be made only at the request of the package of measures related to ownership, long-term Federal Executive Council. economic development, and bank and enterprise The 1989 banking law gave the central bank the restructuring (Gaspari 1990). The currency was recal- authority to introduce regulations on capital require- ibrated at a rate of 1 new dinar to 10,000 old dinars. ments and exposure limits. The 145 basic and 9 asso- These measures initially met with great success. ciated banks were to be restructured and consolidated Inflation fell to zero by June 1990, and foreign into universal banks licensed by the central bank. To Yugoslavia: Financial Rrtrtcturing in a Tramition Economy 1,983-90 143 introduce profit-making as the key operating princi- repayment by banks. Other measures involved refi- ple, the law provided for banks to be transformed nancing $400 million of selective credits and resched- into joint-stock companies, which would make them uling some of the $300 million in credits from the independent business entities. Ownership was Soviet Union, thereby allowing some loans to be opened to individuals and foreigners. The law also reclassified into categories with lower loan loss provi- removed many structural impediments to competi- sions. To improve banks' income position, interest tion, including the limitation of one bank account rates on required reserves were raised to 26 percent per enterprise. All banks meeting minimum financial and 16 billion dinars in bad assets were eligible for criteria would be relicensed by the central bank. By replacement by renewable six-month central bank February 1991, 102 commercial banks had obtained bills at 26 percent a year on "conditional" termns. The licenses. In addition, a 1989 accounting law (later banks were not allowed to cash in on these bills to amended) moved the Yugoslav system closer to inter- generate more bad credits. nationally accepted standards and increased the trans- For banks that were to be closed case-by-case parency and meaniingfulness of information. measures included revoking licenses, voluntary real- The 1989 banking law established the principle ization or liquidation of assets and liabilities, assunip- that banks undergoing restructuring should write off tion of good assets and legitimate liabilities by anoth- losses against capital and suspend bank management er bank, and, when all else failed, bankruptcy pro- immediately. A Bank Rehabilitation Agency was cre- ceedings. For banks that were to be rehabilitated ated to restructure banks by arranging mergers or rather than closed, measures involving the Bank takeovers, purchasing assets, investing or recapitaliz- Rehabilitation Agency included writing off capital ing, and attempting to privatize them. Bank restruc- and assuming assets, purchasing bad loans, building turing was intended to unlink restructured banks up capital, and replacing management. from their former problem debtors and bring funda- Financing for the initial across-the-board asset mental change to bank ownership by replacing enter- and income rehabilitation was estimated at 18.5 bil- prises as owners. The commercial and investment lion dinars ($1.4 billion). About two-thirds of it was functions of banks were to be separated from other to come directly from the (federal) budget, 2.3 bil- activities by assigning them to capitalized subsidiaries lion dinars from the central bank, and the rest from (Gaspari 1991). the rehabilitation agency. After this initial effort, it For a key group of strategic banks, holding more was estimated that an additional $6.2 billion (at cur- than half the assets of the banking system, the pro- rent dinar exchange rates) would be needed for case- gram called for a combination of across-the-board by-case restructuring and another $1.4 billion over help and partial recapitalization. For a middle group five years to meet the commitments of the initial of insolvent but salvageable banks the program across-the-board measures. It was argued that these would reduce insolvency. A third group of solvent outlays would: banks was expected to benefit quickly from the * Arrest the erosion of the financial position of reforms. In addition, case-by-case restructuring banks. would deal with individual banks according to their * Prevent relationships between banks and their degree of distress. Prior to implementation of the major borrowers from multiplying. across-the-board measures, it was estimated that * Lay the foundations for a transformation in the group A banks had a negative net worth of 98.7 bil- ownership of banks, ultimately leading to self- lion dinars and group B banks of 1.6 billion dinars. sufficiency. Group C banks had a positive net worth of 2.3 bil- The restructuring of illiquid banks began in May lion dinars. 1990, but progress was impeded by questions about Across-the-board measures consisted of both asset the accuracy of comprehensive audits and the size of and income rehabilitation operations. Asset rehabili- losses. Measures that would force bankruptcy on tation included writing off $962 million in foreign inefficient enterprises were still absent, and the politi- debt by transferring it to the federal budget and cal commitment to financial discipline weakened repaying $4.2 billion in valuation losses on the prin- under severe political strain (Transition 1991). Since cipal of certain foreign obligations. Under the repay- efficient labor and capital markets were not yet in ment effort, the federal government was to provide a place, bankruptcies increased unemployment and specified portion of the dinar value of the annual depleted the resources of the social safety net, placing 144 BANK RESTRucrURING an insupportable burden on fiscal policy. tional environment when external resources were cut Restructuring was disrupted by the advent of civil offin the 1980s. war and the declaration of independence by Croatia In the early 1980s Yugoslav authorities were able and Slovenia in 1991. to maintain external balance through devaluation and spending cuts at the federal level, but they were Who bore the losses? unable to maintain internal balance. By transferring the foreign exchange burden to the central bank The prolonged crisis in the 1980s meant that losses without correcting the structural flaws, the authori- were spread over almost a decade, but were not ties failed to reduce excess demand and enforce finan- widely eliminated from the system. The distor- cial discipline at the regional, enterprise, and house- tionary effects of hyperinflation and devaluation hold levels. A central bank burdened with net foreign make it difficult to establish who bore the losses in exchange liabilities could not maintain national dinar terms. An examination of the banking system's financial discipline even if it were empowered to do balance sheets between 1980 and 1989, deflated in so. Instead, the banking system facilitated (rather dollar terms, reveals that the dollar value of credit to than checked) credit creation, and households enterprises declined by $40.5 billion during the believed that depreciations increased their wealth in 1980s-indicating that enterprises received enor- foreign currency deposits. mous wealth transfers to pay for their inefficiencies. The inflation tax initially transferred resources The dollar value of money and quasi money fell by from dinar depositholders to creditors, thus alleviat- $25.7 billion over the same period, indicating that ing the pain of fiscal adjustment. But financial disci- households bore the brunt of the wealth transfer, pline required the quasi-fiscal deficits of uncovered particularly in the forni of negative real deposit rates capital losses of the central bank-roughly 5 percent and devaluation. of GSP-to be absorbed in the budget. Although it "Other assets" of the central bankc (a proxy for would have been difficult politically to raise taxes to central bank losses) increased by $9.8 billion over the pay for these and other enterprise losses, delaying same period. Losses actually totaled $15 billion as of these fundamental corrections allowed the losses to the end of 1987, but an improvement in the central be monetized. Thus instability in the banking sys- bank's net foreign liability position in 1988-89 sub- tem reinforced fundamental flaws in the real sector, stantially reduced its losses. Thus enterprise losses creating the vicious circle of stagflation that ulti- over the decade were borne roughly two-thirds by mately degenerated into hyperinflation, helped along households, one-quarter by the central bank, and the by the indexation of wages and built-in expectations balance by bank capital. These losses were enormous of inflation. Since loss allocation had regional impli- by any standard. cations, these economic mistakes only compounded regional and political tensions. Conclusion The Yugoslav experience contains important lessons for other countries engaged in enterprise and All transition economies face the complex task of financial sector reforms. It demonstrates the impor- finding a way to distribute real sector losses that are tance of the following conditions for successful embedded in the banking system without contribut- reforms: ing to large-scale macroeconomic instability (Sheng * Property rights and obligations must be clearly 1990). In the Yugoslav case the difficulties were enor- defined, particularly among individuals, enterpris- mous, made worse by an inherently unstable enter- es, banks, and the state. prise and banking structure, as well as by great * Legal and accounting frameworks must be clear regional economic disparities. and transparent to enforce financial discipline. In The worker self-management experiment in addition, a hard budget constraint must be Yugoslavia created fundamentally unstable and ineffi- imposed at all levels. cient enterprise and bank structures. Easy access to * Structural adjustment must be accompanied by domestic and foreign resources encouraged the sys- fiscal and monetary policies that foster competi- tem to grow in the 1970s. But regionalism and inap- tion, growth, and financial discipline. Attention propriate economic policies created an inefficient should focus on policies that promote structural structure unable to adjust to a competitive interna- improvements in competition and stability, as Yugoslavia. Financial Restructuring in a Transition Economy 1983-90 145 well as those that reduce repression and remove Meridor, eds., Lessons of Economic Stabilization and Its price distortions. Aftermath. Cambridge, Mass.: MIT Press. • The institutional framework should develop both Coricelli, Fabrizio, and Roberto de Rezende Rocha. 1991. "Stabilization Programs in Eastern Europe: A Comparative institutions and human skills, with proper incen- Analysis of the Polish and Yugoslav Programs of 1990." tives for efficiency, stability, and equity. Particular Policy Research Working Paper 732. World Bank, attention must be given to relations between the Washington, D.C. central bank and the banking system, especially Edwards, Sebastian, and Francis Ng. 1985. "Trends in Real the central bank's monetary and supervisory Exchange Rate Behavior in Selected Developing tnesponsibilities. moeay n uerloy Countries." CPD Discussion Paper 16. World Bank, responsibilities. Wsigo,DC Washington, D.C. * The right reforms need to be introduced at the Estrin, Saul, and Lina Takla. 1991. "Reform in Yugoslavia: The right time, induding the sequencing of real sector Retreat from Self-Management." Paper presented at a con- reforms with financial sector changes. For example, ference by the World Bank, London School of Economics, early liberalization of the capital account (through and Portuguese Catholic University, January 31, the introduction of foreign currency deposits) Washington, D.C. Gaspari, Mitja. 1989. 'Balance of Payments Adjustment and became a source of instability when authorities Financial Crisis in Yugoslavia." Economic Development tried to correct domestic structural distortions. Institute and European University Institute seminar series * Structural adjustments of the magnitude envis- on financial reform in socialist economies. EDI, World aged cannot be achieved without high losses in Bank, Washington, D.C. employment and real income. Allocation of loss- - 1990. "Recent Developments in Yugoslavia." Paper es within and among regions, institutions, aiid presented at a symposium on central banking issues in emerging market-oriented economies. Sponsored by the groups stretches the limit of political will and Federal Reserve Bank of Kansas City, August 23-25, consensus, without which economic reform can- Jackson Hole, Wyoming. not proceed. - . 1991. 'Bank Restructuring in Yugoslavia." Paper pre- sented at the seminar for senior bank supervisors sponsored Notes by the World Bank and the U.S. Federal Reserve, October 28-November 15, Washington, D.C. IMF (International Monetary Fund). Various years. 1. Until 1991 the Socialist Federal Republic of Yugoslavia com- International Financial Statistics. Washington, D.C. prised six republics and two autonomous republics, with a total Kessides, Christine, Timothy King, Mario Nuti, and Catherine population of 23.7 million people in 1989. In 1991 the Sokil, eds. 1989. Financial Reform in Socialist Economies. republics ofCroatia and Slovenia declared independence. Economic Development Institute Seminar Series. 2. Gross social producr, which is 7 to 8 percent less than Washington, D.C.: World Bank. gross domestic product (GDP), is equivalent to gross material Klaus, Vaclav. 1991. "A Perspective on Economic Transition in product (total value added in the production of goods) plus the Czechoslovakia and Eastern Europe." In Stanley Fischer, services regarded as productive inputs. Dennis de Tray, and Shekhar Shah, eds., Proceedings of the 3. Such credits were the dinar counterpart of individuals' for- World Bank Annual Conference on Development Economics eign currency deposits with the banks, which were redeposited 1990. Washington, D.C.: World Bank. with the central bank. This scheme was discontinued in 1988. Knight, Peter T 1983. "Economic Reform in Socialist Countries: 4. These comprised an export bank that was funded by the The Experiences of China, Hungary, Romania, and basic and associated banks, various savings banks that offered Yugoslavia." World Bank Staff Working Paper 579, loans to households and whose investments were restricted to Management and Development Series 6. Washington, D.C. bank deposits, and investment loan funds, which financed pro- Kos, Miroslav, and Milan Cvikl. 1991. "Banking Supervisory jects in less-developed regions. Insurance companies, with negli- Activities in Yugoslavia." National Bank of Slovenia, gible total assets, offered life, general, and casualty coverage but Ljubljana. could invest only in bank deposits. Lahiri, Ashok Kumar 1991. "Yugoslav Inflation and Money." 5. This sizable outflow led to debt rescheduling in 1987. International Monetary Fund, European Department, Washington, D.C. References Mates, Neven. 1991. "Inflation in Yugoslavia: A Specific Form of Public Deficit Caused by the Parafiscal Operations of the Bicanic, Ivan. 1989. "Systemic Aspects of the Social Crisis in Central Bank." Institute of Economics, Zagreb. Yugoslavia." In Stanislaw Gomulka, Yong-Chool Ha, and National Bank of Yugoslavia. Various issues. Quarterly Bulkltin. Cae-One Kim, eds., Economic Reforms in the Socialist Sarajevo. World Armonk, N.Y.: M.E. Sharpe.. . Various issues. StatisticalSurvey. Sarajevo. Bole, Velimir, and Mitja Gaspari. 1991. "The Yugoslav Path to Rocha, Roberto De Rezende. 1991. "Inflation and Stabilization High Inflation." In Michael Bruno, Stanley Fischer, in Yugoslavia." Policy Research Working Paper 752. World Elhanan Hclpman, Nissan Liviatan, and Leora (Rubin) Bank, Washington, D.C. 146 BANK R1sTRUCrURING Saldanha, Fernando, 'Self-Management: Theory and Yugoslav Bank Restructuring." Economic Development Institute Practice." World Bank, Financial Operations Department, senior policy seminar on financial reform in transitional Washington, D.C. socialist economies, September 10-12, Paris. Sheng, Andrew. 1990. 'Bank Restructuring in Transitional Transition: The Newsletter About Reforming Economies. 1991. Socialist Economies: From Enterprise Restructuring to World Bank, Vol. 2, No. 1. CHAPTER 10 Financial Liberalization and Reform, Crisis, and Recovery in the Chilean Economy, 1974-87 Andrew Sheng Chile is one of the most open, dynamic, and stable private sector-led growth, through the encourage- economies in Latin America today. But to get there, ment of competition and foreign direct investment. Chile had to follow a long path of political and eco- nomic reform, crisis, and recovery (Bianchi 1992). What Went Wrong? Financial liberalization was part of that process, and the techniques of bank restructuring that Chile There is broad agreement on what went wrong in adopted have important implications for developing Chile (Velasco 1991). Financial, monetary, and eco- countries. nomic policies supported rapid expansion and exces- The Chilean case must be viewed against the sive risk-taking by domestic and foreign financial backdrop of the remarkable institutional reforms and institutions, which made large loans to undercapital- policy changes implemented over the past twenty-five ized and highly indebted enterprises. Neither a macro- years. The Allende administration (1970-73) nation- economic nor a microeconomic explanation of the alized the entire financial system, as well as the min- crisis is wholly satisfactory. Viewed from a long-term ing, transport, and communications sectors. The perspective, Chile's experience illustrates how doctri- Pinochet military regime that seized power in 1973 naire swings in policies, on top of external shocks and adopted an aggressive program of liberalization and fragile domestic economic conditions, can exacerbate privatization. It was during this regime (1973-81) the consequences of financial boom and bust. that the conditions for widespread financial instability The Chilean case, particularly in 1980-81, developed (Harberger 1985). Those conditions erupt- exhibits many of the characteristics associated with the ed into economic crisis in 1981-83, exacerbated by business-cycle theory of escalating enterprise indebt- enterprise failures, a sharp decline in Chile's terms of edness leading to speculation and eventually crisis trade, a withdrawal of international sources of credit, (Minsky 1988; Kindleberger 1978). It also suggests and sharply rising interest rates, capped by a massive that governments, enterprises, and financial markets currency devaluation. These problems were resolved have long learning curves (Valdes-Prieto 1992). only after the public sector resumed extensive supervi- Policymakers must learn that imperfect markets do sion of and intervention in the financial system. not react like textbook markets. Long-repressed finan- During the second half of the 1980s a pragmatic cial markets are so informationally and institutionally policy approach led the way to recovery through an imperfect that they cannot immediately adapt to the outward-oriented, export-led growth strategy, rein- market-pricing mechanism for credit (de la Cuadra forced by prudent monetary and fiscal policies aimed and Valdes-Prieto 1989). Similarly, enterprises must at maintaining low inflation and a healthy balance of learn that speculative bubbles do burst and that exces- payments. Recovery was aided by an emphasis on sive indebtedness cannot be sustained forever. 147 148 BANK RESTRUcTruRNG A number of factors converged to cause the crisis. nal tariff from 105 percent at the end of 1973 to a Although some were more important than others, uniform rate of 10 percent by 1979 (Cortes-Douglas each played a part in the economic downturn. 1990). This rapid liberalization created large losses for the previously protected import-substituting Realsectorfactors enterprises and sparked an import boom for con- sumer durables, financed by the first-time availability High commodity concentration. The economy was of consumer credit under financial liberalization highly vulnerable to fluctuations in international (Schmidt-Hebbel 1988). demand for a single commodity, copper, which An overvalued exchange rate, coupled with excessive accounted for 83 percent of total exports in 1973. wage rzgidity. The use of the exchange rate as a stabi- Chile's boom and bust cycle, especially in the early lization tool led to a large overvaluation, which years, was closely related to the international price increased demand for nontradables and reduced the of copper. competitiveness of domestic industries, especially International developments. Like many other among exporters and previously protected import- developing countries, Chile was severely affected by competing firms (Galvez and Tybout 1985). Between the international recession of 1981-82 (table 10.1). the second quarter of 1979 and the end of 1981 the When the OECD countries decided to tackle infla- real exchange rate appreciated by 45 percent. In addi- tion by using tight demand management, interna- tion, wage indexation, introduced by the 1979 labor tional interest rates were driven sharply upward. The law, increased enterprise losses at a time when Chile subsequent international liquidity squeeze was com- was suffering sharp falls in export prices and sharp pounded by the debt crisis ignited by Mexico in increases in international interest rates (Edwards 1982, when international banks withdrew credit to 1985). In August 1981, with inflation near zero and many highly indebted countries, especially those in export prices down because of the international reces- Latin America. sion, wages were increased by 14 percent as required Trade reform. An aggressive trade liberalization by law, worsening international competitiveness and program initiated in 1975 brought the average nomi- the financial position of enterprises. Table 10.1 Macroeconomic indicators, 1970-90 (percent unless otherwise specified) Real Current Terms of Nominal interest rate growth accountl trade Inflation Government Exchange rate Year in GDP GDP (1980=100) rateb balanceGDP Lending Deposit (pesosUS. dollar) 1970 2.1 -1.1 238.3 32.5 -3.0 - - - 1971 8.9 -1.9 189.9 20.0 -8.0 - - - 1972 -1.2 -3.9 174.3 74.8 -13.0 - - - 1973 -5.6 -2.7 211.4 361.5 -7.0 - - - 1974 1.0 -2.6 153.0 504.7 -5.0 - - 0.8 1975 -12.9 -6.8 108.1 374.7 0.1 - - 4.9 1976 3.5 1.5 108.7 211.8 1.4 - - 13.1 1977 9.9 -4.1 100.7 91.9 -1.0 163.1 93.8 21.5 1978 8.2 -7.1 95.2 40.1 -0.2 86.1 62.8 31.7 1979 8.3 -5.7 109.9 33.4 4.8 62.1 45.1 37.2 1980 7.8 -7.1 100.0 35.1 5.4 47.1 37.4 39.0 1981 5.5 -14.5 86.4 19.7 2.6 52.0 40.8 39.0 1982 -14.1 -9.5 79.7 9.9 -0.1 63.9 47.9 50.9 1983 -0.7 -5.6 85.8 27.3 -2.6 42.8 27.9 78.8 1984 6.3 -11.0 80.8 19.9 -2.9 38.3 23.3 98.7 1985 2.4 -8.3 79.3 30.7 -2.3 41.3 32.2 161.1 1986 5.7 -6.8 72.8 19.5 -1.0 29.7 20.1 193.0 1987 5.7 -4.3 77.9 19.9 0.5 38.3 26.6 219.5 1988 7.4 -0.8 94.2 14.7 -0.2 21.2 15.1 245.0 1989 9.9 -3.0 97.9 17.0 1.5 35.9 27.7 267.2 1990 2.2 -2.8 82.3 26.0 0.8 48.8 40.3 304.9 - Not available a. Ratio of the index of average export prices to the index of average import prices. b. Consumer price index. Source: IMF, various years. Financial Liberalization and Reform, Crisis, and Recovery in the Chilean Economy 1974-87 149 Premature opening of the capital account, which Financialsectorfactors allowed excessive foreign borrowing by the private sec- tor. This point is controversial. Larrain Garces (1988) Financial liberalization without adequate supervi- argues that foreign banks were inexperienced in recy- sion. Financial liberalization was wide-ranging and cling funds from the oil-exporting countries and lent included bank privatization, licensing of new domes- large amounts to countries like Chile without ade- tic and foreign banks, desegmentation of financial ser- quately assessing their debt servicing capacity. In addi- vices, progressive reduction of banks' reserve require- tion, Chilean firms were not experienced with overseas ments, creation of index-linked financial instruments, borrowing and accumulated excessive external debt. gradual freeing of capital flows, and elimination of Valdes-Prieto (1992), on the other hand, argues that it interest rate ceilings, credit allocation controls, and was the implicit foreign exchange guarantee originat- barriers to entry. In 1974 nineteen domestic banks- ing in the pegged exchange rate policy that was the representing all the domestic banks except Banco del major policy error. Estado-were privatized. "Free banking" (with little Rapid denationalization, which resulted in sub- government supervision) reigned, with new foreign stantial concentration of industrial wealth. entrants spurring heated competition. The freeing of Beginning in 1974, banks and enterprises were sold interest rates in 1975 promoted the development of to the private sector on credit, with 10 to 40 per- finance companies (financieras), under little superviso- cent down and the balance at ten-year (two-year for ry control and in direct competition with banks. banks) indexed loans at 10 percent real interest Between 1974 and 1981 the number of financial rates. Groups (grupos, or economic conglomerates) intermediaries nearly tripled to twenty-three domesti- bought banks on credit and then borrowed from the cally owned banks, eighteen foreign banks, and thir- banks to buy privatized firms. By 1979 the groups teen finance companies. controlled more than half of the largest private cor- Rapid expansion in financial intermediation. porations (Cortes-Douglas 1990). Harberger (1985) Liberalization was extremely successful at expanding and Larrain Garces (1988) point out that many financial intermediation, with total financial assets state-owned enterprises were insolvent when they rising from 20 percent of gross national product were sold to highly geared private owners and so (GNP) in 1975 to 48 percent in 1982 (Velasco had no defenses against the crisis when it broke. 1991). Domestic savings did not expand, however, Speculation in shares and property. Between but averaged only 12 percent of GDP in the high- 1977 and 1981 gross domestic product (GDP) growth 1977-81 period, compared with gross grew by an average of 7.9 percent a year in real domestic investments of 19 percent of GDP. Foreign terms, while the inflation rate fell steadily. The real savings rose from 4 percent of GDP in 1977 to 15 index of stock prices, however, increased by 323 percent in 1981, serving as the unsustainable cause of percent between 1977 and 1980 (Edwards 1985). growth in financial assets. With the onset of the crisis, market capitalization High concentration in the banking system. The five fell by 38 percent in peso terms from its peak in largest banks controlled more than half the credit 1980 to its trough in 1983. In dollar terms market market. The banks controlled by the two largest capitalization declined over the same period from groups (Banco de Chile and Banco de Santiago) $9.4 billion to $2.6 billion and continued its drop accounted for 42 percent of credit and 59 percent of to $2.0 billion in 1985. The conglomerates, in nonperforming loans. The five banks requiring official particular, benefited from low financial costs and intervention in 1983 (Bancos de Chile, de Santiago, high returns on share transactions during the share Concepcion, Internacional, and Colocodora) account- boom (Galvez and Tybout 1985). Between 1979 ed for 45 percent of total credit and 66 percent of and 1981 misguided optimism, first about future nonperforming loans (Hinds 1987). income and later in anticipation of a devaluation, Excessively high real interest rates. Real interest rates spurred accumulation in construction and import- on loans averaged 77 percent a year during 1975-82 ed capital goods (Corbo 1985). The compression (table 10.2). Since such rates were clearly unsustain- in share and property values reduced the net able, there is considerable evidence that distress bor- wealth of enterprises, extinguished collateral val- rowing was going on (Galvez and Tybout 1985). ues, and resulted in large bank losses to overgeared Excessive external debt. As the economy boomed borrowers. during the second half of the 1970s, it relied more 150 BANK RESTRUCTURING Table 10.2 Real peso interest rates, 1975-83 tion. The policy debate encompassed the full range of (percent) options from free banking to protection through Year Loan rate Deposit rate financial repression, with support as well for the pru- 1975 164.9 68.7 dential regulation approach (Cortes-Douglas 1990). 1976 176.4 125.1 Most observers agree that there was inadequate bank- 1977 92.9 58.9 ing sysrem supervision in the period leading to the 1979 23.1 30.6 crisis and that this contributed to the magnitude of 1980 15.7 8.5 the problem. 1981 42.4 32.2 During the early stages of financial liberalization, 1982 42.4 29.7 1983 15.9 f39 for instance, when the free-banking approach pre- 1975-82, 76.6 46.3 vailed, the Superintendency of Banks focused more Note: Rates correspond to short-term (30-90 day) bank transactions, on whether banks were in compliance with quantita- a. Average. tive standards than on whether they were acting pru- Sourct'.Veasco 1991. dently. Only in 1980-81 did the superintendency and more on external financing. In 1981 the current establish a system for risk-based loan classification, account deficit rose to 14.5 percent of GDP, financ- provisioning rules, and exposure limits. Not until ing two-thirds of gross domestic investment. External 1981-82 did it crack down on excessive lending by debt in 1981 climbed by 35 percent-to nearly $15 banks to related companies and approve a more pre- billion-and reached 70 percent of GDP in 1982 cise definition of loan limits to single enterprises. The (table 10.3). By 1982, 73 percent of net foreign debt superintendency also obtained powers to regulate the had been incurred by the private sector. investments, lending, credit rollovers, and collateral Moral hazard. Private sector liabilities, expressed requirements of banks classified as unstable or poorly as domestic deposit liabilities of the banking system managed. But these initiatives came too late to avert plus the external debt of private enterprises, were financial disaster. assumed by the government once the crisis began to The government compounded the error of inade- unfold. The free-banking policy, which assumed that quate regulation and lax enforcement by encouraging bank failure and external debt were private contracts the perception of government readiness to intervene that could be resolved privately, was abandoned. The and underwrite deposits in failing financial institu- Chilean government ended up assuming all private tions. In 1977 regulators took over control of Banco foreign debt and implicitly guaranteeing all deposits Osorno, the first bank failure since 1926. The central in the banking system. Thus a private sector debt cri- bank supported deposits, and the government sis was transformed into a national debt crisis. announced a limited but explicit guarantee scheme for small deposits in banks, supervised finance com- Inappropriate responses panies, and supervised savings and credit coopera- tives. The government also covered the deposit losses Thus by the end of the 1970s economic policy had that had occurred at two major finance companies, swung from wholesale nationalization of the banking although it offered no cover for deposits in failed system to pursuit of unfettered financial liberaliza- informal institutions. In 1977 the government also intervened and protected the deposits in several Table 10.3 Capital inflows and external debt, cooperatives, bailing out all creditors. 1977-82 These structural subsidies had another unintend- (millions of U.S. dollars) ed effect when the exchange rate policy shifted to fixed nominal rates to combat inflation (de la Cuadra and Valdes-Prieto 1989). The deposit guarantee }'ear Capital inflow Total of GDP (percent) scheme protected not only the nominal value of 1977 568 4,862 36.3 deposits but also their external value in real terms. At 1978 1,946 6,407 41.7 a time (1981) when Chilean bank regulators were 1979 2,248 8,201 39.6 unable to impose stringent loan loss provisioning on 1980 3,160 10,746 39.1 1981 4,469 14,483 44.0 banks, the government, through its fixed exchange 1982 1,304 16,953 70.3 rate policy, was assuming vast contingent liabilities Source:Velasco 1991. on the foreign exchange exposure of banks. "Above Financial Liberalization and Reform, Crisis, and Recovery in the Chilean Economy 1974-87 151 the line" government deficits appeared to be elimi- failed in November 1981, and the government inter- nated through orthodox fiscal restraint, with budget vened in their operations. Banco Espanol was among surpluses averaging 4.2 percent of GDP between those taken over, only months after the change in 1979 and 1981. In reality, however, the government's ownership. Significantly, these failures occurred quasi-fiscal deficit was growing rapidly as the govern- before the economic crisis and devaluation of 1982, ment absorbed large contingent liabilities through but were followed by intervention in two more banks implicit or explicit deposit and foreign exchange and a finance company in 1982. guarantees. The magnitude of the impending problem was not perceived early enough. In-mid 1982, at the time The Crisis of the maxi-devaluation of the peso, the new loan classification procedure introduced by the The six-year boom ended abruptly in 1981 as copper Superintendency of Banks in 1980 showed that only prices fell sharply in the wake of the international 6 percent of loans were at risk. Provisions amounted recession. Employment shrank by 22 percent to only 1.7 percent of loans, but the level of capital between mid-1981 and the end of 1982. The trade and reserves suggested that at end-1981 the banking deficit reached 17 percent of GDP following the system still had substantial positive net worth (de la rapid appreciation of the peso. In the first half of Cuadra and Valdes-Prieto 1989). 1982 capital inflows fell to just more than a third of As the recession worsened during 1981-82, bor- their volume in the second half of 1981. Sustained rowing from banks increased sharply to meet domes- appreciation of the peso led to serious problems for tic liquidity needs. Financial system loans to the pri- industries in the tradable goods sectors, increasing the vate sector increased to 62 percent of GDP in 1981, demand for credit and pushing up interest rates. compared with 51 percent in 1980, and jumped to Indeed, between 1978 and 1981 the volume of peso 76 percent by late 1982. International interest rates and foreign currency loans (excluding overdue ones) also rose sharply during 1978-81, affecting the debt increased from 162 billion to 800 billion pesos equiv- servicing of floating-rate debt. Net external debt rose alent. A large share of domestic currency debt was 66 percent in 1981 and nearly 24 percent in 1982. short-term maturities, increasing the liquidity Delinquent loans increased from 2.3 percent of total squeeze on enterprises. loans at the end of 1981 to 6.3 percent by May 1982. Share prices peaked in the first quarter of 1981 The business conglomerates increasingly shifted the and began a sharp decline. Real lending rates climbed collateral for their bank loans from their indebted from 7 percent in the third quarter of 1980 to 44 limited liability corporations and societies to entities percent in the first quarter of 1981. With continued beyond the reach of their banks. Lack of consolidated access to foreign funds, banks borrowed $3.1 billion accounting concealed the extent of interest rate and as construction activity increased 16 percent (follow- foreign exchange exposure of highly leveraged con- ing a 26 percent rise in 1980) and high real wages glomerates. Small and medium-size firms also fell kept domestic consumption rates high. deeply into debt in 1981. Signs of problem loans began to emerge as early Capital flight accelerated in the last two months as 1980. Auditors for Banco Espanol qualified their of 1981, and international credit (particularly to report for 1979 by stating that 37 percent of loans domestic banks) dried up in the first quarter of could not be evaluated because of lack of information 1982. The central bank lost $1.4 billion in reserves on the debtors' ability to pay-even though the loans in 1982 as a whole. Following a June 1982 18 per- had been rolled over repeatedly. Anticipating govern- cent devaluation of the peso, the exchange rate float- ment intervention, the owners sold Banco Espanol to ed for a while, and then moved to a crawling peg. another business group in early 1981. The govern- The peso sank steadily, and by June 1983 it had fall- ment-authorized sale did not require the new owners en by 99 percent. Exchange controls were reintro- to inject new capital in the bank. duced, tariffs were raised to a uniform 20 percent, During the second half of 1981 an important and in early 1983 the peso was declared inconvert- company in good standing with international banks ible (Edwards 1985). At this point the position of went bankrupt. Three banks, four finance compa- Chilean companies was so precarious that authorities nies, and a development bank, together accounting acted swiftly to introduce a preferential exchange for 6 to 7 percent of loans in the financial system, rate for dollar debtors. 152 BANK RESTRucrURING The crisis worsened in 1983. As share prices fell, gramming of enterprise debt that enabled both enter- ten mutual funds collapsed in January, their net prises and the financial system to survive the crisis. book value shrinking from a peak of 27 billion pesos The restructuring effort lasted about five years- ($693 million) at the end of 1980 to 4.3 billion from 1982 to 1987-and spanned the period of eco- pesos ($56 million) in February 1983. Their collapse nomic recession and recovery. Legal and institutional resulted in bankruptcies and large losses to the relat- changes were made to the regulatory and supervisory ed insurance companies, induding the three largest, infrastructure. The thrust of the effort, however, was which had invested up to half their assets in these the socialization of bank losses from bad enterprise funds. In January 1983 the Superintendency of debt through the central bank, financed partly with Banks intervened in eight banks, while the central cash but mostly through official debt instruments of bank had to provide 87.5 billion pesos ($1.1 billion) the central bank. to support bank liquidity. By the end of 1983 the banking system was revealed to be severely insolvent Diagnosis and supervision (table 10.4). It faced losses of 70 billion pesos and had more than doubled the volume of delinquent In 1976 the Superintendency of Banks had only ten loans it was carrying-even after massive sales to the inspectors to supervise the country's fourteen banks central bank. and twenty-six finance companies. Subsequent regu- latory changes improved bank accounting, mandated Crisis Resolution bank auditing, and levied penalties against fraudulent accounts. In 1980-81 authorities introduced long- By early 1983 the need for wholesale government overdue regulations to improve diagnosis and super- intervention and support of the banking system was vision of the system. The central bank also sent forty widely recognized. The government intervened in staff members to bolster the superintendency's seven banks and one finance company (accounting for inspection capacity. A number of other supervisory 41 percent of total deposits), immediately liquidating and regulatory changes were introduced during the three of them. Two of the commercial banks, later crisis. returned to the private sector, were Chile's largest and Risk classification and provisioning. In 1980 the the flagships of two of the largest conglomerates. superintendency set up a risk classification system for Whereas earlier rescues had come with a 100 percent loans. Banks were required to classify loans according guarantee of all deposits and interest, this time the to risk level (A to D) for their 30 largest borrowers. government guaranteed only 70 percent of deposits to This requirement eventually was extended to the 400 domestic creditors (Larrain Garces 1988). The gov- largest borrowers, covering about 75 percent of total ernment did fully compensate foreign creditors' portfolios. Previously, credit rollovers-particularly claims, however, to preserve Chile's international among loans to related companies-tended to hide financial standing. the extent of nonperforming loans. Provisioning was This second round of interventions followed on improved in 1981 by requiring general provisions efforts in 1981-82 to tighten and enforce prudential amounting to 0.75 percent of total loans, in addition regulations, reflecting a belated but important turn- to specific provisions against bad loans. around in official policy. In addition, the government Early warning system. Banks were classified as developed an extensive program for subsidized repro- low-, medium-, or high-risk based on capital adequa- cy, asset quality, management quality, earnings, and Table 10.4 Delinquent loans and bank capital, liquidity (the CAMEL system). 1982-83, selected months Diversification of loans. In 1981 the superinten- (billions of pesos) dency introduced measures forcing banks to shed Loans sold Capital Loan loans from related persons and companies in order to Delinquent to central and loss Net reduce concentration. The concentration limit was Month, year loans bank rrserves provisions worth set at 100 percent of bank capital and reserves. December 1982 41.1 41.6 105.0 34.7 57.0 Banks, however, evaded these limits by channeling January 1983 47.7 37.5 102.9 35.7 53.0 June 1983 126.9 36.9 104.3 55.4 -4.1 funds through newly created overseas subsidiaries. December 1983 92.2 109.7 127.7 61.9 -12.2 Regulatory forbearance. Because of the severity of Source: dela Caudra and Vald&s-Prieto 1989. the crisis, the authorities had to exercise some regula- Financial Liberalization and Reform, Crisis, and Recovery in the Chilean Economy, 1974-87 153 tory leniency. They extended from thirty to ninety to inflation and had to be repaid at an 8 percent real days the threshold for declaring delinquent loans interest rate. nonperforming and stretched the time that banks The 1984 rescheduling lowered interest rates and had to make provisions for these loans from three to expanded the amounts eligible for rescheduling, five years. The authorities also allowed banks more including more favorable terms for small borrowers. flexibility in crediting interest accruals, meeting capi- Real interest rates on the rescheduled portion were tal adequacy standards, disposing of physical assets lowered to 5 percent for the first two years, 6 percent acquired from debtors, and accounting for the effect for years three through five, and 7 percent for the of devaluation on foreign currency liabilities. sixth year. Debt could be rescheduled for no less than Banks, which had accumulated inordinately large five years, and up to ten years for large borrowers and stocks of physical assets as part of their debt recoveiy fifteen years for small borrowers. Across-the-board efforts, were given three years to dispose of these col- restructuring was curtailed after 1984, and banks lateral assets. The law also allowed three years for were encouraged to swap debt for equity. In addition, profits or losses from these disposals to be recognized incentives such as tax concessions and loan loss defer- in the accounts. To minimize the impact of rapid rals encouraged flexibility and facilitated loan restruc- asset disposals on prices and bank profitability, the turing for small debtors. Mortgage refinancing was superintendency allowed the banks five years to also expanded, and the real interest rate was lowered record the losses incurred from collateral asset sales. to 6 percent for large debtors (4 percent for small The superintendency also allowed the banks to debtors). The central bank also provided special lines phase in (between 1982 and 1983) the negative effects of credit at below-market interest rates for working of the currency devaluation on their foreign exchange capital needs, payroll financing, hiring incentives, liabilities. Moreover, the capital adequacy requirement construction and public works, and reforestation. was eased by allowing 25 percent of loan loss provi- Foreign exchange. To cushion the impact of the sions to be included in the capital base. And starting devaluation on distressed firms and banks with net in 1983, cumulative losses at year-end could be writ- foreign exchange liabilities, the central bank in 1982 ten off against capital and reserves within five years. established a preferential exchange rate scheme for debt servicing on certain dollar-denominated debt. Damage control Until mid-1985 preferential rates were indexed to inflation, so that real devaluations increased the The four main types of damage control assisted bor- spread between preferential and market rates and rowers, alleviated foreign exchange pressures, helped thus the size of the central bank subsidy. The subsidy depositors, and aided financial institutions. was paid in the form of central bank notes with a Borrowers. Two major rescheduling exercises cov- maturity of six years. This subsidy was gradually ering an estimated 25 percent of the financial system's phased out. In 1984 the central bank assumed a one- loans were undertaken in 1983 and 1984. In 1983 a time loss when it allowed borrowers who were mandatory 30 percent of the debts of viable borrow- rescheduling their loans to convert dollar liabilities ers in "productive" sectors (which excluded con- into pesos at a favorable rate. sumer, mortgage, foreign, trade-related, and certain T'he bank also provided swap facilities to domes- other loans, as well as loans to investment companies) tic debtors owing foreign liabilities. After an interrup- were rescheduled. Debts were rescheduled for a term tion in foreign loan servicing and repayments in of ten years at 7 percent real interest, with a grace 1983, banks continued to collect on dollar loans. period of five years on principal and one year on Since there was little domestic demand for dollar interest. Debtors in foreign currency obtained the loans, banks deposited their dollars with the central same terms, but the interest rate was computed on bank under repurchase agreements. The central bank the rescheduled debt in foreign currency. Mortgage paid a premium over international market rates on and consumer loans were also refinanced and these accounts, and in mid-1985 it converted the lia- rescheduled by the central bank. Rescheduling cov- bilities into a special account. Peso devaluation ered unpaid installments since 1981 and a declining increased the domestic currency value of the dollar share of installments due during 1983-87. The deposits, creating a central bank loss; however, the rescheduled amounts were denominated a special devaluation adjustments were not paid in cash, but unit of account (UnidaddeFomento, or UFs) indexed accrued. Banks could draw from the currency 154 BANK RESTRUcrURING accounts only for external debt servicing, but they for zero-coupon central bank notes under a ten-year could draw on central bank credit lines denominated loan repurchase agreement (at their initial value) or in the inflation-adjusted UFs at predetermined inter- for cash under a ten-year repurchase agreement at 5 est rates. At the end of 1986 the amount of special percent (real) interest. Shareholders had to commit to accounts was equivalent to 22 percent of financial repurchasing all loans sold to the central bank. In system loans. The premium on the special accounts other words, the loans that were carved out by the was eliminated in 1987. central bank became future liabilities of past share- Depositors. To prevent a systemwide run on banks holders rather than of the general taxpaying public. following its January 1983 intervention in the bank- In 1985 loan sales were opened to all banks (not just ing system, the state issued an explicit guarantee of all those in which the government had intervened), but domestic deposits. This explicit guarantee was lower as of August 1985 banks under intervention account- than the implicit 100 percent protection that had ed for 75 percent of the $2.36 billion of central bank prevailed before. Domestic depositor losses from the loan purchases. three financial institutions that were liquidated aver- In early 1985 the Superintendency of Banks was aged 30 percent. The guarantee was extended until given the authority to require banks subject to inter- April 1985, when the deposit insurance scheme was vention to increase their equity. New shares were overhauled. offered first to existing shareholders, and then to third Ailing banks. The most innovative feature of the parties at favorable discounts under extended easy Chilean restructuring was the part-flow, part carve- payment terms, with tax incentives. Bank creditors out solution used for the banks. To ease liquidity could capitalize the amounts owed by converting during the crisis (the flow solution), the central bank them into equity. Finally, the state development bank, provided emergency lines of credit to banks and sub- Corfo, could take on a bank's outstanding central sidies on rescheduling of loans and other credit pro- bank emergency credits and convert them to equity grams. To facilitate the rescheduling, the central bank shares. Corfo's stake could not exceed 49 percent of sold six-year notes to the commercial banks at a 12 bank equity, and it had to sell its shares within five percent real rate and bought ten-year commercial years or transfer them free of charge to shareholders bank notes at a 5 percent real rate (with five years' who had initially subscribed to new capital. The trea- grace on principal and one years' grace on interest sury would reimburse part of any loss the central payments). This provided banks with a 7 percent bank might eventually assume from these transac- annual risk-free spread. tions. These sales popularized share ownership of banks and attracted more capital than was anticipat- Recapitalizing banks ed. All the banks under intervention were returned to the private sector, although Colocadora lost its auton- The carve-out efforts by the central bank involved omy when it was merged with Banco de Santiago. direct purchases of substandard loans. In June 1982, immediately following devaluation, the central bank Who bore the losses? purchased the worst loans (up to 100 percent of capi- tal and reserves) of the banks in which it had inter- Looking back, it is apparent that the exceptional eco- vened, using ten-year zero-coupon notes (letras). This nomic performance of 1979-81 was unsustainable, gave banks a ten-year period in which to write off consumption-led, and financed principally through loans and avoided costly losses that would have erod- foreign currency borrowing. These factors ensured a ed bank capital. This scheme was replaced in early particularly hard landing when Chile's terms of trade 1984 by one that allowed the central bank to pur- worsened. chase loans up to 150 percent of capital and reserves. Chiiean banks expanded their loan portfolios by The banks used the cash to pay off their outstanding about 500 percent during 1977-81, while the ratio central bank credits and to purchase interest-bearing of bank loans to GDP increased from 15 percent to central bank promissory notes (pagares). This method 55 percent. Credit expansion was driven by unsus- stanched the monetary effect of the central bank's tainably high interest rates. The interest rate issue is cash infusions into the banking system. important. The excessively high interest rates meant Banks could sell additional substandard loans (up that enterprises could not sustain debt servicing pay- to 100 percent of capital and reserves) in exchange ments. Annual real rates of 40 percent implied that Financial Liberalization and Rfiorm, Crisis, and Recovery in the Chilean Economy 1974-87 155 Table 10.5 Potential gains from fully leveraged As bad loans accumulated, banks sought high stock market speculation, 1976-82 spreads to compensate for lost income and to support Stock market Investment fillyfinanced dividend payments. The authorities realized too late price index at the nominal Gain that the state would have to bear the burden of private Year (1980=100) ptso loan rate' orloss sector speculation: the government was responsible 1976 3.7 3.7 0.0 nor only for its own domestic debt but also for guar- 19778 26.9 97 8.8 anteeing all external debt and the banking system's 1979 52.3 29.4 23.9 deposit base. Thus the state paid high real interest 1980 100.0 43.2 56.8 rates and the burden of state obligations was eventual- 1981 75.5 65.7 9.8 1982 68.3 107.7 l3994 ly passed back into the economy through higher taxes, a. See table 10.1I for nominal lending rates, devaluation, inflation, and unemployment. Source. World Bank and IMF data. The first indication of financial imbalance occurred in 1977, when lending to the private sector lenders intended to recover their principal within began to exceed total money and quasi-money; the two-and-a-half years, suggesting that banks would banking system became overextended and started require borrowers to have an even shorter time hori- borrowing abroad. But net private wealth-that is, zon for capital payback. the value of stocks, money, and quasi-money, less There may be some validity to the view that the loans from the banking system-continued to rise sharp policy swing from nationalization to liberaliza- from 1978 and peaked in 1980 at $5.7 billion (table tion and privatization under highly leveraged condi- 10.6). Net foreign liabilities did not materially tions for the privatized banks and enterprises was ill- change during that period, but gross foreign liabilities timed, given Chile's circumstances. Light supervision, rose to $5 billion-the same amount as foreign pegged exchange rates, and implicit deposit guaran- assets-at the end of 1980. tees during 1974-80 allowed excessive risk-taking In 1981, with the decline in copper and share and external borrowing-moral hazard behavior at its prices and the emergence of banking problems, stock worst. And as many other countries have found, market capitalization dropped by $2.4 billion, reduc- overvaluation of the exchange rate encourages ing net private liquid wealth (table 10.7). M3 contin- resources to move into nontradables. A fully lever- ued to grow, however, and domestic borrowing aged purchase of shares on the Santiago stock market increased by $4.9 billion and net foreign debt by in 1976 would have made substantial profits until $2.5 billion. The net loss to the system was $4.6 bil- 1980 and would not have shown heavy losses until lion, or 14 percent of GDP The devaluation in 1982 1982 (table 10.5). A Kindleberger-Minsky bubble caused further losses in wealth as the value of domes- was clearly at work. tic money declined by $2.7 billion and capital fled Table 10.6 A balance sheet approach to private sector wealth, 1978-90 (billions of U.S. dollars) Type of wealth 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Market capitalization, 4.8 6.6 9.4 7.0 4.4 2.6 2.1 2.0 4.1 5.3 6.8 9.6 13.6 Money and quasi-money 3.0 4.6 7.9 10.6 7.9 8.0 6.9 5.8 6.5 7.5 9.2 10.1 10.9 Privatesectorcredit 3.8 6.3 11.6 16.5 13.8 12.8 11.1 9.8 10.3 10.8 12.6 12.2 12.6 Net private liquid wealth (market capitalization + money and quasi-money -private credit) 4.1 4.9 5.7 1.1 -1.5 -2.2 -2.1 -2.0 0.3 2.0 3.4 7.5 11.9 Foreign assets 1.8 3.3 5.0 5.0 3.5 3.6 3.8 3.2 3.5 3.6 4.3 4.4 7.4 Foreign liabilities 1.8 2.9 5.0 7.5 8.2 9.4 11.4 10.9 11.7 11.3 9.5 6.8 6.6 Net foreign liabilities (foreign assets - foreign liabilities) 0.0 0.4 0.0 -2.5 -4.7 -5.8 -7.6 -7.7 -8.2 -7.7 -5.2 -2.4 0.8 Note:.The balance sheet is priced in U.S. dollars to remove the distortingeffects of domestic inflation and devaluation. In an open economywith full capital mobility, a private investor evaluates changes in wealth against an international numeraire, such as the U.S. dollar. a. Capitalization of the stock market is used as a proxy for real asset growth since other data are not available. Source: Calculated from IMF, various years. 156 BANK RESTRUCTURING Table 10.7 Change in private liquid wealth, percent of GDP, more than three times the domestic 1981-84 money supply (table 10.8). The central bank also had (billions of U.S. dollars) assumed external liabilities equivalent to 38 percent Type of cange 1981 1982 1983 1984 of GDP and had a net foreign liability position Stock market capitalization -2.4 -2.6 -1.8 -0.5 equivalent to 18 percent of GDP By the end of 1986 M3 2.7 -2.7 0.1 -1.1 the central bank had accumulated a portfolio of bad Credit 4.9 -2.7 -1.0 -1.4 and doubtful loans amounting to 640 billion pesos- Private liquid wealth' -4.6 -2.6 -0.7 -0.1 equivalent to 31 percent of total loans, 185 percent (stock market capitalization (-14.1) (-10.7) (-3.5) (-0.5) of capital and reserves of the entire banking system, + M3 - credit) or 20 percent of GDP. Two large banks, Bancos de Financed in part by Chile and de Santiago, accounted for 62 percent of net foreign liabilities -2.5 -2.2 -1.1 -1.8 Chil ant a. Numbers in parentheses arc shares ofGDP (percent). b. A minus sign denotes an increase in net foreign liabilities. The carve-out postponed the full monetary Sourre:Calcurated from table 10.6. impact of the banking crisis, but the process of increasing credit to the banking system undermined the country (net foreign liabilities rose by $2.2 bil- economic and monetary stability. The size of the lion). The loss amounted to 10.7 percent of GDP. intervention eroded public confidence in the finan- The devaluation reduced the real value of domestic cial system and undermined the stability of the debt (in dollar terms), to the loss of domestic exchange rate while raising inflationary expectations. depositholders, whose deposits were guaranteed only The central bank's credit expansion complicated in nominal, domestic currency terms. Accordingly, macroeconomic stabilization efforts by boosting the there was a massive redistribution of wealth between quasi-fiscal deficit. The bank had to walk a fine line borrowers and depositors. between stabilizing inflation and the exchange rate The carve-out effectively transferred the losses of while providing sufficient credit to prop up weak the private sector (particularly foreign exchange lia- banks and let the enterprise restructuring process bilities) and the banking system to the central bank. work itself out. The cost was a ballooning of central Although the government's fiscal deficits remained bank obligations. The stock of outstanding central small (1.0 percent of GDP in 1982, rising to 3.0 per- bank bonds and short-term securities rose from $277 cent in 1984 but declining to 2.4 percent in 1985), miUlion in 1982 to $3.8 billion in 1988, accounting the quasi-fiscal deficits in the central bank's books for 62 percent of total outstanding bonds in Chile. were large. In 1981 total central bank claims on the However painful it was, the carve-out had a num- central government, private sector, and banks totaled ber of positive effects. It bought sufficient time for only 12 percent of GDP. By 1985 the total was 124 the real economy to turn around, and by the end of Table 10.8 Monetary aggregates, 1976-90 (percentage of GDP) Item 1976 1977 1978 1979 1980 1981 Claims on central government 29.1 31.3 19.6 16.2 10.2 4.7 Central bank 28.2 30.5 18.9 15.1 10.1 4.6 Banks 0.8 0.8 0.7 1.1 0.1 0.1 Claims on privatesector 15.6 20.0 26.5 31.9 42.1 50.5 Central bank 6.1 0.7 7.1 6.9 5.1 4.7 Banks 9.5 19.3 19.4 25.1 37.0 45.8 Central bankclaimson banks 2.5 2.1 3.7 2.7 2.4 3.1 Foreign liabilities 12.5 15.5 12.4 14.9 18.0 22.9 Central bank 8.5 10.2 6.6 6.4 4.3 2.6 Banks 4.0 5.2 5.8 8.5 13.7 20.4 Capital accounts 15.2 14.7 16.0 19.9 19.6 17.1 Central bank 6.9 7.7 10.5 14.2 13.3 10.9 Banks 8.3 7.0 5.5 5.8 6.2 6.2 Moneyandquasi-money 20.1 16.0 21.3 23.3 28.6 32.3 Currency 3.5 3.2 3.4 3.2 3.3 3.5 Source: IMF, various years. Financial Liberalization and Refonn, Crisis, and Recovety in the Chilean Economy, 1974-87 157 1986 bank restructuring was virtually complete. By Rescheduled credits as a share of total outstanding financing the purchase of bad loans primarily with credits fell from 22 percent in 1985 to 11 percent in debt instruments rather than with cash, the central 1988. As domestic confidence improved, the stock bank deferred the monetary impact while maintain- market revived and private liquid wealth expanded ing the viability of the financial system. Restrictions from -13 percent of GDP (in U.S. dollar terms) in imposed on bank shares also blunted the monetary 1985 to 43 percent in 1990. effects of the loan purchases by channeling banks' As the authorities encouraged higher levels of future profit streams to repurchases rather than to equity capital, capital-asset ratios in the banking sys- dividend payments. In this way the central bank tem rose from 5.4 percent in 1985 to 7.0 percent in avoided the wholesale bailout of shareholders in 1986, and equity went from 9 percent of assets in restructured and assisted banks. 1985 to 11 percent in 1988. Bank profitability improved considerably as real returns on capital rose Rebuildingprofitability from -5.1 percent in 1985 to 8.4 percent in 1988. The crisis and its resolution left Chile's banking Although the 1982 devaluation had a massive impact structure largely unchanged, although foreign banks on private wealth, it was the return to international made some gains in market share at the expense of competitiveness that generated the resources for recov- domestic institutions. Competition came not only ery. The trade balance went from a deficit of $2.7 bil- from foreign banks, but also from the securities mar- lion in 1981 to a surplus of $1.0 billion in 1983. kets and from securitized products. The number of Tight monetary and fiscal policies, greater openness to domestic banks and finance companies decreased trade and foreign investment, and greater attention to from twenty-five in 1983 to nineteen in 1988, while enterprise and bank restructuring paid off. Between the number of foreign banks rose from nineteen to 1985 and 1989 Chile cut its external debt by $8 bil- twenty-two. Private domestic banks' market share fell lion by permitting citizens to use foreign exchange from 68 percent of the banking system assets in 1983 resources to retire debt, by allowing foreign creditors to 61 percent in 1988. The Banco del Estado, the to swap debt for equity, and by repurchasing debt in one state-owned commercial bank that emerged from the international secondary market. the crisis unscathed, accounted for 18 percent of total Furthermore, economic recovery after 1985 great- system assets at the end of 1988, roughly the same ly improved the aggregate corporate balance sheet, share as in 1983. which permitted a reduction in long-term indebted- As the crisis ebbed and a new banking law was ness. As real growth revived in 1985, the quality of enacted in late 1986, bank supervision reverted to an banks' assets improved. Past-due loans dropped from orthodox stance. The new law promoted self- 8 percent of total assets in 1983 to 2 percent in 1988. regulation through stiffer requirements for market 1982 1983 1984 1985 1986 1987 1988 1989 1990 8.9 15.2 27.9 51.3 52.6 49.1 40.7 32.7 33.2 4.3 10.4 22.1 43.9 46.8 45.2 40.2 32.4 33.0 4.6 4.8 5.8 7.4 5.8 3.9 0.4 0.3 0.3 81.7 71.9 75.4 70.1 65.0 62.2 57.6 53.4 50.2 14.2 14.7 14.7 13.9 12.7 11.2 9.5 1.3 1.3 67.4 57.3 60.7 56.2 52.3 51.0 48.2 52.1 49.0 12.3 38.5 46.3 65.8 58.1 49.0 33.7 27.7 25.6 48.8 52.8 75.5 77.9 74.3 64.6 43.5 30.0 26.3 5.8 17.6 30.3 38.4 35.7 35.8 24.2 16.1 13.9 43.0 35.2 45.2 39.5 38.6 28.8 19.3 13.9 12.4 30.1 32.5 14.1 35.9 34.6 58.2 22.2 23.0 23.7 22.1 26.5 9.1 8.7 8.1 34.2 7.0 7.1 7.4 8.0 6.0 4.9 27.2 26.5 24.0 15.2 15.9 16.3 46.6 45.0 47.1 41.2 40.9 43.1 42.1 44.1 43.6 3.5 3.3 3.4 3.1 3.3 3.3 3.4 3.3 3.4 158 BANK RESTRUcruRlNG information. The Superintendency of Banks was institutions in a few highly leveraged groups. But required to publish information about the nature prolonged overvaluation of the domestic currency to and quality of bank assets three times a year so that stabilize inflation perpetuated market distortions, depositors, investors, and supervisors would be bet- fueling the shift toward nontradables-which was ter informed about financial institutions. The law financed by external borrowing with the opening of also attempted to eliminate the practice of related the capital account. Domestic and international lending and defined more precisely the concept of indebtedness rose to undesirably high levels, made "client." It tightened capital adequacy standards and possible by a favorable external environment and the compliance sanctions, reinforced the leveraging inexperience of international banks in recycling sur- ratios applied to banks, and set out a mechanism for plus petrodollars to developing countries and that of bank rescues by other banks. Banks were also domestic firms in assessing the true costs of foreign required to dispose of acquired assets quickly or to borrowing under an overvalued exchange rate. write them off fully, and loans to related companies Another key policy error lay in the presumption were to be reduced to no more than 5 percent of the that private external borrowing is ultimately a state total portfolio by the end of 1990. obligation. External borrowing financed enterprise Finally, the scope of state deposit guarantees was indebtedness and consumption-led growth while scaled down to protect mainly small depositors. temporarily keeping the exchange rate artificially Demand deposits continue to carry a 100 percent high. Facing pressure from international creditors, guarantee in full, but only 90 percent of time the government assumed the external debt obliga- deposits per depositor are protected up to UF 120. tions and, in addition, bailed out distressed enterpris- State insurance on larger deposits was phased out. es through a preferential exchange rate subsidy Consequently, while the state guarantee on time administered through the central bank. deposits covers only 7 percent of total system bal- The Chilean economy was already in a fragile ances, it protects 80 percent of savings deposits and state when the terms of trade deteriorated rapidly in covers two-thirds of all depositors. 1981. A sharp jump in international interest rates and the abrupt withdrawal of international credit Lessons ensured a downward spiral of asset prices. Could bet- ter regulatory and supervisory vigilance have averted The Chilean case shows the immense difficulties of the financial crisis? Chile's experience provides useful shifting from a highly nationalized and repressed lessons on the dangers of financial liberalization with- economy into a fully open, liberalized, and privatized out strong supervision and enforcement to avoid system. Highly leveraged and decapitalized banks and such market abuses as related lending, risk concentra- enterprises do not necessarily seek long-run, produc- tion, and capital flight. tive investments. They are more likely to engage in Free banking with few controls on capital mobili- short-run, high-risk activities in order to recapitalize ty reduced the capacity to tax entrepreneurs and themselves. The presumption that the market would wealthholders, while state guarantees on domestic correct structural deficiencies through high real inter- deposits and external borrowing worsened moral haz- est rates proved wrong so long as the state provided a ard and free-rider problems. Large concentrations of guarantee on bank deposits and private sector exter- wealth were built up and used for speculative purpos- nal borrowing. Large bank spreads between deposit es. When losses occurred capital fled the country and loan rates weakened the enterprise sector and dis- through the open capital account, so much of this couraged household savings without improving the wealth evaded taxation. Thus the losses were borne long-run profitability of banks. by the central bank. The Chilean case has important lessons for the The Chilean case also demonstrates the impor- post-centrally planned economies. Because of hyper- tance of national risk management. Policymakers inflation, radical economic stabilization measures should focus on preventing excessive risk-taking and were needed at the same time as the market was the accumulation of large financial imbalances, being liberalized. For example, rapid privatization whether in domestic currency or in foreign exchange, transferred loss-making public enterprises to the pri- in both state and private sectors. Because of a doctri- vate sector. Simultaneous bank privatizations led to naire belief that a free market would correct itself, the the concentration of both enterprises and financial authorities ignored the dangers that capital flight Financial Liberaliztion and Rferm, Crisis, and Recovery in the Chilean Economy 1974-87 159 could destabilize the economy by eroding public con- Edwards, Sebastian. 1985. "Stabilization with Liberalization: An fidence in the banking system-at a time when the Evaluation of Ten Years of Chile's Experiment with Free- state had to absorb all losses and had limited Market Policies, 1973-83." Economic Development and state had to absorb all losses and had llmlted Cultural Change 32 (January): 223-54. resources to deal with the losses. Galvez, Julio, and James Tybout. 1985. "Microeconomic Thus macroeconomic stabilization and financial Adjustments in Chile during 1977-81: The Importance of liberalization policies have to take into account the Being a Grupo." World Development 13(8): 969-94. structural deficiencies in real and financial markets. Harberger, Arnold C. 1985. "Observations on the Chilean The institutional framework for regulation and Economy, 1973-83." Economic Development and Cultural Change 33(3): 451-62. supervision must be in place to allow the restructur- Hinds, Manuel. 1987. "Financial Crises in Developing ing of enterprises, labor, and financial markets with- Countries." World Bank, Country Economics Department, out disrupting monetary stability. Bank restructuring Washington, D.C. was ultimately a success in Chile, but it came at con- IMF (International Monetary Fund). Various years. siderable cost following the successful opening of the International Financial Statistics. Washington, D.C. Chierablean economy tol ingthernationl competition and Kindleberger, Charles P. 1978. Manias, Panics, and Crashes: A Chilean economy to international competition and AHistory of Financial Crises. New York: Basic Books. massive changes in tax, pension, and labor structures. Larrain Garces, Mauricio. 1988. "Treatment of Banks in Difficulties: The Case of Chile." Background paper pre- References pared for World Development Report 1989. World Bank, Washington, D.C. Bianchi, Andres. 1992. "Chile: Economic Policies and Ideology Minsky, Human P. 1982. "The Potential for Financial Crises." During the Transition to Democracy." In Pedro Aspe, Department of Economics Working Paper 46. Washington Andr6s Bianchi, and Domingo Cavallo, eds., Sea Changes University, St. Louis, Missouri. in Latin America. Washington, D.C.: Group of Thirty. Schmidt-Hebbel, Klaus. 1988. "Consumo e Inversi6n en Chile Corbo, Vittorio. 1985. "Reforms and Macroeconomic (1974-82): Una Interpretaci6n real del Boom." In Felipe Adjustments in Chile during 1974-84." World Morande and Klaus Schmidt-Hebbel, eds., Del Auge a la Development 13(8): 893-916. Crisis de 1982. Santiago: ILADES. Cortes-Douglas, Hernan. 1990. "Financial Reform in Chile: Valdes-Prieto, Salvador. 1992. "Financial Liberalization and the Lessons in Regulation and Deregulation." Paper presented at Capital Account: Chile, 1974-84." Paper presented at the the seminar on financial sector liberalization and regulation, conference on the impact of financial reform sponsored by "Changing the Rules of the Game." Sponsored by the World the World Bank, April, Washington, D.C. Bank's Economic Development Institute, Washington, D.C. Velasco, Andres. 1991. "Liberalization, Crisis, Intervention: de la Cuadra, Sergio, and Salvador Valdes-Prieto. 1989. "Myths The Chilean Financial System, 1975-85." In and Facts about Instability in Financial Liberalization in Vasudevan Sundararajan and Tomas Balino, eds., Banking Chile: 1974-83." Universidad Catolica de Chile, Institute Crises: Cases and Issues. Washington, D.C.: International of Economics, Santiago. Monetary Fund. CHAPTER 1 1 Argentina's Financial Crises and Restructuring in the 1980s Luis A. Giorgio and Silvia B. Sagari Argentina began and ended the 1980s in financial 1992)-the government repudiated a large portion of crises. Financial sector restructuring proceeded con- its private sector debt, decapitalizing many banks in tinuously, with the central bank intervening in, liqui- the process. dating, or forcibly merging about 10 percent of the The economic and financial deterioration of the financial system's institutions each year between 1981 1980s had transformed the financial sector, so the cri- and 1988. One hundred sixty-eight formal financial sis of 1989, although related to the 1980-82 crisis, institutions closed. Deposits in the financial system was qualitatively different. In 1982, 80 percent of became highly volatile, and financial institutions' financial sector assets involved credit to the nonfinan- assets were frozen. Restructuring and the financial cial sector; thus the condition of the financial sector instability that it reflected fueled the internal disequi- was inextricably linked to that of the nonfinancial librium that characterized the decade. This in turn sector. The 1989 crisis, by contrast, was precipitated contributed to the demonetization and disintermedi- by unsound and poorly managed public finances, ation that traditionally accompanies high-inflation coupled with the closing of international credit mar- environments. kets to highly indebted countries. This closure elimi- This vicious circle was caused primarily by the nated one of the few noninflationary instruments government's inability to bring spending in line with that the Argentine government had to finance its current revenues and its attempts to finance this gap deficits in the short term (Cavallo and Pefia 1983). by printing money and issuing more debt. This led to Over the decade the banking system was used to a growing debt service burden and, ultimately, to a finance the central bank, which onlent primarily to quasi-fiscal debt in the central bank (Rodriquez state-owned and other public banks at preferential 1991). The debt had to be financed at higher and rates. This process substituted for direct central bank higher interest rates given the continual erosion of financing of the treasury, made the banking system a demand for money or debt instruments denominated captive of public finance, and displaced private, mar- in local currency. The rising cost of borrowing ket-based banking activity. Banks basically stopped increased public sector borrowing needs and under- lending to the nonfinancial sector and instead trans- mined exchange rate stability, which led in turn to ferred resources to the central bank. By the end of higher interest rates. The tight monetary policy of 1988 the central bank held 65 percent of the deposits the 1980s ultimately backfired, since high interest of the financial system. rates eventually fed back into inflation. Caught in a debt-distress trap, the system exploded into hyperin- Macroeconomic Background and Financial Sector flation, which approached an annual rate of 5,000 Framnework percent in 1989. In early 1990-through the forced conversion of commercial bank time deposits into Extensive government intervention in the mobiliza- ten-year, dollar-denominated treasury "external tion and allocation of resources during the 1970s bonds" that traded at a deep discount (Beckerman helped turn Argentina into a high-inflation, slow- 161 162 BANK RESTRUCTURING growth economy (table 11.1; Rodriquez 1991). inflation, demanded higher and higher remuneration During the 1980s real gross domestic product (GDP) for holding financial assets denominated in local fell by an average of 0.9 percent a year. Savings as a currency. Since dollarization increased pressure on the share of GDP dropped because of the sharp fall in exchange rate, the government continually had to raise public savings. Capital formation also steadily interest rates. High real interest rates hurt the prof- declined, from 25 percent of GDP in 1980 to 9 per- itability-and ultimately the solvency-of enterprise cent in 1989 (Beckerman 1992). borrowers, which in turn undermined the stability of banks. External credit also increased the overall The foundations offinancial sector instability indebtedness of the productive sectors of the economy. In 1978 the government adopted an economic Since 1961 the government had been running a pri- program that used the exchange rate as the anchor for mary deficit (Rodriquez 1991). Until 1976 the bud- domestic prices. Preannounced daily devaluations getary shortfall was financed by printing money. In were intended to reduce the gap (more than 1 per- 1976 debt began to play a role in financing the fiscal cent a month) between domestic and international deficit, greatly facilitated by access to foreign borrow- interest rates by minimizing the exchange risk implic- ing in the late 1970s. Borrowing eased the inflation- it in the high domestic rates. Domestic inflation ary impact of deficit financing, but when external meant that international interest rates were negative, financing was cut off in 1982 it inaugurated a fund- causing a massive inflow of foreign capital in 1979. ing crisis that eventually led to hyperinflation Large enterprises increased their gearing ratios and (Beckerman 1992). suffered an erosion of liquidity; those with access to Starting in 1982 government revenues increased external borrowing increased foreign debt as a share and spending grew more slowly than GDP Despite of total assets. Other firms, however, did not switch this relative decline in the primary deficit, the cost of to foreign currency borrowing because of the general financing it remained high, unleashing a public lack of confidence in the appreciating peso exchange finance crisis that had been building for more than a rate-fueling the boom in domestic lending. decade. The stock of debt continued to mount. Savers, Inflation slowed sharply toward the end of 1979, anticipating devaluation and accustomed to endemic but uncertainty about the economic program under- Table 11.1 Macroeconomic indicators, 1970-90 (percent unless otherwise specified) Terms of trades Exchange rate (local Year Realgrowth in GDP Current account/GNP (1980=100) Inflztion rateh currency/US. doLlar) 1970 5.18 -0.71 138.4 13.6 n.a. 1971 4.47 -1.57 141.1 34.7 n.a. 1972 1.19 -0.89 126.8 58.5 n.a. 1973 3.58 1.82 166.8 61.3 n.a. 1974 5.74 0.22 134.4 23.5 n.a. 1975 -0.66 -3.36 121.4 182.9 0.000003 1976 0.00 1.73 119.6 444.0 0.00001 1977 6.32 2.72 109.9 176.0 0.00004 1978 -3.52 4.37 104.3 175.5 0.00007 1979 6.75 -1.01 107.0 159.5 0.0001 1980 0.62 -8.50 100.0 100.8 0.0001 1981 -9.80 -8.43 97.9 104.5 0.0004 1982 -6.66 -4.52 90.2 164.8 0.002 1983 2.18 -4.10 96.2 343.8 0.01 1984 2.63 -3.45 96.9 626.7 0.06 1985 -3.28 -1.57 89.7 672.2 0.60 1986 7.28 -3.84 85.3 90.1 0.94 1987 2.15 -5.54 81.7 131.3 2.14 1988 -2.83 -1.78 86.2 387.7 8.75 1989 -8.00 -2.40 89.6 4923.6 423.0 1990 3.48 1.80 91.5 1343.9 4876.0 n.a. Not applicable. a. Ratio of the index of average export prices to the index of average import prices. b. Consumer price index. Source: National Institute of Statistics and Census; Central Bank of Argentina. Argentina's Financial Crises and Restructuring in the 1980s 163 pinned the 31 percent real interest rate reached by peso was allowed to float. Controls were briefly reap- the beginning of 1980. As early as 1979 the real sec- plied when the Falkland Islands conflict with Great tor of the economy had started to experience prob- Britain began in April 1982, but were lifted again lems from the incompatibility between real interest toward the end of 1982. Appreciation of the real rates and the profit rates of companies that were effective exchange rate had been halted, and decline highly indebted in domestic currency. had begun. The nominal anchor exchange rate policy resulted Nevertheless, the erratic nature of exchange rate in a steep real appreciation of the peso, which distort- policies contributed to a general lack of confidence in ed relative prices by encouraging import substitution stabilization efforts. Devaluation brought increased and production of nontradables. The import boom foreign exchange to the central bank, with an atten- and export slump caused a current account deficit at dant impact on the money supply. The banking sys- the end of 1979, exacerbated by the effect of the tem bore the brunt of ensuing efforts at monetary con- increasing external debt service burden on the invisi- trol, while the increasing need for public sector bor- bles account. The onset of an international recession rowing crowded out the private sector. Devaluation in 1981 affected the terms of trade of internationally also increased the debt service burden of both the pub- traded goods as well. High real interest rates, compe- lic and private sectors. Enterprises had such problems tition from imported goods, and an increase in real servicing their debt that the central bank assumed the wages lowered profits and increased the incidence of private sector's external debt. Domestic saving was insolvency for a wide range of enterprises (Balifio declining, boosting internal financing costs. And 1987). Expectations of devaluation merely increased though government expenditures were cut in the early the pressure on the financial system. 1980s, tax revenues began to decline, so no real There were two large devaluations in the first half progress was being made in economic restructuring. of 1981. After the second one the exchange market was split into a commercial market-for which the Financial sector growth and instability rate was set by the central bank-and a free-floating financial market. The government also introduced Financial institutions expanded rapidly throughout an exchange insurance scheme to encourage the this period of uncertainty and mounting real sector rollover of longer-term loans (more than eighteen problems (table 11.2). Despite demonetization and months) and compensated borrowers for the effects disintermediation, the number of banks increased of the June 1981 devaluation on loans rolled over for because of the demand for more intensive intermedi- at least one year. These initiatives were augmented ation brought about by the general internal disequi- later in the year by a swap facility for six-month librium. By the end of 1989 real money was worth operations. By the end of 1981 the foreign exchange about one-third its value at the end of 1979. At the market was reliberalized and most of these short- same time, the average term of deposits fell from 310 term stabilization techniques were eliminated. In to 10 days and the velocity of circulation of sight addition, the exchange market was unified and the deposits doubled as the public acted defensively Table 11.2 Financial sector branches and staffing, 1979-89 Government-owned banks Private banks Total banks Nonbankfinancial institutions Personnel Personnel Personnel Personnel Year Branches (thousands) Branches (thousands) Branches (thousands) Branches (thousands) 1979 1,813 79 2,032 63 3,845 142 261 10 1980 1,840 80 1,999 66 3,839 146 280 11 1981 1,876 79 2,046 63 3,922 142 277 11 1982 1,877 78 2,215 63 4,092 141 272 9 1983 1,892 79 2,481 66 4,373 145 268 7 1984 1,914 81 2,645 68 4,559 149 231 6 1985 1,935 81 2,568 64 4,503 145 193 4 1986 1,953 80 2.540 61 4,493 141 164 4 1987 1,969 80 2,411 58 4,380 138 138 3 1988 1,990 82 2.458 61 4,448 143 92 3 1989 2,003 82 2,480 65 4,483 147 44 2 Nate. Excludes the central bank. Sourcre: Central Bank of Argentina. 164 BANK RESrRUCrURING against high inflation. Though the operational effi- also stopped collecting its charge rate on banks, ciency of many banks declined steadily and bank clo- although the regulation account continued to pay sures escalated, the number of bank branches rose by interest on reserves. Although the central bank had 638 between 1979 and 1989. prevented a complete decapitalization, the banking With the nationalization of deposits in 1973, the system was once again highly repressed and suffered central bank became the only depositholder until extensive disintermediation. 1977. During that time financial entities played the role of deposit collectors, for which they received a The Austral Plan commission; depositors collected interest from the central bank. Banks could lend capital, reserves, and The Austral Plan, introduced in mid-1985, affected rediscount funds, but the deposits were in effect sub- the structure of the financial system through the end ject to a 100 percent reserve requirement. The finan- of the 1980s. The plan included a wage and price cial system contracted as depositors looked for alter- freeze; devaluation, followed by a fixed exchange rate; natives to bank deposits, which they found in the creation of a new currency, the austral, to replace the form of indexed government bonds. peso; and fiscal reforms, including a commitment by In 1977 changes were introduced to the laws the central bank not to print money to finance trea- affecting financial institutions, especially with respect sury deficits (Beckerman 1992). The central bank to solvency and liquidity, monetary and credit policy, began using the financial system as the main supplier and entry and bank branching. Deposit insurance of funds, and in so doing (temporarily) lowered the also was introduced, initially covering 100 percent of inflation rate from 672 percent in 1985 to 90 percent peso-denominated deposits.' Deposit insurance and in 1986. Eventually, though, the plan also became a the freeing of deposit interest rates aided the remone- source of hyperinflation because the central bank tization of the economy and boosted activity in the assumed several fiscal and quasi-fiscal activities: formerly repressed financial system. * Servicing of the nationalized public and private Selective credit practices were lifted under the external debt, for which it purchased surplus cur- new system, leaving most credit and investment rent account funds. decisions with the financial intermediaries. Banks, * Financing of real sector operations through loans however, were required to pay into a central bank to the banking sector, mainly government-owned monetary regulation account" based on the loan institutions. capacity generated by their sight deposits. The cen- * Redemption of maturing government securities. tral bank used this account to compensate the banks Thus the path to hyperinflation was open. Improving for the reserves they were required to hold against the stability of the banking system was impossible time deposits. Although the reserve requirements without radical fiscal reform, which was not forth- were imposed to counter the expansionary monetary coming. effects of the other changes in policy, these arrange- In June 1985 the financial system began supply- ments later contributed to the central bank's quasi- ing financing through the reserve requirements and fiscal deficit. forced investments imposed by the central bank. Foreign and domestic borrowing accounted for Forced investments took the form of so-called inac- 65 percent of total government financing over the cessible deposits. Unlike reserve requirements, these 1977-81 period. But from 1982 to mid-1985 the deposits could not be drawn down even if a bank's central bank was the sole financier, through expan- deposits declined, and the interest earned on them sion of the monetary base, thus thwarting the effect could not be withdrawn but had to be capitalized of anti-inflationary initiatives. By 1983 the fiscal into the deposit balance. They had the advantage, deficit had increased to 16 percent of GDP. therefore, of not adding to high-powered money. Moreover, the devaluations in 1982, and ensuing And unlike the previous system of reserve require- assumption of private external debt by the central ments, the inaccessible deposits did not lead to disin- bank, led to a reversal of public policy regarding the termediation of funds from the formal sector.2 banking sector. Deposit interest rates once again The system once again became tightly repressed, became subject to controls, while lending rates were not unlike it was before the liberalization of the mid- depressed by a compulsory rescheduling of private 1970s. It was also inherently unstable. Whereas in loans at negative rates. At this point the central bank June 1982 reserves had amounted to only 3 percent Argentins Financial Crises and Restructuring in the 1980s 165 Table 11.3 Consolidated financial system balance crises of other Latin American countries at the time. sheet The rapid drop in inflation in late 1979 and 1980 (percent) pushed real interest rates to more than 31 percent a Balance June D-cember December July December year. This worsened the impact of the recession sheet item 1982 1982 1985 1988 1988 caused by competition from imports, which itself was A rets caused by real effective exchange rate appreciation. Total reserves 3 14 12 19 25 The extent of nonperforming loans and the number othercredits 78 73 71 60 46 of bankrupt firms had a direct impact on the prof- Fixed assets 7 6 8 9 7 itability and solvency of financial institutions. Loan Other assets 12 7 9 12 22 portfolios were deteriorating in quality just as finan- Total 100 100 100 100 100 cial institutions were experiencing a steep increase in Liabi&ities Deposits 39 17 33 32 33 liabilities, resulting in severe capital losses to banks Rediscounts 5 16 9 12 13 and other financial institutions. The first indication Other liabilities 47 60 46 39 36 of a crisis came in March 1980 with the bankruptcy Capital and reserves 9 7 12 17 18 of one of the largest private banks in Argentina, Total 100 100 100 100 100 Banco de Intercambio Regional (BIR). Note: Exdudes institutions in liquidation and the central bank. Sourre: Central Bank of Argentina. The seeds of crisis of assets, by December 1988 reserves held by the cen- To finance the progressively heavier loan portfolio, tral bank accounted for about a quarter of the assets institutions such as BIR initiated aggressive cam- of the consolidated financial system (table 11.3). The paigns to attract depositors by offering increasingly nature of credit risk also changed as loans and other higher interest rates (Arnaudo and Conejero 1985). credit fell from 78 percent of assets in June 1982 to The supervisory and institutional framework at the 46 percent in December 1988. Loans to the nonfi- time encouraged this practice in a number of ways. nancial private sector fell by half in real terms during First, deposit insurance covered 100 percent of local this period. currency deposits, making depositors risk-neutral The stock of the financial system's liabilities with investors for whom the solvency of their financial the central bank also increased: rediscounts alone rose intermediary was irrelevant (Fernandez and from 5 percent of total liabilities in June 1982 to 13 Rodriquez 1983). Moreover, there was a perception percent in December 1988. Rediscounts were ear- that public banks-whether national, provincial, or marked to both economic activities and financial municipal-were implicitly insured despite the fact intermediaries. In fact, 80 percent of rediscounts were that most did not take part in the central bank channeled to public sector banks, which increased deposit insurance scheme. their share of loans from 35 percent at the end of Second, although the regulatory regime was 1979 to 61 percent by the third quarter of 1989. designed to promote discipline, soundness, and safe- Not surprisingly, the quality of banks' loan port- ty in the financial system, monitoring and enforce- folios deteriorated, with the system's share of nonper- ment could not keep pace with the rapid increase in forming assets increasing from 11 percent at the end the number of financial institutions. Furthermore, of 1982 to 27 percent by December 1988. The posi- supervisors tended to focus on compliance with the tion of private banks improved slightly, with nonper- rules rather than on qualitative assessment of assets forming loans falling from 11 percent to 8 percent, (Balifio 1987). while the nonperforming portion of public banks' portfolios rose from 11 to 37 percent. Restructuring The Banking Crisis of 1980-82 Between the end of 1979 and 1981 twenty-five pri- vate domestic banks and thirty-four nonbank institu- The Argentine banking crisis of the early 1980s was tions went bankrupt. Even financial institutions that different from the crisis of 1989-90 and-in that it survived the crisis had extremely poor-quality portfo- was caused primarily by the private sector's inability lios. At least half the aggregate portfolios of all finan- to service its financial commitments-similar to the cial institutions were nonperforming. The crisis, 166 BANK RESTRUCTURING which started with the collapse of the BIR, was reached $3.2 billion, equivalent to about 7 percent of resolved by liquidating financial entities, reimbursing GDP (table 11.4). all depositors, and transferring resources from the Part of the subsidy ultimately was borne by hold- depositors to the debtors of the system. This was ers of regulated-rate deposits. Since regulated rates done mainly through central bank operations (see were negative in real terms, depositors were taxed below). As a result of these efforts the nonfinancial implicitly. Moreover, the government restricted the private sector was no longer the main debtor of the amounts and terms of deposits yielding market financial system by the end of 1982. (unregulated) interest rates, effectively creating a During 1980 alone the central bank provided liq- niche market for regulated-rate deposits. Although uidating institutions with loans equal to 5 percent of there was a 100 percent reserve requirement on regu- GDP (or 12 percent of the total deposits of the finan- lated deposits, the central bank paid interest to banks cial system). The bank also took over administration on the reserves at the same rate that banks paid to of all institutional assets. The repayment of deposits depositors. The implicit tax paid by depositors was was financed by expanding the monetary base. The equivalent to $2.3 billion. The $0.9 billion difference balancing entry on the asset side of the central bank between this amount and the total subsidy was balance sheet was "credit to financial institutions in absorbed by the central bank through the spread liquidation," but the only assets these institutions had between the inflation rate and the interest rate on the to repay with were their buildings and their loan refinancing of restructured corporate loans. portfolios, which were of extremely poor quality and Institutions that remained in operation also refi- low market value. nanced their nonperforming loans through the cen- These early steps did not stem the tide of banking tral bank, so they were not disadvantaged by the failures, however. By December 1982 eighty-three restructuring operations. Still, the structural changes institutions had been closed down. Central bank under way by the end of 1982 did have a profound loans to the institutions undergoing liquidation impact on them. The real value of monetary aggre- equaled 3 percent of GDP that year. gates had shrunk by half compared with three years During 1981-82 the government assumed private earlier. At the same time, investors began switching foreign debt liabilities. Part of the foreign debt was to dollar-denominated assets and substantially short- exchanged for pesos, converted at a below-market ened their investment horizons for local currency exchange rate The balance was to be paid over time investments. (in pesos) at an interest rate that after 1984 fell below Financial entities ultimately refinanced nearly all market. The foreign debt of public enterprises also the credit granted by the central bank to the nonfi- was transferred to the treasury. The debt service costs nancial private sector. This had a temporary salutary on this exposure were to be financed through money effect on commercial banks, causing the share of creation by the central bank, combined with nonperforming loans in the financial system to drop increased reliance on the financial system to absorb sharply to 11 percent of the total portfolio at the end part of the additional liquidity-thereby crowding of 1982. But the process was a fundamentally infla- out productive investment. tionary remedy. Both the dollarization of the finan- The monetary and financial reform package cial system and the increase in money velocity made introduced in July 1982 was the first systematic it difficult to control the growth of the monetary base attempt to address the solvency crisis. The plan and to counter the upward pressure on interest rates, focused on alleviating the indebtedness of private two dilemmas that persisted throughout the decade. enterprises in order to forestall further banking fail- ures. Debts were liquified through refinancing at neg- The Road to the 1989 Crisis ative real interest rates with terms extended beyond five years. By this time, however, the average term for The apparent stabilization of the financial system was deposits was only about forty days-a highly unsta- short-lived. In June 1985 the Austral Plan introduced ble situation for banks. To stabilize the banks' posi- a direct assault on inflation-but without providing a tion and to dampen upward pressure on interest strong commitment to fiscal reform. Declining rates, loans were refinanced by the central bank at the money demand undermined efforts to restrain mone- same rates and terms as were extended to the private tary growth in 1986 and early 1987, adding to infla- borrowers. The subsidies implicit in this restructuring tionary pressure. Moreover, public spending increased Argentinas Financial Crises and Restructuring in the 1980s 167 Table 11.4 The transfer of resources between debtors and creditors of the regulated financial system, July-December 1982 Credit to Implicit subsidy Interest- Real thepinvate Real to debtors, bearing deposits deposit Implicit tax on savers Net result' sector (mil- lending of the public interest lions of interest rate (millions (billions (millions of rateh (millions (billions (millions (millions Month australes) (percent) ofaustrales) of dollars) austrates) (percent) ofaustraks) of dollars) ofaustraks) of dollars) July 34.2 17.1 5.9 1.0 20.9 -18.3 3.8 0.7 2.1 346 August 37.1 8.7 3.2 0.7 21.5 -9.8 2.1 0.4 1.1 231 September 38.6 9.4 3.6 0.8 22.3 -10.8 2.4 0.5 1.2 254 October 40.4 1.7 0.7 0.1 24.0 -3.3 0.8 0.1 -0.1 -18 November 47.3 4.4 2.1 0.3 26.4 -6.1 1.6 0.3 0.5 76 December 47.8 4.9 2.4 0.3 29.6 -6.7 2.0 0.3 0.4 56 Total - - - 3.2 - - - 2.3 - 945 a. At the free market exchange rate. b. Weighted average rate, with weights given by the amount of each type of deposit. Computed using the general wholesale price index. Sourre: Central Bank of Argentina and authors' calculations. prior to elections in September 1987. The growth in The monetary base was also expanding as a result money supply was directly related to public financing of government redemption of its outstanding of external debt service, rediscount facilities, and securities. As the internal public debt of the treasury amortization of the domestic public debt. Monetary increased, economic agents demanded higher real expansion was absorbed largely through reserve interest rates (up to 40 percent a year) and shorter debt requirements and forced investments, but because maturities. When the securities fell due, some were these controls were imposed on the financial sector redeemed with funds borrowed by the treasury from there were quantity and price impacts on borrowed the central bank, and some were refinanced with new funds. Funds to support productive activity dimin- bonds. The bond debt increased in real terms as the ished, and interest rates rose. In addition, a funda- government borrowed to service its existing debt in a mental redistribution of income between creditors grand Ponzi scheme. and debtors was under way. The authorities tried to absorb the immediate Whenever the Ministry of Economics lacked the impact of the expanding monetary base through funds to service its foreign debt, for example, the cen- reserve requirements and forced investments imposed tral bank expanded the monetary base to buy foreign on the financial system. This process repressed the exchange from the export sector. At the same time, it banking sector and redistributed funds from the pri- increased its indebtedness to the financial system in vate to the public banks. As a result, between order to absorb part of the additional liquidity it was December 1985 and December 1988 the central creating. Between the third quarter of 1985 and the bank debt to the financial system in terms of remu- end of 1988 interest payments to service overseas nerated reserve requirements increased by 59 percent debt totaled $13.5 billion. in real terms. By the end of 1988 nearly two-thirds Similarly, the central bank financed domestic real of total deposits were held in the form of reserve sector operations by using rediscounts (table 11.5). requirements at the central bank, compared with From 1985 to 1986-87 the real outstanding value of only one-third at the end of 1983. This led to fur- these instruments increased 80 percent, resulting in ther deterioration in bank portfolios, which was an increase in the money supply. When interest rates were freed in October 1987, the indebtedness of the Table 11.5 Balances of total rediscount line central bank to the banks rose along with interest (millions of australes) rates because it continued to pay interest from the regulation account even though the account accrued no income. Money creation could be sterilized only 1985 4.6 4.6 1986 13.1 7.8 by issuing more interest-bearing debt to commercial 1987 38.4 8.2 banks-debt that would have to be serviced and ulti- 1988 214.4 6.7 mately redeemed. Sourre: Central Bank of Argentina. 168 BANK RESTRUCTURING compounded by the unregulated activities of banks the volume of funds channeled through the formal in the interfirm market. institutions. About $1 billion equivalent-15 per- cent of the interest-bearing deposits in local currency Interfirm market at the time-were brought from the interfirm mar- ket. The main disadvantage of the new system was The noninstitutional, interfirm market developed its inflexibility in terms of deposit withdrawals, with a highly interconnected infrastructure. which proved a serious problem in the first half of Operations, which consisted mainly of the absorption 1989 when hyperinflation led to a generalized run and placement of short-term local currency funds, on deposits. typically were collateralized with government securi- ties. Market participants avoided reserve requirements Structural change and regulations on related lending, portfolio concen- tration, and minimum capital. This less-rigid frame- Between 1982 and the end of 1988 banks decreased work gave intermediaries a higher credit capacity per their lending to the private sector from 362 billion unit of borrowed funds, allowing them to charge australes (at constant December 1988 values) to 118 lower lending rates and to pay investors higher deposit billion australes. Funds instead were increasingly rates-and making them more profitable than their going to the central bank in the form of reserve formal sector counterparts. For borrowers, the inter- requirements. By the end of 1988 the central bank firm market represented easy access to credit uncon- had a quarter of the total assets of the consolidated strained by central bank regulations. financial system, essentially in the form of these Formal and informal channels of credit were still reserve requirements. These assets, however, were linked, however. Some of the financial institutions effectively frozen. The stock of total liabilities of the that participated in the interfirm market dedicated financial system held by the central bank-much of themselves to it entirely, allocating all managers, which had been channeled to the public sector staff, branches, and so on. The high level of risk in through public banks-also rose sharply. This meant interfirm operations led to solvency crises in some that a large portion of liabilities was tied to local, institutions when massive withdrawals of interfirm municipal, and state public finance, raising the deposits could not be met by the resources from potential for substantial losses. counterpart loans. In these circumstances interme- As a consequence of the higher real cost of the diaries would use the funds from the institutional scarce credit available to the private sector, banks' market, book imaginary loans, and pay off the loan portfolios again started to deteriorate. This time, interfirm deposits. In this manner the solvency however, it was not the health of the aggregate finan- problems of the informal system spilled over to the cial system that was at stake, but the viability of spe- formal system, at which point the central bank's cific, poorly managed financial intermediaries- deposit insurance system rescued depositors. Thus among them, public institutions. Portfolios began to the monetary base expanded. leading to additional deteriorate during the six months following the intro- reserve requirements. duction of the Austral Plan. In May 1985, 23 percent A new system discouraging interfirm market of the total portfolio was nonperforming, but by operations was adopted in October 1986 to break December that figure had risen to 30 percent, largely this vicious circle. Marginal reserve requirements because real interest rates were averaging 5 percent a were sharply lowered and were no longer remunerat- month. By the end of 1988 the nonperforming part ed. In addition, the stock of remunerated reserve of the aggregate portfolio had fallen slightly to 27 requirements held at the central bank was immobi- percent, but the position of government-owned lized and consolidated into a deposit account that banks had worsened (table 11.6). And rather than yielded an interest rate equivalent to the market make provisions against these exposures, the banks average. As a result the average reserve requirement accrued the outstanding interest, even where part of was significantly higher than the marginal one. In the principal was known to be lost. Realized losses the context of demonetization, the immobilized were covered by the national government through deposits translated into an effective increase in capitalizations, financed through monetary emissions reserve requirements. This system of high average by the central bank. Such techniques merely and lower marginal reserve requirements increased obscured the extent of bank losses. Argentinas Financial Crises and Restructuring in the 1980s 169 Table 11.6 Nonperforming assets in the banking system, December 1982-December 1988 (percent) Type of bank December 1982 May 1985 December 1985 July 1988 September 1988 December 1988 Government-owned 10.8 33.0 39.4 29.6 35.0 36.7 Private 10.9 9.5 11.4 8.3 8.6 7.9 Total 10.9 23.2 30.3 21.4 25.7 27.1 Source: Central Bank of Argentina. Banking operations become a net creditor to the public sector. Public lia- bilities totaled $7.5 billion-$5.8 billion in reserve Private and government-owned banks differed sub- requirements held in the central bank, $1.7 billion in stantially in terms of operating efficiency. The gov- national treasury bonds. These liabilities severely ernment-owned banks concentrated on financial restrained economic policy because they were short intermediation. They absorbed 41 percent of the term and yielded high positive real interest rates deposits of the system and granted 63 percent of (averaging 30 percent a year). In addition, there was loans, but by the end of 1988 they also were receiv- little propensity in the economy to hold narrow or ing 76 percent of the total rediscount lines of the broad money. The central bank could thus service its central bank. Their easy access to credit made gov- debt only by creating money that no one wanted to ernment-owned banks less diversified than their pri- hold or by creating interest-bearing debt that could vate counterparts, which went into new lines of busi- only be sold at exorbitant interest rates. ness such as capital market operations, corporate The authorities had hoped to lift restraints on the finance, and investment banking. banks by reducing reserve and inaccessible deposit In addition, private banks had an average of requirements. Instead, as inflation rose, they had to twenty-three employees per branch, compared with increase these burdens. An inflow of foreign forty in public banks. Private banks also invested in exchange, caused by a favorable trade balance in grain computers, which allowed them to achieve far greater exports to the United States, added to inflationary operating efficiencies than public banks. Moreover, pressures. Rising inflation put further pressure on the they participated in the formation of computer ser- free market exchange rate, widening the difference vice companies, which provided fee-based services to with the official rate. smaller financial institutions. Public banks eventually lost the advantage of The Primavera Plan implicit deposit insurance when some public provin- cial banks were forced to close. And emerging private It was under these circumstances that the authorities mega-banks fostered a "too big to fail" impression introduced the anti-inflationary Primavera Plan in among depositors-which gave the banks the protec- July 1988. The essence of the plan was a devalua- tion of implicit deposit insurance, eroding the privi- tion, followed by a multiple exchange rate system leged position of government-owned banks. designed to fill the central bank's coffers. The plan was supported by external aid, appropriation of a The 1989 Crisis portion of the surplus generated by the higher value of exports, maintenance of high interest rates to By December 1988, 206 financial institutions were make it more attractive to invest domestically than under central bank liquidation. Of these, 84 institu- externally, and revaluation of the real exchange rate tions had been transferred to the central bank before through a system of selling foreign exchange at the December 1982. The debts of these 206 institutions, parallel market rate and purchasing it at an official in terms of loans granted by the central bank for rate, which was expected to earn revenue for the cen- repayment of deposits, reached $7.8 billion. In real tral bank. Monetary policy would be aimed at what- terms this was nineteen times the debt balance in ever interest rates were necessary to maintain a free December 1982. The central bank made provisions market exchange rate spread of 20 to 25 percent covering 93 percent of the value of these credits. above the official rate. There were also wage agree- The situation in late 1988 was vastly different ments, substantial price increases, and further mea- from that in 1982 because the private sector had sures to decrease fiscal imbalances. These measures 170 BANK RESTRUCTURING included cuts in public employment and investment gross national product (GNP). The M6 aggregate, and a 50 percent reduction in the financial assistance including all types of deposits and government secu- offered by the central bank to Banco Hipotecario, rities, fell from $13.2 billion equivalent in September the national housing bank and a major recipient of 1988 to $3.9 billion (valued at the free exchange rate) rediscounts. In addition, measures were taken to lib- in June 1989. Private credit was unavailable and, eralize imports. therefore, its cost was infinite. In June 1989 tax col- Although this program was similar to earlier lections were only half those of a year earlier in real efforts, the economic team behind the Primavera terms. GDP fell an estimated 13 percent in the sec- Plan had far less public credibility, partly because ond quarter of the year, and the purchasing power of interest rates had risen to very high real levels. T here salaries decreased 30 percent between March and was also doubt about the viability of wage agreements June. Consumer prices increased 1,877 percent and the response to public offerings of foreign between February and July, while the free exchange exchange, since demand might outstrip supply. rate increased 3,723 percent. By the end of June Nevertheless, during the first two months of the plan international reserves were not sufficient to cover inflation dropped into single digits and the foreign short-term debts. exchange spread between official and market rates Hyperinflation substantially altered the opera- was maintained at 20 to 23 percent. The program's tions of the financial system. Commercial banks suf- continued success depended entirely on exchange rate fered heavy deposit withdrawals as investors rushed expectations. There was no reduction in the fiscal to dollar-denominated assets. During the first half of deficit. 1989 there were eight banking holidays and twelve The decrease in the exchange rate spread was exchange holidays, and limits were frequently set on achieved by means of a real interest rate that was the withdrawal of sight and time deposits. The vol- highly positive in terms of dollars. During the plan ume of deposits decreased in real terms and there was the accumulated deposit interest rate reached 78 per- an increase in deposit withdrawals. To avoid iestric- cent, while the exchange rate devaluation of the free tions on withdrawals, people opened multiple dollar was only 24 percent. This stability was fragile accounts, boosting the ratio of bank accounts to bank from a macroeconomic viewpoint, given that the personnel 10 percent between December 1988 and interest rates were supported by the central bank July 1989. As bank illiquidity worsened, the central (through the remunerated reserve requirements on bank released reserves and inaccessible deposits, pro- interest-bearing deposits). Private sector demand for vided rediscounts, and ultimately permitted wide- credit was practically nil. The exchange rate spread spread reserve shortfalls and informal overdrafts by ultimately was maintained through the quasi-fiscal commercial banks. In April the central bank raised deficit of the central bank. the interest it paid on reserves to the commercial The Primavera Plan ended when expectations banks. Further changes in the exchange rate failed to about the exchange rate changed in January 1989 staunch the hyperinflationary spiral because of the and the public increased its demand for dollars increase in base money the central bank was provid- offered for sale by the central bank. Faced with the ing to banks to forestall closure. possibility of losing a significant amount of interna- The financial system's consolidated balance sheet tional reserves, the central bank stopped selling for- continued to reflect the real decrease in the volume of eign exchange, forcing those who had foreign debts financial intermediation and the increasing indebted- to finance them through exchange purchases in the ness of the public sector (including the central bank) free market. The exchange rate spread widened, dis- to the banks. Despite this, however, banks' net worth abling the only brake on the increase in internal rose because of the indexation of their fixed assets- prices. In February the government introduced the central bank's accounting norms require revalua- another devaluation and hyperinflation began. tion of fixed assets in line with inflation-and the profits accrued by the banks during this period. Even Crisis in thefinancial system in the hyperinflationary environment of the first half of 1989, in fact, the accrued profitability of the The economic crisis was evident from a number of financial system was positive, and higher than during economic indicators. Demonetization reached a his- the previous July when the inflation rate had been toric peak-in June 1989 Ml was just 1.6 percent of stable, albeit high. This performance was attributable Argentinas Financia1 Criie andRestructuing in the 1980s 171 to two factors: the distinctions made among deposit had again reached 15 percent a month. The authori- interest rates according to deposit size (smaller ties admitted defeat as they devalued the official rate deposits received a lower interest rate) and an increase from 655 australes to 1,010 australes to the dollar. in central bank remuneration on forced investments. The team that had designed and implemented the BB Still, the financial system was hardly viable because Plan resigned, and a new economic team took office. the government-which was effectively bankrupt- At this point complete dollarization of the econo- once again started to use it as a financier. In July 1989 my seemed plausible. To avert this possibility the the Menem administration implemented the BB Plan, government announced dramatic measures on whose sharp devaluation (more than 110 percent) was January 1, 1990 to halt the hyperinflation that had intended to improve the fiscal account by encouraging plagued the economy during the 1980s. The Bonex exporters to surrender foreign exchange.3 There were Plan forced the conversion of commercial banks' time massive increases in public prices and price freeze deposits into cash (up ro $500 equivalent) and exter- agreements with major private enterprises. These nal bonds (Bonex) with a maturity of ten years, a moves were intended to restore public confidence so two-year grace period, and an interest rate slightly that the government could borrow from the financial above LIBOR. Treasury and central bank obligations system at reduced interest rates as part of its structural to the banks also were converted, thereby eliminating adjustment program. the impact of this large domestic debt on public sec- But until the primary surplus rose sufficiently to tor borrowing requirements. The Bonex Plan also cover interest on the domestic debt, stability depend- included a unilateral writedown of public debt to the ed on monetary policy. The central bank decided not private sector since Bonex bonds would trade at a dis- to resort to inaccessible deposits, but to sell its own count that would widen as the outstanding debt bills ("certificates of participation") to financial insti- stock increased. The effort to restore central bank sol- tutions in open markets in order to control liquidity vency by eliminating the bank's quasi-fiscal deficit (Beckerman 1992). The initial (monthly) 5 percent meant unprofitability and widespread insolvency for rate offered on the certificates was too low to deter commercial banks. Moreover, it further eroded confi- time-deposit withdrawals, which put additional pres- dence in the Argentine financial system and killed sure on the exchange rate. So the central bank raised demand for local currency or public securities. At the the rate to 15 percent. Because of the volume of out- time, however, there were few other options, and standing public debt, however, the cost of borrowing implicitly declaring public bankruptcy may well have at this rate was more than the central bank could been the best of a bad lot. support. The rates dropped to 4 percent between July and October, however, as commercial banks volun- Conclusion of the Crisis tarily absorbed these new instruments-about $3 bil- lion worth-without opening the spread between the The 1989 crisis was the final unwinding of an unsta- official and parallel exchange rare. ble financial situation that began in the late 1970s But the BB Plan went the way of the Primavera with financial sector liberalization. The basic cause of Plan as the private sector defected. The interest rate the crisis in the early 1980s was the rapid expansion on the certificates was too low to maintain the target of private financial institutions at a time of liberalized exchange rate, so the central bank again raised the interest rates and extensive external borrowing, cou- rate to deter deposit withdrawals from commercial pled with large fiscal deficits and overvalued exchange banks. The government was caught in its own debt rates. But the cutoff of international resources during trap: it raised rates on certificates of participation to the 1982 debt crisis raised domestic real interest rates competitive levels to staunch further flight to the dol- and forced successive devaluations that fueled infla- lar, but it could not service its debt at competitive tion and increased disintermediation. The solution to rates without borrowing more to do so. Depositors the private sector debt distress of the early 1980s- recognized the futility of this effort, and commercial assumption by the central bank of private external banks were thrust into panic because liquidation of debt-placed the bank in a perilous net foreign lia- certificates of participation to meet deposit with- bility position, thus reducing its capacity to manage drawals would only depress the value of their assets. monetary policy. Emergency measures adopted in November 1989 During the second half of the 1980s, as public accomplished nothing. By December interest rates deficits again ballooned, nonfinancial public borrowing 172 BANK RESTRUCTURING requirements reached unsustainable levels, climbing Table 11.7 Financial system assets and liabilities, from 2.5 percent of GDP to 15.9 percent. The cen- December 1981-June 1990 tral bank tried to stem disintermediation from the December June June banking system by paying higher and higher real Item 1981 1989 1990 interest rates, which generated large quasi-fiscal losses Public sector credit 40.3 292.3 99.3 on its books. By 1989 the central bank's quasi-fiscal Private sector credit 119.9 91.3 43.6 balance was estimated at 6 percent of GDP, with total Private sector liabilities 100.0 84.3 25.0 central bank interest-bearing liabilities reaching 37 N Note: Private sector liabilities in December 1981 are used as the base. percent of GDP in the first quarter of 1989. Sourre: Central Bank of Argentina The 1989 financial crisis therefore was funda- mentally a case of public sector debt distress. The pri- liabilities to the private sector as a base. Credit to the vate sector was not a major player in the crisis public sector rose by more than seven times, while because it could not compete for resources at the private claims on the financial sector declined 15.7 high real interest rates paid by the public sector. The percent in real terms between December 1981 and public sector losses had to be passed on to other sec- June 1989. After inflation and debt conversion public tors. It was no longer possible to pass the losses to the real debt dropped by two-thirds. Private debtors external sector because external debt was already in gained by more than half, while private depositors arrears. Repudiation of the external debt was ruled and currencyholders lost 70 percent of the real value out. Attempts to generate a primary surplus sufficient of their claims on the financial sector. to pay for the combined public (including central In effect, the Argentine economy had to deal with bank) interest burden also failed. The only alternative systemic debt distress. It had to stop financing public was to repudiate internal debt. deficits by borrowing from the financial sector- The January 1990 conversion of austral-denomi- which was the primary cause of inflation-and deal nated time deposits and public sector debt into $3.5 with the suspension of debt with external creditors. billion in dollar-denominated Bonex (at an account- The Argentine case was the most severe instance of ing exchange rate of 1,800 australes per dollar) financial distress (other than that of Yugoslavia), and reduced central bank debt and-since interest rates on its solution the most extreme. australes were considerably higher than rates on Bonex securities-debt servicing interest rates. To the extent Lessons that commercial banks reduced their time deposits and remunerated reserves with the central bank, the The Argentine crisis in the early 1980s demonstrates central bank eliminated its interest-bearing liabilities the dangers of financial liberalization without ade- and hence its capacity to generate further quasi-fiscal quate attention to supervision in an environment deficits. The conversion was a stock adjustment that lacking fiscal discipline. An overvalued exchange rate, passed the losses on to depositholders. high domestic interest rates, and external borrowing Since Bonex traded at a discount of almost two- added to the private enterprise distress that resulted thirds, the conversion created large losses for holders of in deterioration of the portfolio of the financial sys- time deposits. The immediate impact was a reduction tem. This pattern was common to a number of other in M3 of roughly 8.5 percent of GDP The debt con- countries in the early 1980s, notably Chile and version did not stop inflation, however. Partly because Colombia. of policy uncertainties, deposits for tax and wage pay- The assumption of private sector distress by the ments, and exemptions for senior citizens and on public sector, without first restoring fiscal discipline, deposit conversions for withdrawals of up to 500,000 led to a vicious circle of deteriorating public finance australes, more than half the deposits at the end of that gradually spiraled into public sector debt dis- 1989 were still available to depositors. After another tress. Even the rescuer of the financial system, the short burst of inflation in March 1990, when inflation central bank, was caught in the debt trap-and thus reached 96 percent a month, inflation began to sub- became part of the problem. side as the fiscal position and the balance of payments The Argentine case also demonstrates the need for showed improvement in the first half of 1990. careful monitoring of all quasi-fiscal deficits, whether The losses borne by the private sector during the in public enterprises, state banks, or the central bank. 1980s are shown in table 11.7, using financial sector Failing to address these deficits ultimately escalates the Argentinas Financial Crises and Restruring in the 1980s 173 costs of rescuing the financial system through an percent of the first I million pesos, and 90 percent of deposits implicit or explicit deposit insurance scheme. If the above that. Foreign currency deposits were not covered. state's capacity to finance the total fiscal deficit is lim- 2. The previous system had provided an incentive to disin- ited, and the primary surplus is insufficient to,pay the termediation because the yield on remunerated reserves was leas ited and the primary surplus is insufficient to pay the than could be earned on alternative lending operations. interest burden of the public sector, then public sector Financial groups thus allowed their commercial banks' deposits debt distress is a mathematical inevitability. to decrease and invested their funds in the interfirm financial marker (Beckerman 1992). Postscript 3. The plan was called "BB" because President Menem drew his economic team from the executive ranks of the Argentine corporation Bunge y Born (Beckerman 1992). Although the January 1990 debt conversion did not solve the crisis, it created the political and economic conditions that allowed the government to put in place a package of policy changes that enabled recov- Arnaudo, Aldo, and Rafael Conejero. 1985. "Anatomia de [as ery and improved general stability. During 1990-91 Quiebras Bancarias de 1980." Desarrollo Economic6 24 the government initiated a major effort in fiscal (January/March): 605-16. reform by widening and making uniform the value- Balifio, Tomas. 1987. "The Argentine Banking Crisis of 1980." added tax and improving tax collection. In addition, IMF Working Paper. International Monetary Fund, the bureaucracy was reduced and privatization was Central Banking Department, Washington, D.C. accelerated. Through debBeckerman, Paul. 1992. 'Public Sector Debt Distress in accelerated. Through debt negotiations, the external Argentina, 1988-89." Policy Research Working Paper 902. stock of debt was reduced by $7 billion, partly World Bank, Washington, D.C. through debt-equity swaps in the privatization Cavallo, Domingo, and Angel Pefia. 1983. "Deficit Fiscal process. Major efforts were also made in trade and Endeudamiento del Gobierno y Tasa de Inflaci6n: tariff reform and deregulation. In the financial sector Argentina 1940-1982." Estudios26 (AprilJune): 39-78. efforts were made to restructure and downsize the Fernandez, R., and C. A. Rodriquez. 1982. 'lnfl2Ci6n y Estabilidad." Ediciones Macchi. major public banks, including significant staff reduc- ______. 1983. "Las Crisis Financieras Argentina: 1980-1982." tions. The April 1991 Law of Convertibility guaran- Jornadas de Economia Monetaria y Sector Externo, Banco teed convertibility of the peso at a rate of one peso to Central de la Republica Argentina, Buenos Aires. one dollar, thus making the central bank a de facto Gaba, Ernesto 1981. "La Reforma Financiera Argentina: Lacciones currency board. de una Experiencia." Ensayas Economicos 19 (September): 1-52. Piekarz, Julio. 1987. 'El Deficit Cuasi Fiscal del Banco Central.' Banco Central de la Republica Argentina, Buenos Aires. Notes Rivas, E. 1984. 'Costos Bancarios, Produccion Multiple y Rendimienros a Escala." Serie de Esrudios Tecnicos 61. 1. In 1979 the deposit insurance scheme was changed to require CEMYB, Banco Central de la Republica Argentina, Buenos monthly contributions from member banks, assessed at a fixed per- Aires. centage rate on their average liabilities subject to reserve require- Rodriquez, C. A. 1991. "The Macroeconomics of the Public ments. Banks were not compelled to join the scheme, which in any Sector Deficit: The Case of Argentina." Policy Research event was not fully funded. Depositors were protected up to 100 Working Paper 632. World Bank, Washington, D.C. I CHAPTER 12 Post-Liberalization Bank Restructuring Andrew Sheng Financial markets are derivatives of the real economy. had systemic impacts on other highly indebted coun- Restructuring banks in the post-liberalization world tries. In January 1995 the knock-on effects of capital ignore this at their peril. withdrawal from emerging markets caused a ripple of speculation-successfully beaten back-against a Most developing countries have emerged from the number of East and Southeast Asian currencies. In devastation of the 1980s, a decade of high debt and February the U.K.'s Barings Bank was brought down low growth. In the past several years increasing num- by losses of more than $1 billion through speculation bers of developing countries have introduced market- of more than $26 billion in the Nikkei stock index oriented reforms, producing great strides in trade and Japanese government bond futures. The turmoil reform, financial liberalization, and the development in the currency markets continued in March with a of capital markets. Still, the transition to market sharp slide in the U.S. dollar and devaluation of the remains a major challenge for most of the post-cen- Spanish peseta and Portuguese escudo. Banking trally planned economies. problems resurfaced in Mexico, while French banks The banking problems of the 1980s were almost began to show large losses from the managerial forgotten by 1993, when banks earned record profits. autonomy given to the state-owned banks over the Low interest rates and renewed faith in the global past five years. recovery fed bubbles in both bond markets and As funds fled in search of security, banks in emerging-market share prices. Privatization, financial emerging markets again witnessed liquidity and innovation, and investment by industrial country potential solvency problems. In defending their cur- pension and mutual funds led to enormous capital rencies against speculative attacks, policymakers dis- flows to the developing countries, especially in Asia covered that their vulnerable spots were not fiscal and Latin America. Around the world, bank profits deficits, but potential quasi-fiscal deficits in their rose as margins increased, lending recovered, and banking systems, whose fragility was exposed in the proprietary trading offered new sources of income- newly liberalized, global financial environment. especially in foreign exchange, money market instru- Thus-despite considerable progress in the ments, and derivatives. 1990s-banking fragility has not gone away, and the The momentum began to slow in 1994 as interest debates over bank restructuring remain. Should gov- rates in the United States started rising. Problems ernments devote considerable resources to bailing out with derivatives surfaced with well-publicized losses, failed banks? Why can't the market take care of itself? most notably those of Metallgesellshaft, Proctor & Should state resources be used to bail out failing Gamble, and Orange County. The Mexican currency enterprises, the sources of the bad loans for the bank- devaluation in December revived the specter of sover- ing sector? What is the correct sequencing of inter- eign risk, forcing the U.S. government and multilat- ventions? Has financial liberalization impaired the eral agencies to piece together a $50 billion rescue ability of bank supervisors to manage the financial package to stem a potential default that would have institutions of the 1990s? Are the lessons of the 175 176 BANK RESTRucTURING 1980s relevant to bank distress in the 1990s? This the (highly regulated) banks. The rapid emergence of chapter reviews the lessons of bank restructuring cov- such extensive holdings in newly open financial sys- ered in the introduction and considers alternative tems complicates the supervision and regulation of approaches to the issues developed there. emerging markets and their relationship with the banking system. Bank Restructuring in the Post-Liberalization World Are the principles of oversight of capital markets different from those of the banking system? Banking Bank restructuring should not be confused with bank systems and capital markets are both complementary crises. Crises are events; restructuring is a process. and competitive channels of resource allocation in a The variety and origins of crises change, and the market economy. Developing country policymakers techniques of reform and resolution must adjust have always concentrated on supervising the banking accordingly. But the basic processes of bank restruc- sector, because that has been the primary channel of turing identified in this volume-diagnosis, damage resource allocation, savings, and credit-which control, loss allocation, and rebuilding-remain as explains many governments' desire to own and strict- valid in the 1990s as they were in the 1980s. ly regulate banks. The emergence of capital markets, both equity markets and money and debt markets, The new world has changed that landscape and challenged these con- ventional perspectives. With the emergence of financial globalization and Domestic and global markets today thrive on the innovation, aided by improved telecommunications arbitrage of information, skills, taxes, regulations, and and technology, developing country financial mar- even differential transaction costs. There are a num- kets have changed rapidly to keep up with U.S. and ber of reasons for the phenomenal growth of capital European bankers-with what Kaufman (1994) markets-and the corresponding move away from refers to as the "Americanization of financial mar- more traditional banking-and most relate to kets." These changes occurred partly under the wave increasing levels of regulatory and tax arbitrage: of deregulation that swept the globe in the 1980s, Developing country governments have imposed led by London's 1986 "Big Bang" financial liberaliza- extensive regulations on the degree of financial tion, and partly in response to the need to synchro- innovation in the banking system, partly for fear nize financial market reforms with real market of exposing it to higher risks, but also for fear of changes. Even the most conservative financial losing control over the monetary base. authorities in developing economies realized that * In almost all countries with rapid capital market there were limits to financial repression of traditional growth, there has been no capital gains tax on banking markets as a means of financing growth and securities transactions, whereas interest on bank industrialization. deposits is taxable, particularly for corporations. The progressive dismantling of exchange controls * Because of securitization and financial innova- and the increasing inflow of portfolio and foreign tion, the cost of intermediation through capital direct investment pose new challenges for developing markets is significantly lower than through bank country central bankers used to controlling their credit. U.S. experience suggests that spreads on markets. The most marked change has been the funds obtained through the debt market are increasing strength of nonbanks relative to banks, only 50 basis points, compared with nearly 200 powered mostly by capital markets. Short-sighted basis points for bank loans. Indeed, the remark- banking specialists often fail to recognize the chal- able growth of derivative markets can be attrib- lenge posed by these markets. For example, in five uted to the ability to hedge risks at very low Asian economies-Hong Kong, Malaysia, the transaction costs. Philippines, Singapore, and Thailand-market capi- * As more developing countries discover the bene- talization in the equity market is already larger than fits of privatizing through capital markets, bank- the total domestic currency assets of the banking sys- ing risks actually drop because the overall gear- tem. A Malaysian stockbrokering firm can earn more ing of borrowers declines with higher capital profits in one year than many of the five largest bases raised through equity markets. And banks. Nonbank financial intermediaries in the national risk management improves with effi- Republic of Korea have more deposits and assets than cient capital markets. Post-Liberalization Bank R&structuring 177 Resource allocation is more efficient through cap- The speed and complexity of markets makes the ital markets because capital markets enable enter- post-liberalization world dazzling to bankers and prises to swap assets and liabilities through equity policymakers. Markets have become more liquid and debt, rather than through bank financing. and volatile as a result of liberalization and innova- But while private sector banking systems and tion. Technology has made banking services faster, securities markets have moved fast, the regulatory cheaper, and more convenient. In Hong Kong, for and supervisory systems in many developing coun- example, funds can be transferred into twelve differ- tries have not caught up in terms of resources and ent foreign currency accounts by telephone. understanding, particularly regarding the complexity Deregulation-particularly the lifting of capital and volatility of markets. As a result many policy- controls-has removed barriers to inflows and out- makers have found themselves with less freedom and flows. Portfolio choices are no longer restricted to fewer instruments to deal with the crises arising from domestic currency assets. Financial innovation, par- market flows. ticularly in derivatives, has created low-cost services and products that can be tailored to an investor's Derivatives and systemic risks needs. Banks and securities houses thrive on selling such products. A significant development in the 1990s (led by the Investors, for their part, can now quickly shift United States and the European Union) has been the funds into a variety of products to protect their total merging of banking and financial markets. Such con- return. The result is that any perception that solidation of different product markets into large investors will suffer a deterioration in total return, financial groups is meant to improve economies of for whatever reason, may trigger large portfolio scale in product delivery and reduce risks through shifts. Because news travels fast-regardless of diversification. Steps have been taken in the United whether it is true-investors can protect their States to remove the Glass-Steagall restrictions on investments using a variety of hedging instruments, bank participation in securities markets. In the particularly forwards, futures, and options, so long European Union insurance companies have begun to as a counterparty is willing to bet the other way. merge with banks, banks have taken over securities Programmed trading would allow computer-gener- houses, and building societies have upgraded them- ated trading to take advantage of minor variations selves to banks. In Malaysia and Mexico emerging in prices through "anomaly trading" or to activate securities houses have generated so much capital that selling programs once prices touch certain trigger they have absorbed banks within their groups. With levels. With global markets linked by Reuters, the rapid growth of emerging markets and derivative Telerate, and Bloomberg news services, and trading instruments, money center banks in the United on a virtual twenty-four-hour basis, any bad news States have transformed themselves into investment can generate a tidal wave of fund flows. houses, while securities houses have become con- So can the market take care of itself, since there glomerate investment banks with subsidiary banks, will always be another investor willing to bet the credit card companies, derivatives traders, and asset other way? Unfortunately, the answer is no since the managers. The eight money center banks now world still operates on imperfect information. When account for more than 85 percent of the total volume bad news prompts capital flight, the only counterpar- of derivatives activities by U.S. banks (Moody ty willing to absorb the other side of the hedge may Investors Service 1994). be the authorities, seeking to establish stability in the The melding of financial institutions makes the markets. If the authorities are unwilling or unable to separation of banking from financial markets increas- intervene, then prices may freefall, creating a major ingly difficult. As banks engage in securities business price shift that could devastate the economy. Thus and securities businesses engage in banking, cross- there is sympathy for a Mexican manufacturer who border transactions increase in size and volatility. was profitable and efficient a year ago, but who today Bank supervisors now find themselves needing to talk faces considerable difficulties because currency depre- not only to domestic financial market regulators- ciation has pushed import prices up 80 percent. Of such as the insurance commissioner or securities course, there is always the response that free markets exchange commission-but also to the regulators' imply that the manufacturer should have anticipated foreign counterparts. the price adjustment. 178 BANK RESTRucruRING One policy response to market volatility is to same amount of capital as providers of loans. "throw some sand into the wheels," that is, make Together with the low transactions costs associated fund flows less volatile by reimposing or threatening with derivatives, this accounts for the rapid growth to reimpose capital controls. Variations of this in gross outstanding derivatives--reaching as much approach have been used by Chile and by Malaysia as $7 trillion by the end of 1991, as much as the in early 1994 when massive capital inflows threat- total cross-border assets of Bank of International ened to negate domestic monetary policy objectives. Settlements reporting banks. Another policy response for those who fear the Some analysts argue that regulators should not be "dark side" of high-speed, high-power financial mar- concerned with the gross values of derivatives, but kets is to ban or impose restrictions on such instru- only with the net replacement cost of derivatives ments as derivatives. The Barings Bank disaster and trading. Net replacement costs could be quite low other well-publicized losses by corporations that when the volatility of markets is low. But in unusual experimented with derivatives have created an aver- markets, such as the wild swings in currency markets sion to these exotic products. But derivatives, like all during the exchange rate mechanism crisis of the financial tools, have their advantages and disadvan- European monetary system in 1992, underlying mar- tages, depending on how they are used. The Group kets can dry up and bid-ask spreads can widen con- of Thirty and the Bank for International Settlements siderably, resulting in many derivatives models not agree that there should be greater transparency in the being able to price their derivatives correctly and disclosure of derivatives products and transactions, placing their buyers and sellers in a high-risk environ- that issuers and users should have adequate risk man- ment. The failure of Barings Bank indicates that agement systems, and that regulators should closely bankers must be concerned not only with their net examine the legal issues of netting and collateral and exposure, but also with their gross exposure. The cur- the greater use of clearinghouse arrangements for rent lack of transparency in derivatives trading means over-the-counter products. that many regulators are not aware of the true size of The derivatives market evolved in response to the risks being assumed by their banking systems. the need to hedge and to manage risks better. This blind spot, caused by financial innovation and Derivatives can improve the liquidity of markets liberalization, may be the biggest problem for regula- and reduce transaction costs in portfolio manage- tors in the post-liberalization era. ment. But because of the leverage factor, users of derivatives can suffer large losses from market Why do banksfail ifthe market can take care ofitse#f' volatility and defaults of counterparties. Advocates of the derivative market often forget that while indi- Banks fail because of bad policies, poor banking prac- vidual institutions can reduce risks through deriva- tices, and weak institutional frameworks. In a global- tives, the system as a whole is a zero-sum game. ized financial market the management of banks and Systemic risk is not reduced simply because one part financial systems must be placed in not only a sectoral, of the financial system is active in derivatives hedg- but also a national and international context. The ing activity-the risks spread to other parts of the Barings failure involved a British bank speculating on financial or nonfinancial system. Indeed, reducing Japanese index futures in the Singapore International risks in one part of the economy may even increase Monetary Exchange. Since Barings was a major inter- risks in other parts. national asset manager, the failure affected investors Advocates also tend to forget that bank loans are and depositors all over the world; its rescue involved a derivatives of corporate earnings. Banking has been Dutch banking conglomerate. Mexico's problems highly regulated because banking is a highly lever- threatened Latin American and emerging markets as aged business and consumer deposits need to be OECD investors sought to reduce their exposure to sheltered from losses. The dangers of the modern emerging markets as a whole. Policymakers therefore derivative business stem more from the fact that sell- must consider their exposure to international risks ers of derivatives are offering highly leveraged (low when designing structural adjustments to their margin) products and assuming large credit risks in economies and banking systems. the meantime. Since derivatives are mostly account- Current problems also have roots in the misguid- ed "below the line" and are off-balance sheet, the ed policies of the 1980s. This volume has attempted sellers of derivatives are not required to provide the to show how such policy mistakes, including poorly Post-Liberalization Bank Restctsuring 179 designed financial liberalization programs and inade- Lessons for Post-Liberalization Bank Restructuring quate supervision, led to banking crises. In the rush to liberalize as a correction to central planning and This section reviews the relevance of the lessons of excessive regulation, many policymakers and their the 1980s to bank restructuring in the 1990s. advisers lost themselves in the herd instinct, assuming that markets will always correct themselves. There Lesson I Financialstability rests on thegovernments was universal underestimation of the volatility of ability to maintain a stable currency markets in a global world. Market analysts used to say "the trend is your The speed of market reactions in the post-liberal- friend." For policymakers the trend can also be the ization world does not allow policymakers to reflect end. The herd instinct can cause market prices to on their policy mistakes. A stable financial system vary tremendously, mainly because of information rests on monetary stability, without which good asymmetry. Markets do not have the information or business decisions cannot be made. But in the post- the confidence in the economy or banking system liberalization world monetary stability needs to be that policymakers have. In addition, financial liberal- redefined. A simplistic view of monetary stability is ization programs have placed limits on the instru- that the government maintains fiscal discipline by ments, such as capital controls, that policymakers running a budgetary surplus. But since banks can have to deal with such swings. Central banks, even in create money through their credit operations (and the OECD countries, are not as capable as market by absorbing counterparty risks in derivative players of resisting market trends through interven- trades), there is also a need to maintain banking tion. The classic advice, that policymakers in a mar- discipline. Some observers have suggested that poli- ket economy must make sure that the fundamentals cy credibility can be maintained through a currency are sound, is more valid than ever. But in the real board system. But a currency board system cannot world most countries do not get all their fundamen- generate foreign exchange on its own, while a bank- tals right at the same time. ing system can generate domestic currency through The sharp swings of the 1990s suggest that poli- its credit operations. Consequently, so long as cymakers and bankers alike must pay attention to domestic banks are willing to absorb credit risks by risk management at the institutional, sectoral, and lending (either by acting as the counterparty to a national levels. Simple rules, such as never using derivative trade involving domestic currency or short-term resources to finance long-term invest- through direct loans in domestic currency), the for- ments, apply equally to banks and economies. The eign exchange reserve-domestic currency ratio can reliance on short-term portfolio flows to finance cur- always be diluted to erode the credibility of the rent account deficits exposed many countries to external value of the currency. financial instability when capital flows changed For example, speculation against a country's cur- direction. rency will always succeed if domestic banks are will- When national balance sheets were rewritten in ing to lend domestic currency to the speculators. the 1980s in response to financial liberalization and When foreign investors have better credit ratings other reforms, many policymakers underestimated than domestic clients, banks may be tempted to lend the risks posed by the global economy. Volatile capi- to fuel speculative attacks against their own domestic tal flows, made possible by the removal of capital currency. Such behavior implies that policymakers, controls, can reward good reforms, but they can also including bank regulators, must take bank supervi- punish policy mistakes rapidly and severely. sion issues into account in the management of their Policymakers have little time to correct such weak- monetary policy. nesses because once the market perceives that there There is an inherent conflict between monetary are policy errors, capital flight puts pressure on the policy and bank supervision policy. At a time when exchange rate, which in turn puts pressure on asset the central bank is concerned with maintaining a sta- prices such as stock and property prices and ultimate- ble currency, there may be a need to rescue banks, ly domestic inflation. Policymakers who seek to which creates a quasi-fiscal burden. Monetary cre- ensure overall price stability in order to foster growth ation through lending to rescue banks negates the suddenly find that speculation could destabilize all ability of tight monetary policy to combat inflation their reform efforts. or capital flight. 180 BANK RESTRUcrURING Removing exchange controls also removes a precipitating a currency crisis that eventually caused major tool of policymakers, leading to conflicting large losses in the real economy. In a number of coun- targets of maintaining either the internal value of tries bank losses were compounded by an overconcen- currency (domestic prices) or the external value tration of assets geographically, sectorally, or in terms (exchange rate). As the basic Mundell (1967) model of ownership, which encouraged connected lending shows, in a world of free capital flows the central and credit abuses. At both the microeconomic and bank must choose between stabilizing the exchange macroeconomic levels, bank managers and policy- rate or stabilizing the interest rate-it cannot do makers have not managed their risks very well. both. Central banks can still exploit certain ineffi- At the macroeconomic level, diversifying national ciencies in the market to influence both rates, but export income and attracting foreign investment help only in the short term. reduce risks, but certain economies and banking sys- As market information improves and investors tems have limited space in which to adjust. If an become more sophisticated, disintermediation from economy is landlocked or has inherited a highly con- the banking system occurs quickly once wealthhold- centrated industrial or agricultural base, its ability to ers foresee future asset losses because of bad policies, adjust is necessarily limited. Such rigidities are not in ineffective management, or weak institutional frame- themselves defects-it depends on whether policies works. Restoring financial discipline in the economy are devised to compensate for the economy's rigidi- begins with restoring fiscal and monetary discipline. ties. The ability to think about national risk manage- ment is important. Are there risks in the economy Lesson 2 Banks fail because of losses in the real sector, that policymakers should take account of? Countries compounded by poor risk management andftaud participating in the global market may have to main- tain significantly higher foreign exchange reserves Because money is a derivative of real value in an econ- given the volatility of capital flows. omy, tampering with the value of money risks tamper- ing with the entire financial structure. At the same Lesson 3 Liberalization programs often fail to take into time, the value of bank money is a derivative of the account the wealth effects of relative price changes, and performance of the real economy. Problems or losses inadequate supervision creates further losses in the real sector ultimately surface in the banking sys- tem. Thus bank restructuring cannot be achieved This is perhaps the major lesson of financial sector without reform or restructuring of the real sector. liberalization in the Southern Cone (Argentina, Real sector losses show up in the banks through Chile, Uruguay) economies in the 1980s, but some the failure of counterparties to honor their contracts, of the problems of the 1990s show that the lesson has such as credit losses. Thus bank restructuring requires not been fully learned. The overly rapid trade liberal- that bank regulators understand the ability of both ization of the 1980s created trade shocks for previ- banks and their clients to manage their risks. For ously protected enterprises in many countries, lead- example, the large capital inflows of the past several ing to large bank losses. Where these enterprises years led some policymakers to allow domestic enter- belonged to business groups that also owned banks- prises, as well as the government, to borrow short- and bank supervision was inadequate-these banks term flows to finance long-term investment. The bor- often financed distress borrowing at unrealistically rowers incurred excessive maturity mismatches, as well high real interest rates. as foreign exchange mismatches, since the borrowers In the 1990s the failure to adequately address may not have access to foreign exchange earnings. bank supervision resulted in weak banks that were Such poor risk management occurred because bor- unprepared to compete in global markets. Credit rowers had no experience with risks of this sort, never risks were not evaluated properly, and banks had high having operated in a liberalized environment. Taking levels of bad loans remaining on their books and advantage of their newfound freedoms, and fearful inadequate capital relative to such risks. If policy- that capital controls would be reimposed, enterprises makers raised interest rates to protect the domestic engaged in extensive external borrowing. When the currency from capital flight, banks simply raised prospect of devaluation developed, the fear of suffer- interest rates and transmitted the losses to enterprise ing devaluation losses on loans prompted both lenders borrowers. The failure of borrowers threatens to and borrowers to engage in hedging activities, thus become a systemic issue, infecting the banks and Post-Liberalization Bank Restructuring 181 requiring public rescue. A credible bank restructuring Lesson 6 Stopping theflow offuture losses is critical program requires strong bank supervision and enforcement, together with laws that encourage debt Damage control involves changing incentives within discipline and avoid bank owner-borrower conflicts. the real and financial sectors. Consider the example of the excessive leveraging of buyers of privatized Lesson 4 Bank losses ultimately become quasi-fiscal banks, who are then more concerned with paying off deficits their debts than with maximizing profits. There is a danger that such owners will engage in risk-taking Widespread bank failures simply cannot occur in behavior and the financing of capital flows in order most countries. Allowing banks to fail during a peri- to pay off their foreign currency obligations. Where od of capital outflows encourages domestic and for- incentives are distorted and cause losses, ownership eign capital flight. And most policymakers under- or management should change. Changing manage- stand that widespread bank failures suggest that the ment is essential at the bank level. Failure to address government is unable to maintain overall stability. the incentives structure invites a recurrence of bank- The result is that even with limited deposit insurance ing problems in the future. schemes in place, many policymakers have not dared to pass large bank losses on to depositors. Lesson 7 The method of loss allocation determines the This pattern implies that bank losses are ultimate- success of the restructuring program. ly the burden of the state. Thus the state has to ensure that it maintains fiscal and banking discipline. Someone must pay for bank losses. In a world of free Running a small primary surplus is not sufficient if capital flows this burden tends to fall on those who the budget is threatened by the large fiscal cost of are unable to escape it. Loss distribution is a political recapitalizing the banking system. A deposit insurance matter that must be addressed by individual societies. scheme is only a temporary means of calming deposi- Countries differ, calling for case-by-case restructur- tors during a crisis. What is required is to address the ing. Since losses are allocated across society, either to fundamental problems in the banking system and to depositors through an inflation tax or to taxpayers put in place sound, preventive bank supervision. through future taxes, bank supervisors must obtain public support. Successful loss allocation-while pre- Lesson 5 Failure recognition is important because a serving macroeconomic stability-depends on banking crisis is a solvency problem, not a liquidity issue whether the budget is able to generate a primary sur- plus to service its debts (including debts incurred by Bankers and policymakers tend to attribute funda- the carve-out) without excessive monetary creation. mental problems to a temporary liquidity (flow) In the global market foreigners will not be part of the problem rather than to a crisis of solvency (stock). burden-sharing exercise, except perhaps the multilat- Providing liquidity only buys time for restructuring eral agencies. But if policy credibility can be secured, to occur. Capital outflows or disintermediation from foreign capital flows, particularly foreign direct the banking system are symptoms of bank failure, not investment, can help generate the growth to pay for causes. The basic cause of bank and enterprise losses the losses. is overleveraging with inadequate capital relative to the risks undertaken. Failure recognition requires that Lesson 8 Success depends on sufficient real sector resources a proper diagnosis be made of the causes of bank loss- to pay off losses, adequatefinancial sector reforms to es, including the use of marked-to-market account- intermediate resources efficiently and safely, and the ing. Because of the lack of transparency in bank and budgets ability to tax "winners"and wind down "losers" enterprise accounting in many developing countries, without disturbing monetary stability losses often are significantly larger than anticipated or earlier reported. Insolvent banks can easily hide their Without growth, there are not enough resources to losses with bad accounting (evergreening of bad pay for bank losses. Without the right incentives and loans). The failure to deal with hidden losses can cre- social stability, there is no growth. Recession, unem- ate perverse incentives in the banking system, leading ployment, and banking crises all threaten social sta- to bad resource allocation and adding to macro- bility. Thus policymakers have to establish policy economic instability. credibility-in the form of public confidence-in 182 BANK RESTRucrURING order to create the preconditions for eradicating ing laws and court processes must be put into place. bank losses. Because financial markets are derivatives Banking laws and regulations should be enforced. of the real economy, bank losses are rooted in real Accounting frameworks should encourage the mea- sector financial imbalances-inextricably linking surement and disclosure of economic performance bank restructuring to fiscal and enterprise sector using international accepted accounting standards. reforms. In the post-centrally planned economies The payments and operating systems within the this phenomena is known as the troika problem. financial sector must work efficiently and robustly. Recapitalizing banks without addressing the underly- ing enterprise losses, inefficiencies, or fiscal problems Lesson 10 Time and timing are of the essence risks repeating banking problems in the future. Banking crises may seem to develop overnight, but Lesson 9 Rebuilding a safe and profitable banking system the structural problems of bank weaknesses generally requires goodpolicies, reliable management, and a strong are longstanding, hidden by a lack of transparency. institutionalframework The sooner problems are recognized and dealt with, the lower the costs to the economy and the banking A key question in the new world of financial innova- system. In the 1 980s-when economies were closed tion is whether bank management and policymakers and markets were less efficient-policymakers and are capable of supervision in such a volatile market bankers had more time to adjust to changes in world environment. Restructuring banks requires strong markets because of information asymmetry and mar- policy diagnosis, effective damage control, and effi- ket rigidities. But there is little time for protracted cient loss allocation in order to rebuild safe, sustain- decisionmaking in the current world of open capital able, and profitable banks. This requires strong min- accounts and high-tech financial markets. The failure istries of finance, central banks, and bank supervisors, to act quickly and decisively causes the swiftly mov- as well as a salvageable banking system. ing market to punish the policy error. At the micro- Building strong government and market institu- economic level, the failure of the Barings Bank sug- tions will necessarily take time. Thus overzealous pri- gests that market volarilities could be significantly vatization without first introducing appropriate regu- riskier than was anticipated in what appeared to be lation and national risk management measures may low-risk hedging strategies. Nevertheless, the more result in costly financial crises. Resources will have to you delay, the more you pay. be found to strengthen the supervisory process, as well as to develop risk management skills in the Conclusion financial community. Getting the incentives structure right is critical to While the instruments and circumstances are differ- rebuilding stable and sustainable markets and institu- ent, the basic principles of bank restructuring have tions. Incentives should not induce moral hazard not changed much since the 1980s. The important behavior. Neither should they inspire overregulation. issue of market and policy fundamentals remains the Balancing the risks and rewards of the marketplace same. Those who ignore the fundamentals will likely requires considerable reregulation efforts. For exam- have less time to deal with their shortcomings. pie, the current global financial system clearly is not a Supervisory frameworks must be expanded to level playing field. Investment banks, mutual funds, encompass both banks and capital markets. and hedge funds are not required to maintain the Furthermore, the emergence of a global financial net- same regulatory regimes and capital adequacy levels work means that there are large gaps in global super- as banks. Derivatives do not attract the same capital vision. The uneven playing fields of global financial requirements as credit instruments. Increasing the markets are fenced by domestic regulations that the transparency of instruments, institutions, and regula- market is rapidly finding ways to circumvent. tions would help to level the playing field. Banking Domestic banking problems now have global dimen- systems must be restructured to fit into a more mar- sions, and international cooperation is required to ket-oriented and volatile world. solve them. Moving toward this goal will require credible pol- The basic objective of oversight of capital markets icy measures that encourage stability, competition, is no different from that for banking systems: efficient and growth. Equitable and clear contract and bank- and sound resource allocation. Secondary objectives Post-Liberalization Bank ARsntrcturing 183 include protecting small savers, minimizing fraud, * Improving access to domestic capital markets, and protecting the channels of macroeconomic policy not only for smaller and newer domestic compa- management. The same instruments are used to nies but also for foreign (including regional) regulate market access, price, quantity, and product companies to widen the range of financial prod- innovation. The same regulatory principles apply: the ucts and investment opportunities for better risk supervision of solvency, through evaluation of capital management. adequacy and creditworthiness; liquidity, through * Adapting the legal framework to enable bank and proper treasury and risk management; earnings, capital market supervisors to cooperate and coordi- through costs of intermediation and efficiency; and nate better both domestically and internationally. quality of management, the human element. * Strengthening the financial infrastructure, such as Although significant progress has been made in recent the integrity and robustness of securities settle- years through the efforts of the Basle Committee on ment, clearing, and payments systems. Banking Supervision and the International * Deepening over-the-counter and exchange mar- Organization of Securities Commissions, much kets for new financial products, including futures, remains to be done to promote international coopera- options, and other derivatives, and creating a level tion in supervision. playing field for these tools in the regulatory Global markets would be much more transparent, framework. efficient, and better managed for risk if multinational * Promoting international cooperation in the closer corporations were less focused on their headquarters integration of regional and international markets, and domestic regulators were less concerned with to allow domestic markets to trade financial prod- protecting national capital markets. Local markets ucts on a global basis. selling local financial products to all consumers, Given the rapid changes in the global economy including some foreigners, is not what constitutes a and the market's ability to find ways around national global market. Global markets mean that the entire barriers, bank restructuring will continue as range of international financial products will be avail- economies and financial systems adjust to these new able for transaction locally. This has profound impli- realities. cations for the national and international regulatory framework. References The challenges of supervision and bank restruc- turing in a globally liberalized financial setting Kaufman, Henry. 1994. 'Structural Changes in the Financial include: Markets: Economic and Policy Significance." Federal Reserve Strengthening sectoral and national risk manage- Bank ojKansas City Economic Review 79(2): 5-15. ment, particularly by avoiding risk mismatches Moody Investors Service. 1994. "Derivatives Activity of U.S. that could be exploited by speculative arbitrage. Commercial Banks." New York. This requires paying considerable attention to Mundell, Robert A. 1967. International Economics. New York: getting the fundamentals right. Macmillan. getting the fundamentals right. Sheng, Andrew, and Yoon Je-Cho. 1993. "Risk Management * Creating a more level playing field between banks and Stable Financial Structures." Policy Research Working and other financial institutions. Paper 1109. World Bank, Washington, D.C. ,"., ~~~~~~~~~~~~~~~J3LA \U\\\\\\ ,,., _ " * *, l l _ | ! || l i