/28 4 Z - WORLD:ANIK L&TIM ASmfl"*N v V~~~~~~~ND- OARIBBLAN STUDIElS SE p ,qqe V ewpoints Work in progress for public discussion Consens ~~~~~~~~~~~~~~,,Ir s Yir Masar ~~~~~~~~~~~~~~~~H m ei-grcii Gefre__epir X~~~~~~~Kfe La Po- Doal Wi-iiklii r I) 'aniel Leder3aii I 10..t I )ailLeemI WORLD BANK LATIN AMERICAN AND CARIBBEAN STUDIES Viewpoints BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER by Shahid Javed B urki and Guillermo E. Perry THE WORLD BANK WASHINGTON, D.C. Copyright C 1998 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 September 1998 The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility wharsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. 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 per- mission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to copy portions for classroom use is granted through the Copyright Cleamnce Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923, U.S.A. Shahid Javed Burki is vice president and Guillermo E. Perry is chief economist of the World Bank's Latin America and the Caribbean Regional Office. Cover illustration by David McLimans Burki, Shahid Javed. Beyond the Washington consensus: institutions matter / by Shahid Javed Burki, Guillermo E. Perry. p. cm. - (World Bank Latin American and Caribbean studies. Viewpoints) Includes bibliographical references. ISBN 0-8213-4282-7 (alk. paper) 1. Latin America-Economic policy. 2. Caribbean Area-Economic policy. 3. Economic stabilization-Latin America. 4. Economic stabilization-Caribbean Area. I. Perry, Guillermo. II. Title. III. Series. HC125.B7665 1998 98-39051 339.5'098-dc2l CIP Contents Acknowledgments ......................................................................... vii Introduction ............................................................................. 1 Part 1. Institutional Reforn: Why and How ....................................................... 9 Chapter 1: Institutions Matter for Development ................................................... 11 What Do We Mean by Institutions and Organizations? ...................................... ......... 11 Why Do Institutions Matter for Development? ......................... ............................. 12 Information and Enforcement Problems in a Market: The Financial Sector ............ .................... 13 Information and Enforcement Problems in Hierarchies: The Education Sector .. 14 Institutions Matter: The Evidence ................................................................ 15 Issues of Measurement and Some Illustrations ..................................................... 1 5 A Brief Review of Recent Empirical Studies ......... .............. .. ............................ 17 LAC's Progress in Institutional Development .. 17 Notes ................................................................................... 24 Chapter 2: Institutional Reform Is Possible ...................................................... 25 Factors Increasing the Demand for Institutional Change ................................................ 25 The Supply of Institutional Change: The Role of Societal Actors, Interest Groups, and Collective Action .... ........ 26 Societal Interests in Change ................................................................... 26 Collective Action .......................................................................... 27 Latent Interest Groups ....................................................................... 27 Coalition Building ......................................................................... 28 Compensation Schemes ...................................................................... 28 Time-Inconsistency and Other Complications ........................ ............................. 29 The Intermediation of Societal Interests ............................................................ 3 1 Governmental Institutions in the Change Process .. 32 Some Guidelines for Reform ..................................................................... 34 Note ................................................................................... 37 Part 2. Institutional Reform in Markets: The Case of the Financial Sector ................................. 39 Chapter 3: Institutions, Governance, and Incentives in Banking Safety Net Arrangements ..................... 41 Information and Incentives in Banking ............................................................ 42 Loan Contracts ........................................................................... 42 Deposit Contracts ......................................................................... 43 Mitigating Agency Problems ................................................................ 43 Banking Regulation and Safety Nets .............................................................. 44 1ii BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Prudential Regulation without a Safety Net ............. ......................................... 45 Pressures for Bank Safety Nets before Central Banks . ................................................ 46 The Creation of Central Banks without Ex Ante Safety Nets ........................................... 47 Ad Hoc Safety Nets and Externalities ............... ................................. 47 Financial Repression ............................................ 48 Financial Liberalization ................................................ 48 Crises ................................................ 49 The Safety-Net Tripod: Capital, Monitoring, and Closure ............................................ .. 50 Capital ............................................ 50 Monitoring ................................................ 51 Internal Governance ................................................ 51 External Governance: Market Discipline ................................................ 52 External Governance: Official Regulation and Supervision . ......................................... 52 International Governance ................................................ 53 Closure and the Complexities of Exit Policies ................................................. 55 Fire Protection and Fire Fighting: Designing Exit Policies .............................................. 56 Early Warnings, Prompt Correction, and Intervention ............................................... 56 Bank-Failure Resolution ................................................ 57 The Role of Deposit Insurance in Bank Failure Resolution ............................................ 58 Safety Nets, Systemic Risk, and Catastrophe Insurance ................................................ 60 Summary and Conclusions ................................................ 61 Notes .................................................. 62 Chapter 4: Capital Markets and Legal Institutions .............................................. 67 Legal Protection to Investors ................................................... 68 Shareholder Rights ........................................ 68 Creditor Rights . . . .............................. ......................................... 74 Enforcement of Laws ................................ 75 External Finance and Legal Institutions .... ............................ 77 Data ................................ 79 Results ................................ 79 Regression Results ................................ 81 Conclusion and Policy Implications .... ............................ 84 Notes ................................. 85 Part 3. Institutional Reform in Hierarchies ............................ 87 Chapter 5: Reforming the School in Latin America and the Caribbean: An Institutional Analysis .... ............. 89 Educational Institutions and the School ...................................................... 91 Actors in the School ...... 92 Local External Actors ............................. 92 National External Actors ............................ 94 Problems with Educational Institutions . . . . . . 95 The Theory and Practice of Education Reforms ............................. 96 Second Phase of Education Reforms ............................. 98 The Political Economy of Reform .................................... 103 Prospects for Reform .................................... 105 Notes ..................................... 107 Chapter 6: The Challenge of Judicial Reform ................................ 109 Recent Changes in Latin American Judiciaries ..................................... 11 Improving the Current Approach .................................... 112 A Different Paradigm .................................... 113 CONTENTS An Incentive-Based Approach to Redirecting Judicial Behavior . .. 114 Attracting the Right Candidates ....................... 115 Incentives for the Best Judges ............................................................... 115 The Remuneration Issue .................................................................... 116 Boosting Productivity ..................................................................... 118 Drawbacks .............................................................................. 118 Conclusions ..............................................................................119 Notes .................................................................................. 119 Chapter 7: Public Administration in Latin America and the Caribbean: In Search of a Reform Paradigm .... ...... 121 The Problem of Public Administration: An Institutional Perspective .. .................................... 122 Constraints on Collective Action of Voters ..................................... ...................... 123 Special Interests and Other Problems ........................................ ....................... 123 "Models" of Public Administration ........................................ ...................... 124 The Hierarchical Model .................................................................... 124 The New Public Management .......................................................125...... 125 Voice .................................................................................. 126 Public Administration in Latin America and the Caribbean ....... ................. ... ............... 126 The Performance of Public Administrations ... 126 Informality in the Public Administration ....................................................... 128 The Origins of Informality .................................................................. 131 The Regional Reform Experience ...................................... .......................... 131 The Successes: "One-Off' and Enclave Reforms ................................................... 131 Reforms of the Core Public Administration: A Mixed Picture ........................................... 134 Reform: Constraints and Possibilities ....................................................... 135 Informality and Reform ....................................................... 135 Reforms that Address Informality ....................................................... 135 Reform Options ....................................................... 136 Notes ....................................................... 137 Technical Appendix Concepts for Analyzing and Designing Institutions .................................. 139 Data Appendix: Institutional Indicators Used in Figures 1-17 ......................................... 145 References ....................................................... 149 Boxes Box 1.1 Examples of Formal and Informal Institutions .12 Box 1.2 Examples of Organizations .12 Box 2.1 Time-Inconsistency Problems in the Political Economy of Financial-Sector Reform .30 Box 3.1 Market Discipline, Lender-of-Last-Resort Facilities, and Exchange Rate Systems .53 Box 3.2 International Harmonization of Regulatory and Supervisory Standards .54 Box 5.1 Minas Gerais, Brazil: Building Capacity in Schools .98 Box 5.2 Education Reform in Chile: The Whole Exceeds the Sum of Its Parts .100 Box 5.3 Colombia: A Battle Between Giants .101 Box 5.4 New Zealand's Education Review Office .102 Box 5.5 The Dominican Republic: Reform Captured by the Bureaucracy .104 Box 5.6 The Case of El Salvador: Stability, Consensus Building, and Sustained Reform .105 Box 6.1 Measuring the Performance of aJudicial System .117 Box 7.1 Public-Sector Reform: A World Bank View .122 Box 7.2 New Zealand: A Leading Example of New Public Management .127 v BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Box 7.3 Informality in the Peruvian Government ............................................... 129 Box 7.4 Budgeting and Informality in Developing Countries ...................................... 130 Box 7.5 Chile's Long March to Sustain an Effective Public Administration ............................ 132 Box 7.6 A Set of Reform Issues Emerging from New Public Management ...... ........... ... 137 Tables Table 1.1 Empirical Evidence of the Role of Institutions for Economic Growth, Financial Development, Inequality, and Poverty ............................................................ 18 Table 2.1 The Political Economy of Reform and Compensation Mechanisms ................. ........... 29 Table 2.2 Summary of Constellations of Presidential Powers over Legislation in Latin American Constitutions ... . 33 Table 2.3 Presidents' Parties' Mean Share of Congressional Seats in Latin America ......................... 34 Table 2.4 Relationship Between Presidents' Constitutional and Partisan Powers in Latin America ............. 35 Table 4.1 The Variables . .................................................................. 69 Table 4.2 Shareholder Rights around the World .................................................. 72 Table 4.3 Creditor Rights around the World .................................................... 76 Table 4.4 Enforcement of Laws . .............................................................. 78 Table 4.5 External Finance and Legal Institutions ................................................. 80 Table 4.6 Regressions of External Finance and Legal Origin ......................................... 82 Table 4.7 Regressions of External Finance and Shareholder and Creditor Rights ........................... 83 Table 5.1 Focus and Goals of Two Generations of Educational Reform in LAC ............................ 99 Figures Figure 1.la Institutional Development and Economic Growth ......................................... 16 Figure 1. lb The "True" Partial Coefficient Between Growth and Institutional Development ................... 16 Figure 1.2 Composite Institutional Index by Regions, 1984-97 ....................................... 20 Figure 1.3 Risk of Repudiation of Contracts Index by Regions, 1984-97 ................................ 20 Figure 1.4 Risk of Expropriation Index by Regions, 1984-97 ........................................ 20 Figure 1.5 Corruption in Government Index by Regions, 1984-98 ..................................... 20 Figure 1.6 Law-and-Order Index by Regions, 1984-98 ............................................. 20 Figure 1.7 Quality of the Bureaucracy Index by Regions, 1984-98 ..................................... 20 Figure 1.8 Composite Institutional Index by LAC Sub-regions, 1984-97 ................................ 21 Figure 1.9 Risk of Repudiation of Contracts Index by LAC Sub-regions, 1984-97 ................ ........ 21 Figure 1.10 Risk of Expropriation Index by LAC Sub-regions, 1984-97 .................................. 21 Figure 1.11 Corruption Index by LAC Sub-regions, 1984-98 .......................................... 21 Figure 1.12 Law-and-Order Index by LAC Sub-regions, 1984-98 ....................................... 21 Figure 1.13 Quality of the Bureaucracy Index by LAC Sub-regions, 1984-98 .............................. 21 Figure 1.14 LAC: Southern Cone Composite Institutional Index, 1984-97 ................................ 22 Figure 1.15 LAC: Northern Cone Composite Institutional Index, 1984-97 ................................ 22 Figure 1.16 LAC: Central America and Panama Composite Institutional Index, 1984-97 ..................... 22 Figure 1.17 LAC: Mexico and the Caribbean Composite Institutional Index, 1984-97 ........................ 22 Figure 1.18 Insecurity of Property Index by Regions ................................................ 23 Figure 1.19 Unpredictable Changes in Laws and Policies Index by Regions ................................ 23 Figure 1.20 Unreliable Judiciaries Index by Regions ................................................ 23 Figure 5.1 Preschool Gross Enrollment Rates in Costa Rica and Chile, Poor vs. Non-poor .................... 90 Figure 5.2 Achievement Test Results, 1992 ...................................................... 90 Figure 5.3 Average Mathematics Achievement Test Scores, Grade 8 .................................... 91 Figure 5.4 Principal Actors in the Traditional School ............................................... 91 Figure 5.5 LAC Education Reforms ............................................................ 96 Figure 5.6 Assessment of Educational Quality: Latin American Testing Systems, 1986-97 .................... 97 vi Acknowledgments T_ HIS REPORT IS THE PRODUCT OF TEAMWORK. THE IDEAS PRESENTED HEREIN WERE INI- tially discussed with staff members from the Latin American and Caribbean (LAC) regional office and from the Public Sector Management unit of the World Bank, as well as with spe- cialists affiliated with other organizations, who participated in a retreat on November 23-24, 1997. We are grateful to all of those who participated in that fruitful event, especially to Douglass C. North, Andres Velasco, Allen Schick, Phil Keefer, Florencio L6pez-de-Silanes and Alberto Chong. The research and writing of this report were conducted by a group coordinated by the Office of the Chief Economist for LAC, led by Guillermo Perry. This group was composed of Daniel Lederman (Chapters 1 and 2), Robert Ayres (Chapter 2), Augusto P. de la Torre and Yira Mascar6 (Chapter 3), Donald Winkler and Benjamin Alvarez (Chapter 5), Linn Hammergren and Richard Messick (Chapter 6), and Geoffrey Shepherd (Chapter 7). Philip Brock wrote a paper that was commissioned to serve as the foundation for Chapter 3. The chap- ter extensively incorporates text from that paper. Chapter 4 was commissioned to Rafael La Porta and Florencio Lopez-de-Silanes. We thank Conrado Garcfa-Corado and Pushan Dutt for their assistance in preparing the data used in Chapter 1. Marta Cervantes and Gladys Guerricagoitia provided secretarial support in the preparation of Chapter 2. Charles Griffin, Juliet Literer, and William Mayville assisted in the preparation of Chapter 5. Finally, Chapter 7 draws extensively from a paper written by Geoffrey Shepherd and Sofia Valencia (1996). We are also grateful to Eliana Cardoso, Sebastian Edwards, Carol Graham, Phil Keefer, Danny Leipziger, Ernesto May, Gobind Nankani, Anthony Ody, Lant Pritchett, and many others who took the time to read portions of the report, to provide comments on earlier versions, or to present their work on institutional reforms and political economy issues in our monthly seminar series. Marcus D. Rosenbaum edited the final version of this report. Shahid Javed Burki and Guillermo Perry vii Introduction IN 1990, A GROUP OF LATIN AMERICAN AND CARIBBEAN (LAC) POLICY-MAKERS, REPRESENTATIVES of international agencies, and members of academic and "think-tank" communities participated in a conference sponsored by the Institute for International Economics in Washington. Their purpose was to evaluate the progress achieved by LAC countries in promoting economic policy reforms after the debt crisis of the 1980s. At the conclusion of the group's deliberations, John Williamson (1990) wrote that "Washington" (at least as represented by those attending) had reached a substantial degree of consensus regarding 10 policy instruments.' With but one exception (namely, the protection of property rights) the policy prescriptions of the "Washington Consensus" ignored the potential role that changes in institutions could play in accelerating the economic and social development of the region; it focused instead on the issues of fiscal discipline, liberalization of trade and investment regimes, deregula- tion of domestic markets, and privatization of public enterprises.2 Emerging from the debt crisis, the region's main priorities were to achieve macroeconomic stability and to dismantle the basic elements of the protectionist model of development-priorities that the consensual view saw as necessary to reap the potential benefits from rising global trade and financial flows. Subsequent experience has convincingly demonstrated that the policies prescribed by the Washington Consensus are paying off. The Long March, a document we prepared for than of the distributive consequences of trade and financial last year's World Bank-sponsored conference on develop- liberalization. In particular, the resumption of growth has ment in LAC, which was held in Montevideo, reviewed the not been accompanied by strong labor demand in the for- existing evidence and concluded that the region's resump- mal sector (in many countries either formal unemployment tion of economic growth in the 1990s had been closely or informal employment have increased); export growth associated with the implementation of many of the policies has been concentrated in natural-resource intensive indus- recommended by "Washington." tries; and the wage differentials between skilled and non- The expectation, however, was not only that globaliza- skilled labor appear to have widened. Consequently, tion and the "first-generation" reforms would raise eco- income-distribution problems have not improved in many nomic growth rates, but that they also would significantly countries, and have deteriorated in others, which has reduce poverty and inequality. Indeed, capital inflows and resulted in poverty rates that remain unacceptably high. In export growth were expected to promote the development addition, economic insecurity for the poor and middle of labor-intensive sectors. This has not occurred. To be classes, linked to job insecurity and income volatility, has sure, the reforms have produced a decline in poverty rates, tended to increase. but this development seems to be more a consequence of The Long March concluded, then, that further reforms the decline in inflation rates and modest growth, rather were needed to achieve higher sustained rates of growth I BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER and to make a more significant dent in poverty reduction. analyzing and designing institutions (Chapter 1), and to It identified, in particular, the need to focus on improving evaluate how political-economy concepts can be used to the quality of investments in human development, pro- develop strategies for implementing institutional reforms moting the development of sound and efficient financial (Chapter 2). Employing some of the concepts elaborated in markets, enhancing the legal and regulatory environment Chapters I and 2, the report demonstrates that sound insti- (in particular, deregulating labor markets and improving tutional reform can be technically and politically viable in regulations for private investment in infrastructure and the key sectors we discuss (Chapters 3-7). social services), improving the quality of the public sector (including the judiciary), and consolidating the gains in The Growing Demand for Institutional Reforms macroeconomic stability through fiscal strengthening. The in Latin America and the Caribbean report showed that such reforms would entail considerable The globalization of national economies, the implementa- "institutional" reforms. tion of the first-generation reforms, and the process of This report examines the precise nature of the required democratization in LAC are contributing to a rise in the institutional reforms and provides a framework for their demand for institutional reforms. The combination of design and implementation. We hope that it will help exogenous factors promoting globalization (such as techno- launch a dialogue among policy-makers, civil society, and logical changes and economic and demographic trends in the academic community in LAC on how best to design the developed economies) with the implementation of and reform institutions-that is, on how to "supply" insti- trade and financial liberalization has increased the demand tutional reforms to meet new societal demands. for institutional reforms in LAC along three dimensions. First, they have increased the demand for institutional What Is Meant by Institutional Change? reforms on the part of the private sector, which now com- The terms "institutions" and "organizations" are often used petes in a global marketplace and has realized that its prof- as synonyms. This volume, however, does not treat them as itability or competitiveness is affected by the quality and such, adopting instead the definitions used by what is efficiency of the delivery of financial and public services, being called the "new institutional economics": Institu- the quality of education, and the effectiveness of the judi- tions are rules that shape the behavior of organizations and cial system. In other words, there is a growing awareness individuals in a society. They can be formal (constitutions, that institutional reforms in these areas are needed to laws, regulations, contracts, internal procedures of specific enhance the competitiveness of the private sector, which organizations) or informal (values and norms). In contrast, will in turn contribute to long-term economic growth. organizations are sets of actors who collectively pursue Second, the rapid growth of volatile capital flows has common objectives. Using these definitions, this docu- increased the demand for institutional reforms that may ment focuses on reforming institutions-that is, rules- help mitigate the risks associated with this trend. I'hat is, that determine the non-price incentives for the behavior of globalization, despite its positive features, has increased individuals and organizations. Specifically, in four key sec- the vulnerability of countries to developments occurring tors-finance, education, justice, and public administra- elsewhere, and there is a growing awareness that new or tion-the analysis emphasizes rules that may solve infor- more effective financial institutions (at both the iriterna- mation and enforcement problems. tional and the national levels) are required to reduce this Research conducted for this report revealed that there is vulnerability. much that is still not known about the determinants of Third, globalization has increased the demand for insti- institutional change. The document makes it abundantly tutions that can help reduce income inequality and provide clear, nevertheless, that analytical tools are available that social safety nets for people who are rendered more vulner- can facilitate analysis of this important topic. There is no able in the new competitive environment. This report, how- pretense of covering all aspects of institutional reforms ever, does not focus on this last set of topics, but only on here, nor is it intended to produce a detailed manual for those associated with the competitiveness and vulnerability action. The more modest objective is to examine how the issues-although educational institutions examined in concepts of the new institutional economics are useful for Chapter 5 certainly can affect equity and social protection. 2 INTRODUCTION This is a defining moment in the institutional reform Such developments open opportunities for all. They also process in LAC. In the context of more open economies, create new risks. Most developing countries, and LAC institutional reforms and related organizational changes are countries in particular, have decided to open their greatly needed to enhance the competitiveness of the pri- economies to trade, investment, and financial flows in order vate sector, to reap the potential benefits of the economic to participate from the potential benefits provided by glob- reforms undertaken in the last decade, and to reduce the alization. But often they have done so without an adequate LAC region's financial vulnerability. The greater percep- institutional framework. As the World Bank warned early tion of such needs on the part of many different social last year, in such cases the risks can be high (World Bank actors is creating a rapidly increasing societal demand for 1997d). Also, without adequate institutions, the potential institutional reforms. benefits of globalization in terms of higher growth and This heightened demand coexists with a situation in investment rates will either not materialize or be too con- which most LAC countries are lagging in institutional centrated, thus exacerbating, rather than easing, inequali- development compared with other countries with which ties and social tensions. they have to compete in international markets (see Chapter 1). In the following paragraphs we elaborate on some of Stability in Global Economies: Learning from these issues. Experience Recent international experiences have contributed greatly Demands Emanating from Globalization and First- to the demand for reform by the private sector and politi- Generation Reforms cal leaders. Prime examples are the Mexican peso crisis of The globalization of national economies has been proceed- 1994-95 and the more recent financial crisis in Asia. ing at a rapid pace in recent years. Merchandise trade flows Regarding the latter, economies that until a year ago had have grown three times faster than global production dur- performed extremely well for decades, and that followed ing the 1990s, and foreign direct investment (FDI) to generally sound macroeconomic policies, fell into a pro- developing countries has grown twice as fast as domestic found financial and currency crisis and are today experienc- production in these countries (World Bank 1997c). Also, a ing sharp recessions, rising unemployment, and social growing number of corporations now place their assembly unrest. Unsound financial and corporate-governance insti- plants, establish their offices, and sell their products and tutions created perverse incentives that led the private sec- services across the world. The countries of the LAC region, tor in those countries to misuse their access to huge short- liberalizing their trade and investment regimes to reap the term capital inflows, engaging in massive financial and potential benefits of such a process, have been active par- currency risks that rendered them and their economies ticipants in this worldwide transformation. highly vulnerable to changes in investors' sentiments. Globalization has received a substantial impetus in the In short, we have learned that in a world of global finan- last decade from the enormous increase in international cial integration, sound financial and corporate-governance finance facilitated by the technological revolution in com- institutions (in addition to sound macroeconomic policies) munication and informatics. This information revolution are essential for promoting a healthy intermediation and use has produced such a reduction in transaction costs for the of international capital flows. Without them, integrated financial sector that it has created a truly global financial economies are vulnerable to changes in investors' senti- market, stimulated the rapid development of new financial ments and thus to financial and currency crises, and private products, and increased the speed at which international agents may become so exposed to financial and currency capital flows change directions. This process is likely to risks that a devaluation or an interest-rate hike could have intensify in the years ahead as savings and demographic devastating effects on the real economy. That is, good macro trends in the developed world are likely to provide increas- policy is not enough; good institutions are criticalfor macroeco- ing capital surpluses intermediated by large institutional nomic stability in today's world of globalfinancial integration.3 investors with both the capacity and need to invest them Thus, industrial and financial groups in all emerging across the globe in order to maintain acceptable rates of markets are now realizing that their interests are exposed return and to diversify risks (World Bank 1997d). to high risks by the lack of sound institutions in these 3 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER areas. Such a change of perception has already facilitated brought renewed global attention to issues such as the access substantial progress in the reform of financial institutions to justice for poor people, the protection of fundamental in LAC since the Mexican peso crisis. (The financial crisis human rights, and democratic institutions in general. These of the 1 980s also led a few countries, like Chile and Colom- developments, too, are adding to the new demands for bia, to adopt significant reforms in their financial institu- reform of the institutional fabric of LAC countries. tions.) Authorities throughout the region are busily engaged in further reform in this area, with strong support In sum, globalization (and the powerful demonstration from public opinion and the private sector. Chapter 3 dis- effects of recent financial crises), previous reforms, clemoc- cusses the nature of the institutional reforms needed in the ratization in the region, and the end of the Cold War, have financial system (the so-called financial safety net) to opened a "window of opportunity" for implementing com- respond to this growing societal demand. prehensive institutional reforms aiming to alter funcdamen- It is not yet apparent to what extent the lessons from the tally the behavioral incentives for individuals and organi- Asian crisis will facilitate the reform of corporate gover- zations in the LAC region. These developments have raised nance institutions-i.e., those related to the disclosure of the effective demand for institutional reform. consolidated balance sheets by financial groups, share- Potential demand for reform, however, is not enough. holder rights, and the rights of creditors (including bank- The "supply" of institutional reform requires considerable ruptcy laws) that have proved to be so elusive throughout ingenuity, both in technical and political terms, to solve the region. The importance of such reforms for robust cap- the complex incentive issues involved and to overcome the ital-market development is demonstrated in Chapter 4. resistance of losers as well as the obstacles to "collective The inadequacy of present-day international institu- action" that usually weaken the effective support of poten- tions to deal with the new global financial market was also tial beneficiaries. patently revealed in these crisis. However, such a topic is not the object of discussion in this report. Factors Affecting The Supply of Institutional Change: Technical and Political Challenges Demands Stemming from Democratization What are the main considerations bearing on the supply of The increased demand for institutional reforms stems from institutional change? In the first place, some authors have an array of non-economic and non-financial factors as well. argued that institutional reform is more difficult to imple- With the important exception of Cuba, there is an ongoing ment than "first-generation" reforms (Graham and Nafm process of consolidation and deepening of democracy 1998). But pessimists disregard the fact that the first stage throughout the LAC region. Citizens are demanding trans- of reforms in LAC was accompanied by significant institu- parent and efficient governments, improved access to qual- tional reform in several areas such as central bank in(depen- ity education, and reliable and efficient judiciaries. The dence, a shift from import licensing to dispute-resolution deepening of democracy has also been accompanied by an (related to anti-dumping and other "unfair" trading) rules increased emphasis on governmental decentralization and in trade, and a complex new set of regulations for compe- the devolution of governmental responsibilities to local tition in the provision of infrastructure and public services. governments and communities, which also necessarily The indexes of institutional reform presented in Chapter 1 entails profound changes in institutions. The importance also indicate that substantial institutional reform has taken and complexity of the topic of institutional reform in gov- place since the early 1990s, especially in the protection of ernmental decentralization (the rules that shape intergov- property rights, by reducing the risks of expropriation of ernmental relations and local institutions) warrants a spe- private property and improving contract enforcement. cial treatment. It will be the exclusive subject matter of In addition to technical difficulties, it is undenaiable, our report next year; hence, it is not examined in this however, that a number of key political-economy factors report. may render further institutional-and policy-reform This process of consolidation and deepening of democracy extremely difficult to implement. For example, organized in the region is receiving significant support from the inter- and vocal groups that benefit from the existing institu- national community. Indeed, the end of the Cold War has tional setup normally invest resources to mobilize opposi- 4 INTRODUCTION tion against institutional change (through lobbying, cam- inhibits collective action by potential winners from insti- paign financing, and the press), and, for reasons discussed tutional reforms can be mitigated. below, it may prove difficult to mobilize the support of Similarly, resistance of some losers can be overcome or at potential beneficiaries. least reduced through different types of compensation On the other hand, external or domestic technological, schemes. Reformers may alter the technically optimal economic, or political shocks-plus the advancement of design of the reform in order to improve its political via- knowledge-change the way different actors perceive ben- bility. This approach brings the danger of making reform efits and costs of existing and potential policies and insti- possible at the expense of limiting its potential efficiency. tutions. This may change the resources available to interest However, as discussed in Chapter 2, there are institutional groups to invest in favor of or against reform, and they may reforms (especially those that provide "choice" and "voice" disrupt the existing political economy-enabling leaders to beneficiaries) that are sound from both technical and to exploit the "window of opportunity" for institutional political points of view. Compensation schemes may also reform. be included in careful "bundling" of a package of reforms, in such a way that losers of a particular component of the Dealing with Winners and Losers from Institutional package benefit from another, thereby lessening their resis- Reform tance to the proposed reform that may hurt them. The growing societal demand for institutional reform Finally, in those cases in which the optimal sequencing requires skillful political entrepreneurship to overcome the from a technical point of view may create political prob- resistance of losers and to mobilize the support of winners. lems, governments may attempt to "lock in" their future It is usually more difficult to mobilize winners than losers commitments through international agreements or consti- due to several factors. tutional changes that increase "exit" costs and thus enhance First, potential benefits of reform are usually more the credibility of commitments. For example, participat- uncertain and potential beneficiaries more difficult to iden- ing in international negotiations under the auspices of the tify, while potential costs are normally more certain and World Trade Organization or Regional Trade Agreements losers easily identify themselves. may enhance the credibility of economic and institutional Second, potential beneficiaries of reform are usually reforms (Burki and Perry 1998). numerous, dispersed, poor, uninformed, and unorganized Chapter 2 provides some guidelines for LAC reformers (what we have called "latent" interest groups), while that may help the design of politically viable reform strate- potential losers tend to be more concentrated, rich, gies, taking into account some features related to the polit- informed, organized, and vocal. ical and governmental systems in the region, including Third, appropriate sequencing of reforms from a techni- political parties and the nature of executive-legislative cal point of view may be "time-inconsistent" from a polit- relations. ical-economy point of view. For example, initial reforms may fortify interest groups that later can more effectively Examples of Technical and Political Challenges oppose the additional promised reforms. Chapter 2 exam- to Institutional Reform in Four Key Sectors ines some examples of these types of problems in the Institutional change is not something that can be easily sequencing of financial reforms. achieved, even in purely technical terms. Chapter 3, for example, illustrates the difficult tradeoffs involved in Overcoming Obstacles to Reform designing a safety-net system for the financial sector, which There are, however, ways of overcoming these obstacles to includes the central bank's role as lender of last resort, reform. The mobilization of potential beneficiaries can be deposit insurance, prudential regulation, monitoring, and enhanced by carefully crafted public information cam- intervention and closure rules. In principle, a financial paigns, in which information about the potential benefits safety net should provide reasonable protection against sys- and likely beneficiaries of reform-and even the identifica- temic crises, which may develop because of asymmetric tion of losers-become public knowledge. In this way, the and incomplete information available to depositors regard- "asymmetric information" problem that commonly ing the quality of their banks' operations and portfolios. At 5 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER the same time, however, the design of such a safety net participation) and "choice" (using credit and voucher should mitigate the moral hazard problem-that is, the mechanisms to enable families to choose from a greater incentives, accentuated by the very existence of the safety number of schools). net, that induce excessive risk-taking by financial firms A similar development is occurring with regard to the (which shifts some risks from depositors, and in some cases justice sector. The private sector is increasingly demanding even from equity owners and debtors, to the government or more reliable and efficient judiciaries. At the same time, taxpayers). the poor and the excluded are demanding access to justice. By the same token, Chapter 4 demonstrates that coun- The international community is also strongly urging tries with civil-law systems, especially from the French tra- reforms that provide independent and transparent judicia- dition, such as most of LAC, have generally found it more ries that effectively protect human rights. Substantial difficult to provide adequate legal protection to creditors efforts have already been undertaken in a number of coun- and minority shareholders and thus exhibit small, shallow, tries to give more independence to the judiciary, to mod- and concentrated capital markets that limit investment ernize rhe courts, and to increase the salaries of judges. and growth opportunities. Results, however, are still inadequate. Chapter 6 argues In reference to the low quality and low efficiency of LAC that there is a need to confront the complex political issues public-education systems-as evidenced by the poor per- related to the difficult balance between independence and formance of LAC students in math and science tests and by accountability of judicial systems, and to modify the incen- the region's high repetition and dropout rates, respec- tive structure in which judges and administrators operate. tively-observers historically have argued that such The chapter explores the potential benefits of requiring the neglect was a result of the fact that elites in the region full disclosure of judicial decisions, as well as of alternative tended to educate their children in private schools, and reforms in the system of remuneration of judges and judi- thus had no interest in public-education reform. As men- cial sector administrators. It illustrates, among other tioned above, the rise of global competition has now obvi- things, the technical difficulties involved in designing per- ously increased the interest of the business community in formance-based remuneration policies for judges. the quality of public education. Communities are also Finally, demand for public administration reform is also putting enhanced pressure for improving the quality of accelerating, due to the competitive demands of globaliza- public education. tion and the deepening of democracy in the region, as dis- Nevertheless, Chapter 5 makes it clear that reforming cussed above. Chapter 7 discusses incentive problems that the institutions of basic education to achieve more effi- are extremely complex for public administrations, since ciency and higher quality (in particular, changing the gov- they are intimately related to a chain of "principals" and ernance structure of schools to introduce autonomy and "agents," running from voters to elected legislative organi- parental control; the incentive structure for teachers and zations (congresses, councils) and executive officials (presi- their trainers, including the linking of payment structures dents, mayors), and from there to bureaucrats. Thus, pub- to performance; and mechanisms for increasing school lic administration institutions are inextricably linked to competition) remains a daunting task. This is so partly for political institutions. technical reasons, but also because there are powerful actors The chapter examines why LAC countries have hacl such (teacher associations and bureaucracies in the ministries of difficulty in constructing efficient and effective public education) that are likely to oppose and effectively block bureaucracies-in particular why there is such widespread such reforms, unless adequate inducements (for example, "informality" within the LAC public sector. It finds in terms of status, empowerment, and remuneration for numerous difficulties related to the inadequacy of political teachers) are included in the reform package. Equally as institutions to which administrative behavior is intimately important, the latent interests in favor of educational linked. It shows that as the consolidation of democratic reform (parents and communities) must be effectively institutions advances, especially as political parties mod- mobilized through public-information campaigns and by ernize, opportunities for efficient reform of core public empowering them with "voice" (shifting decisions and administrations will increase. resources to school councils with parental and community Chapter 7 also analyzes the feasibility of establishing 6 INTRODUCTION alternative models of public administration. In particular, with ambitious goals for education, followed by explicit the chapter examines the classic hierarchical model (based support for financial, judicial, and public-sector reforms.4 on checks and balances and highly centralized and strict This "Santiago Consensus," as coined by the president of personnel, budgetary, and monitoring systems, which are the World Bank during the Summit, may play the same characteristic of continental Europe) and the so-called New catalytic role for the reform agenda of the next decade as Public Management model (based on managerial discre- the one played by the Washington Consensus earlier in this tion and accountability, performance contracts, and com- decade. petition for service delivery, which was pioneered by New President Sanguinetti of Uruguay recently underscored Zealand and is being increasingly used, wholly or partially, this point when he emphasized that such a consensus has by more advanced countries). It concludes that some fea- emerged due to the widespread achievement of democrati- tures of the classic "hierarchical" model of public adminis- zation, economic stabilization, and resumption of growth tration will probably need to be maintained and consoli- in the region. He thus highlighted the fact that democra- dated before the region can move to the implementation of tization and the success of first generation reforms, both New Public Management. It is not a surprise, therefore, broadly endorsed throughout the region, are enabling that countries with relatively well-functioning hierarchical regional leaders to focus on the institutional and social civil services, such as Chile and Costa Rica, are precisely agenda.5 As the countries of Latin America and the those in the process of implementing reforms that move Caribbean confront the new millennium, it will be seen them toward the New Public Management model. increasingly that "institutions matter." From Washington to Santiago Notes This report thus recognizes that although institutional 1. According to Williamson (1990), the 10 policy instruments proposed by the "Washington Consensus" were fiscal discipline; pub- reforms have become a priority in LAC and are being lic expenditure priorities in education and health; tax reform; posi- increasingly demanded by societal actors, their implemen- tive but moderate market-derermined interest rates; competitive tation poses significant technical and political challenges. exchange rates; liberal trade policies; openness to foreign direct Successful implementation of such reforms undoubtedly investment; privatization; deregulation; and protection of property hinges on the ability of the region's political and technical rights. ledrhp eesfor optimism: 2. Williamson, however, concluded that, though he would have lheadership. Recdent history parovidelysugrnds inthedesign thought this was part of a tacit consensus, the region in practice had little concern with improving the protection of property rights, and and implementation of the first generation of reforms, and little had been achieved in this area. the recent challenges posed by globalization have been met 3. On the Mexican crisis see the articles in Edwards and Naim decisively, without backtracking on economic reforms. (1997), among others. On the Asian crisis see International Mone- More recently, the region's leadership has explicitly tary Fund (1998), World Bank (1998a), and Perry and Lederman accepted the challenge of responding to the increased (1998). 4. "Declaration of Santiago," Second Summit of the Americas, demand for institutional reforms by adopting many ele- Santiago, Chile, April 18, 1998. ments of this institutional reform agenda in the Santiago 5. Speech delivered at the Inter-American Development Bank, Summit of April 1998. The presidents' declaration begins May 1998. 7 I PART 1 Institutional Reform: Why and How CHAPTER 1 Institutions Matter for Development HE SANTIAGO CONSENSUS ACKNOWLEDGES THAT THE POLITICAL AND ECONOMIC ENVI- T ronment is ripe for launching a new set of institutional reforms in education, finance, justice, and civil service. One of the most difficult challenges, however, is to develop a coherent framework for analyzing and designing these new institutions. In this chapter (combined with the Technical Appendix), we attempt to clarify key concepts-discussing why institu- tions matter for development and examining empirical evidence that links institutional development to economic performance. What Do We Mean by Institutions and Organizations? A variety of meanings are commonly attached to the word "institution." Although it is often used as a syn- onym of "organization," we find useful the distinction made by the "new institutional economics" litera- ture. It defines institutions as formal and informal rules and their enforcement mechanisms that shape the behavior of individuals and organizations in society.' By contrast, organizations are entities composed of people who act collectively in pursuit of shared objectives. Thus, organizations and individuals pursue their interests within an institutional structure defined by formal rules (constitutions, laws, regulations, contracts) and in- formal rules (ethics, trust, religious precepts, and other that attempt to enforce such rules. Some of these rules reg- implicit codes of conduct)-see Boxes 1.1 and 1.2. Orga- ulate actors who provide market-enhancing services, such nizations, in turn, have internal rules (i.e., institutions) to as accountants, auditors, lawyers, and others. deal with personnel, budgets, procurement, and reporting Hierarchies are sets of rules for making transactions procedures, which constrain the behavior of their mem- based on vertical lines of decision-making authority. For bers. Thus, institutions constitute the incentive structure example, organizations often operate under internal rules for the behavior of organizations and individuals. that establish levels of responsibility and accountability, For economic analysis it is useful to distinguish between where some members are entrusted to monitor the perfor- two sets of institutions: markets and hierarchies. Markets mance of others. The specialized literature on institutional are a set of institutions (rules and their enforcement mech- economics has pointed out that hierarchies are set up to anisms) that set the stage for conducting discrete and establish contractual obligations-such as those between impersonal transactions, without requiring a continuous managers and their employees in private and public orga- contractual relationship. These rules can range from defin- nizations-to produce goods and services with lower trans- itions of location and timing for conducting certain trans- action and monitoring costs than would be required in actions, to more complex rules set out by contract law, pure market transactions (see definitions in the Technical financial-credit laws, and courts and arbitration procedures Appendix) (T. Moe 1984, 0. Williamson 1981). 11 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Laws and regulations arey oforsttlio incMuder o legiltvelomet?arbr the inentiv srnkx-eand tbh~fict h-beamar-oetsmmare"prect"; all actosaeasmeaohv m indivdualsand pi~r~iganiz Iossuha rs ii In pete and frel acuie pnorartion gabeomttenqalty aen- ~~di~ operation of markets Ed~picesofthe goodsandiservics~ inall tranvacteos as well - C~inracts f tivi ~ervars as s the aseawihacoutru iother andrs reiieabsiitySoindiiuasanlfim c among private inviduals o orgnatinanose efiia n ty whatos ibuy or sel andfromwh -personnel budgetary, pro:nremen reporting, and itatis,ocs dec s,ision iscmade at i carried ou ste hl - audt prcedors ar the for-reealn tisleealtitatioaffecsuppitheoul incentives within publideeiclivr heorouctanganiie i teatredonsn - ~~dnsti1t~~~~rtionaL ~tit andqualhy,aadoth buyecwoud-pahthe-orrepcnd ri~is~ opoliicias atthe-varous evel 6f~vrinWh DmouIntstinationly Masileorn eeopet cal ~sppsihiitisofa~d t~r~lrio~ a~ong he Trhmrket are "pefect" mallkatos aore assume toi haescritom-i ~oilevls fgo&fienr(i..,~a~a~ ~ov~ic~i,dty plreasnda rely acquiraedwa informationabou tequaisty cand and other governments~~~ prcsof-are thegodsind servlicesoin all9trnsactons easpwell Examples of ~oforal lNSbtUtJOO ca choseal doeffincintlyswatto thuy to clsellyn firomi whom Trus or he tndecy t cooerar aion~idivind neItclsoassiaasumesthttioshere parenoenforcmenth problems; who nconeer ach therinfequetly san~nfrover tat s onew day adecsone tismae itcois carriedior smothlyh rl baced hon evfac that idiaidaly ets iao rules . an c ost-r and I tics idail situation, the s-upimer wo ul gokid trne ctnse sfwa wtype ofrbeaniaiornil andrelsewhe-e ineved deliver thase pouth ade eric in the agrehopree quan-ee thaterallomewhrebtwplysen imporktsant proe hinterarchie. tity anfwdifquality andutermbuyers woud payth norretospfond By way or i alusta son alotnsde toeN co etalnandividual Tfniheoresa aad frewnares. fhorg whic this daescitone mis- vlserviorestablishing infomial srvets nofacndsu t asonably accurate-what i, bnsiu onal eomts calt. i onivstal nr tha are oeimplii' t uishsuerfo ally reatit A spot-mret trad gt wh (t w iliam 199 4. For ak the g exap le sxpaicith beatiora ou tomes,whaic aid complemented by shop pig forv goearie at supea on a periodi basis a sysem fr reortig an montorig peformnce s is usuly domgane in circumstnes that coelyn miretur them Bale and Dale 1998, Schick 199). In this way, public ser- neossicma assu im tions. S ers compensute po duct as makin tn sationese iwt a hie organ tiou an e l nstitutione r even allowed to t M ai e toe Shoppers canheve teath res tha t fall s w e betweenti markets and "ppethiemarhes. try a hfew hderent serm aketws anet insfomadtion frdom aealy o iuestratlion, considerthe New Zerfalrand e m aoe nceibs, and frinds. Thicsh we ayl sasues asso mis- civl sevc refrm wher public serantsXa nowEjaagSMS fgace aset ofg taksabinttheprocesrs,rebasiciySidvdallywsan choorse htwat inentives. tSehaptrewrd or punis pefrmnereaiv o aned wet gtwhat woe opaylorxa werdo tiake)tegos. ihu asystemforreporting andmonitoring performance (see buyItaloassume them agai; ee casesrwe may evforceenretu hem. Bal0e - an D ale 198 Scic 198) Ina a il thiswyaubie- th isuponermarkt deisin turn gaeitis cpnatred onutie asotl vants mus goprt in a hirach, but wit institutional d shopest-aree Inothi allowe stoaleave the spremises witoutd fEaue tha mim .....ic,, theg a2a incentdeives ofe copetitive markets,vipaying,togh tee maybeafew cuassoshpitng,ba naey by esablia,,,2shSfiin{ gg a link bewe pefomac an chks an ceitg carduns (wich ar atually isusasshi-n rewrds (Se Chater 6 ZXa r aqr i an 7.)i ate wTher the morew cmplkexs worl ofic fhinance). itini :. ggg SE a iSsl< g,LdvS o,, g)>;d ol co C° r-~ CO c Oo_Ilr Note: A rise in the index represents a reduction in the risk of repudiation of contracts. FIGURE 1.10 FIGURE 1.13 Risk of Expropriation Index by LAC Sub-regions, 1984-97 Quality of the Bureaucracy Index by LAC Sub-regions, 198498 10.0 - 6.0- 9.0 8.0 - , 5 . 0 5.0 8.0 - 29 7.0 c 4 ^.0 ............- 6.0~~~~~~~~~~~~~~~~~~~. 6 .o a E 30- - -- -- ----- - --t 23.0 4.0 . .. . E ~3.0 t2.0- 2.0- 1.0 .1.0 - 0.0-] 0 1 I III O _ ON ON ON ON ON ON ON ON ON ON ON ON O te N ON< 0 v- N 0 N CO> N N N0 N0 N N N ON ON ON \O ON ON ONONO Note: A rise in the index represents a reduction in the risk of expropriation. -| 4-+- Southern Cone - Northern Cone - Central America and Panama + Mexico and the Caribbean 21 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER FIGURE 1.14 FIGURE 1.16 LAC: Southern Cone Composite Institutional Index, 1984-97 LAC: Central America and Panama Composite Institutional 45.0 - Index, 1984-97 40.0- .E. ........................ <1 4(; p~ 4. 35.0- 30.0 0- 0~~~~~~~~~~~~~~~50 0 25.0- - E ' 0 0 0 00 0 E 20.0-2 C: 0 15.0- 15.0 10.0 - ....... ......... ......10.0 - . ..... .... ~~~~~~~~~~~~~~~~~~~~~~~5 .0 - - .. . .. .- -- - - - - - - - - - -- ---- . ...-.. 0.0- . 5.0-~~~~~~~~~~~~~~~~~~- - ct co cc co co cc C Cs G CI Cs t01 0at t t C CS Cs C ~11 21 co C c, Co CO Cs Cs Cs CO Cs oC C G CO Cs Argentina Brazil + Chile -4--- Costa Rica ---- El Salvador * Guatemala + Paraguay + Uruguay -4- Honduras St Nicaragua 0 Panarma FIGURE 1.15 FIGURE 1.17 LAC: Northern Cone Composite Institutional Index, 1984-97 LAC: Mexico and the Caribbean Composite Institutional Index, 1984-97 35.0- .. 40.0- 30.0- _ l 35.0 - 30.03 25.0 ° 200- . g> . 1 ~~~~~~~~~~~~~~~25.0 ^2/ 20.0 E .E/* 20.0- ................ ~ c 15 .0 a. ........................ . . . . . r 10.0 ..... ........ A 5.0~~~~~~~~~~~~~~~~1. 0.0 . .. ... ....... . .......... 05.0 - 0.0 - I I I I I I 0.0- I -f aS 'C (N t- at Cs 't CO a 0 (N - as IC \0 at a0 Cs O N a. S IA 'C ( a C .0 x x 0G ' NG \G E I, CO CO (z C 0Y ; C, Cl, 0 GI Ct , CG.G Bolivia Colombia Ecuador - Dominican Republic Guyana | Peru + Venezuela -*- Haiti -4-- Jamaica Mexico - Trinidad 22 INSTITUTIONAL REFORM: WHY AND HOW FIGURE 1.18 Insecurity of Property Index by Regions 90 80 70 0 ~'50 r o 6°- I 1 ... .. ... . ....... | 11 - ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. .. . .. . C)0 4 tA1.' :......I|0 1:: t20 F~ 10 10 10- - ',L LK M_|N|l DCs SSEA MENA CEE LAC SSAFR CIS FIGURE 1.19 Unpredictable Changes in Laws and Policies Index by Regions 80 70 60 -i _ O 50 -14 - 40 . 30 9 301z1;; 20 1 0- 10 DCs SSEA MENA CEE LAC SSAFR CIS FIGURE 1.20 Unreliable Judiciaries Index by Regions 80 - II ....~. . . . . . . . . .---5mz---- 70 - 60 40 -~L 20 10 0 DCs SSEA MENA CEE LAC SSAFR CIs Note: Indexes reflect the percentage of survey respondents who thought that this conssderation was an important obstacle to conducting business. Key: DCs = Developed Counteies SSEA = South and Southeast Asia MENA = Middle East and North Africa CEE = Central and Eastern Europe LAC = Latin America and the Caribbean SSAFR = Sub-Saharan Africa CIS = Commonwealth of Independent States 23 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER in the areas of expropriation risk and contract enforcement. tion of institutional development that was not explained by the con- So, much remains to be done to improve the quality of insti- trol variables (again, the residuals). Finally, we estimated the linear tutions-and thus to help accelerate the pace of long-term relationship between the two sets of residuals, which gives us a mea- groh a d. T remaining sure of the "true" relationship between growth and institutional development. It is worth pointing out that to some extent we have question is how policy-makers can promote institutional dealt with the issue of causality by using the initial level of the com- change, which can be a challenging task indeed. posite institutional index (corresponding to the year 1984) in these regressions. Notes 4. The coefficient of the composite institutional index in the sim- 1. North (1990) and 0. Williamson (1985) are well-known con- ple regression shown in Figure lia is significant at the 1 percent tributions to this extensive literature. North (1990, p. 73) defines confidence interval. The "true" partial coefficient shown in Figure organizations as "purposive entities designed by their creators to 1. lb is significant at the 11 percent confidence interval. The residu- maximize wealth, income, or other objectives defined by the oppor- als exercise was also done using data for two sub-periods separately tunities afforded by the institutional structure of the society." (1984-89 and 1990-95), and on the pooled data for both sub- 2. Education is a "public good" in at least two aspects: First, periods. The results of these exercises confirmed the positive and sig- when a public school is set up, all households in that jurisdiction can nificant relationship between growth and institutional development. benefit from it. Second, and more generally, education tends to pro- 5. According to Chong and Calder6n (1998), one potential expla- mote social peace and economic prosperity in the long run, which nation for the apparent positive relationship between formal institu- benefits all households, regardless of whether they have contributed tional development and inequality is that in some developing coun- to the financing or administration of the schools, or even whether tries informal institutions tend to predominate in the business they have children in school. transactions of the poor. Hence it is possible that the development of 3. The econometric procedure was the following: First, we esti- formal institutions benefits the formal sector more than the informal mated the relationship between growth and a set of standard sector. But this is just an untested hypothesis, and the results of this explanatory variables-the initial level of educational attainment in study are based purely on cross-country regressions, which do not tell the population, the initial level of GDP per capita in constant dol- us anything precise about the potential dynamic effects of institu- lars, the initial ratio of investmenr-to-GDP, the average ratio of trade tional development on poor countries over time. over GDP, the average inflation rate, the initial level of M2 to GDP, 6. In last year's The Long March, we also used ICRG data to assess the average variation of the terms of trade, and the standard devia- LAC's progress in institutional reform (or governance), and in that tion of the terms of trade. (See the Data Appendix for a detailed occasion we complemented the analysis with indicators provided by description of these variables.) From this procedure, we calculated the Business Environment Risk Intelligence (BERI). Unfortunately, this the portion of the growth rare that was nor associated with these con- latter service has been discontinued, and thus we were unable to trol variables (the residuals). Second, we ran a similar regression that update that data. estimated the effects of these control variables on the composite 7. See Burki and Perry (1997) and Inter-American Development index of institutional development, and similarly estimated the por- Bank (1996) for a review of progress achieved in structural reforms. 24 CHAPTER 2 Institutional Reform Is Possible HAPTER 1 DISCUSSED HOW IMPORTANT EFFECTIVE INSTITUTIONS ARE FOR ECONOMIC performance, and how urgent institutional reform is for Latin American and Caribbean countries. It demonstrated that efficient institutions should provide clear, widely known, coherent, predictable, credible, and properly and evenly enforced rules. Although pre- dictability and stability of institutions are important for promoting investment, stimulat- ing growth, and reducing poverty and inequality, institutional adaptation and change also are important. Such adaptability ensures that the incentive structure accommodates changes in technology, social preferences, external factors, and institutional innovations elsewhere. Given the importance of institutional change, this chapter turns to the question of how such change occurs and how to accomplish institutional reform. Political economy becomes a crucial consideration in this regard, as it often has a determining impact on the nature and extent of institutional reform at differ- ent times and under different circumstances. The political economy of institutional change in Latin Amer- ica and the Caribbean remains a relatively underexplored area of inquiry. Nevertheless, disparate studies in recent years have pointed to some key factors that appear to facilitate institutional change in the region as well as those that might impede it. This chapter attempts to point out some of the most salient political- economy factors likely to affect the potential for institu- receive greater volumes of foreign investment as a share of tional change, and to show how these factors, if taken seri- domestic investment than they did two decades ago. As ously into account, can enhance both the analysis and the noted in the introduction, this increasing activity in the undertaking of such change. global economy has led entrepreneurs and firms to realize that their international competitiveness is critically affected Factors Increasing the Demand for Institutional by the quality of domestic institutions. Both the spread of Change the Mexican peso crisis of 1994-95 and the more recent An increase in the effective demand for institutional Asian financial crisis demonstrated the increased vulnera- change comes from many sources, not all of them politi- bility of national economies to external shocks in a world cal-technological innovations; external economic shocks; characterized by rapidly increasing financial integration. In natural or manmade disasters; and foreign experiences, addition, globalization has brought with it potentially good or bad, that have powerful demonstration effects. adverse distributive effects, which need to be offset by new One prominent example in contemporary Latin America institutional structures (i.e., social safety nets and improved and the Caribbean that has already been discussed is glob- access to quality education for vulnerable groups). alization. Countries in the region now trade a higher share On another front, the end of the Cold War and the of domestic production with the rest of the world and global and regional emphasis on democratization and the 25 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER advancement of human rights have created demands for The Supply of Institutional Change: The Rolle of more accountable and transparent political institutions and Societal Actors, Interest Groups, and Collective a reformed judiciary. In Latin America and the Caribbean, Action the devolution of power to local governments, the popular The factors affecting the "supply" of institutional election of mayors, the increased importance of national change-including how much change takes place, at what legislatures, and the emergence of grassroots organizations pace, and along what dimensions-are numerous an-d var- and new political parties are all manifestations of these ied, and we do not attempt to present a comprehensive list. increased demands. There are many complex technical issues involved, but we Developments such as globalization act as facilitators of focus here on the political-economy factors affecting the institutional change; they provide the dynamic context in likelihood that increased demands for change actually will which such change takes place. They increase the societal yield results. This is because the nature of the institutional perception that it is "time for a change." For example, the changes undertaken is likely to turn significantly on how economic crises of the 1980s led to the widespread accep- the demands for change previously cited are (or are not) tance and subsequent adoption of liberal trade and invest- processed by the political system. The ensuing discussion ment regimes recommended by technical experts inside focuses on three broad areas: societal actors and the factors and outside the region. In turn, the adoption of these new affecting their ability to act collectively; political organiza- policies required changes in the rules governing interna- tions like political parties that serve an important interme- tional trade, and even led to the dismantling of some orga- diary role between such societal actors and the formal insti- nizations, such as import-licensing agencies, and to the tutions and organizations of government; and some creation of new ones, such as antitrust agencies and those characteristics of the formal institutions themselves. that evaluate the impact of "unfair" trade practices. In short, the prevailing institutional structure can become Societal Interests in Change widely perceived as incongruent with the way the world (or In attempting to understand why and how institutional a country or region or sector) is evolving. Factors that change takes place, it is necessary to identify and accurately might previously have narrowed the parameters for characterize the major societal-level interests at stake in change-most notably the accumulated weight of history such change. Obviously, these will vary from country to and culture-become, or appear to become, less constrain- country (and even from region to region within coLntries) ing. The effective demand for institutional change is thus as well as from sector to sector. increased. In this connection, it is necessary to understand what In the long run, the main factor propelling the demand societal actors-with what interests-are attempting to for institutional change is learning or the accumulation of effect institutional change. It is necessary to assess what knowledge (North 1990). Over time, actors evaluate the resources-financial, organizational, and political-they functioning of institutional arrangements based on their are capable of bringing to the table of reform. And it is own experiences. Accumulated knowledge helps actors necessary to assess who benefits and who loses from current develop more effective rules to overcome the underlying institutional arrangements, and who would benefit and problems that existing rules were designed to solve and lose from the changes. In short, it is essential to have a clear enables them to identify new problems requiring institu- understanding of the major social cleavages surrounding tional solutions. Without social learning a society can be institutional change, and to know the nature and intensity trapped for decades or more in a perverse equilibrium char- of the demands of the main societal actors. acterized by low-quality institutions, slow accumulation of Bardhan (1997) explains how a society's institutional knowledge, and meager growth and institutional change. arrangements "are often the outcome of strategic distribu- The role of social learning in institutional change points to tive conflicts among different social groups," and how, the crucial importance of education over the long run. therefore, inequality in the distribution of power and Regardless of the underlying forces increasing the effective resources can sometimes block necessary reform of these demand for institutional change, education is critical for institutions. The flip side of this, of course, is that changes promoting an adaptive institutional structure. in the relative capacity of organizations to influence insti- 26 INSTITUTIONAL REFORM: WHY AND HOW tution-building may lead to institutional changes. Emerg- are smaller. Hence, the costs of collective action are associ- ing powerful organizations may deem the existing institu- ated not only with the resources needed to pursue a partic- tional structure to be inconvenient for their own interests; ular objective (e.g., political campaign contributions, lob- or existing organizations or interest groups see either their bying, etc.) but also include the costs of monitoring the net benefits derived from the existing institutional struc- behavior of members to prevent free riders. Any factors that ture sharply reduced or their own capacity to oppose insti- may reduce such costs usually enhance the ability of mem- tutional change weakened (both of these effects appear in bers to act collectively as a group. times of economic or financial crisis). In addition, eco- In addition, the pattern of distribution of the costs and nomic crises and other exogenous developments such as benefits of particular institutional reforms also influence technological changes can change the perceptions of inter- the likelihood that affected interest groups will act collec- est groups or organizations about the costs and benefits of tively to support or oppose such reforms. It is often argued particular institutions. Hence, not only can changes in the that reforms that yield benefits to a large number of people relative capacity of organizations bring about institutional but that have negative consequences for specific groups change, but also changes in the perceptions of influential will be especially difficult to implement. When the bene- organizations can lead them to support or reduce their fits are dispersed, according to the argument, they will opposition to institutional reforms. seem small relative to the costs of acting collectively The recognition that organizations may lead the way in (Becker 1983), so support for the reforms will tend to be changing institutions is an important ingredient of insti- weak and poorly articulated. To clear and concentrated tutional change. Of particular importance is the political losers, however, there will be strong incentives to cover the concept we can call the "private interest" theory. This the- costs of collective action to oppose the reforms. ory emphasizes that private interest groups can seek policy and institutional changes (or prevent such changes) to serve Latent Interest Groups their own interests. This approach thus echoes that of An important implication here is that in any society we Bardhan in emphasizing the role of distributive conflicts may find groups of individuals or organizations that have a among interest groups as an important determinant of common interest but are unable to organize themselves as institutional change. an effective interest group; Olson (1965) referred to these groups as "latent" interest groups. That is, for many poten- Collective Action tial beneficiaries of a particular institutional reform the One prominent way of addressing these concerns in the costs of collective action may be too high, and their collec- economic literature on institutions is through a focus on tive voice will not be heard in policy debates. The costs of the costs for individuals and, especially, organizations to collective action and the resources necessary to promote it, act collectively in pursuit of their interests (see Olson however, are not the only impediments to collective action. 1965). "Collective action" is costly for each member of a Lack of information regarding the details and consequences group, yet the benefits from collective action accrue to all of policy prescriptions is often an obstacle, too. In other members. Thus, there is an incentive for individuals or words, voters may clearly understand what their interests organizations to be free riders and let others pay the orga- are, but they may not know the specific details and techni- nizational and coordination costs for group activities, cal issues of policy debates. including political lobbying (Becker 1983). Such latent interest groups seem especially numerous in Olson (1965) identified several characteristics of interest Latin America and the Caribbean and may be linked to groups that affect the likelihood of a group's being formed many factors, including the inequality in the distribution and how effective such a group would be. For example, the of income and wealth for which the region is unfortunately smaller the number of members, the greater the likelihood well known. Indeed, such inequality may be the key bar- that collective action will take place, because the smaller rier to collective action for many actors in Latin American the group, the lower the costs of coordinating and enforcing and Caribbean society. Many potential interest groups have membership requirements. Likewise, if the members are lacked both the incentives and the resources to engage in geographically concentrated, the costs of collective action effective collective action. 27 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Within Latin America and the Caribbean, many such from reforms to reduce their opposition and raise their sup- latent interest groups have historically been "bought off' port (Edwards and Lederman 1998). Table 2.1 describes or co-opted by the clientelistic distribution of ad hoc and five types of compensation mechanisms that have been used piecemeal "pork" and favors on the part of politicians, par- in various situations. ticularly at the local level. Such patron-client ties, arguably * The first type is direct compensation, which is usually in the prevalent form of political representation in much of the form of fiscal transfers or subsidies to losers from the region over an extended period of time, are antithetical specific reforms. For example, adjustment assistance to the development of more structured, broader-based, and schemes are commonly implemented jointly with more effective political organizations. The proliferation of trade liberalization programs. In the case of Chile, for grassroots organizations throughout the region in recent example, the government offered a minimum employ- years, however, has the potential to significantly alter the ment program in the 1970s to alleviate (although pattern just described. It could greatly intensify the effec- meagerly) the unemployment caused by trade liberal- tive demand for institutional change stemming from the ization and cutbacks in public employment. societal level and could greatly increase the possibilities for * Indirect compensation may emerge out of economic constructing pro-change coalitions. forces (such as the devaluation of currencies that usu- ally follows the implementation of stabilization and Coalition Building trade liberalization policies), or may be deliberate pol- An emphasis on the constellation of societal forces at play, icy measures that compensate groups affected by a and on how such forces do or do not engage in collective particular reform through the adjustment of a differ- action in pursuit of their interests, leads to a concern with ent policy instrument that indirectly raises their rev- coalition building and the construction of pro-change or pro- enues or reduces their costs of production or organiza- reform coalitions. The key concem is how to build such pro- tion. In several cases of education reforms, for change or pro-reform coalitions in specific countries and sec- instance, the centralized structure of collective bar- tors. What is the role of compensation mechanisms, and what gaining of teachers' salaries and salary levels have been are the roles played by timing, sequencing, and uncertainty? maintained in order to reduce the opposition of teach- A number of analysts have examined the factors that ers' unions to other education reforms. seem to account for the building of "pro-poor" coalitions in * Cross compensation entails changes in other policies that the context of the political economy of poverty reduction. raise revenues or reduce costs of potential supporters A central conclusion of these analyses is that it is necessary not affected by the change in the policy under consid- to include payoffs for non-poor groups as well (Ascher eration. For example, Bolivia's capitalization program 1984; World Bank 1990). Some institutional changes raised public support for privatizations, while the pri- appear to pit the poor against the non-poor directly, but vatizations themselves would not necessarily have frequently the fortunes of both groups are linked, and negatively affected the population at large (in fact, coalitions can be formed that cut across the poor/non-poor they should provide overall benefits). divide. Coalitions may form, for example, along sectoral * Exclusionary compensation relies on exemptions from lines (e.g., agriculture vs. industry) or geographic lines reforms to certain groups that would otherwise block (e.g., the interests of Brazil's Northeast versus those of the the reform effort. A case in point is the Chilean mili- more developed southern region). Where institutional tary's non-participation in the privatized social secu- reforms to benefit the poor have been effected in Latin rity system. America and the Caribbean, their success has generally * Finally, political compensation can be achieved, for turned on the stance of white-collar workers, professionals, example, by incorporating leaders of opposition and small- and medium-size business interests. groups into reform-oriented governments. The fact that winners and losers can be compensated for Compensation Schemes either not opposing or actually supporting institutional More generally, there are various types of compensation reforms also implies that reform proposals can include sev- schemes that can be used to compensate losers and winners eral policy changes in a single package. Hence, proposed 28 INSTITUTIONAL REFORM: WHY AND HOW reform packages can have "built-in" compensation mecha- TABLE 2.1 nisms, whereby one element hurts the interests of one The Political Economy of Reform and Compensafion group, but another offers compensation. Mechanisms MECHANISMS MAIN FEATURES AND SOME EXAMPLES Time-Inconsistency and Other Complications A. Direct Compensation Groups directly affected by the reform policy are The consideration of compensation schemes, however, compensated through the transfer of cash or financial brings further political complications. For example, are securities. In this way the authorities expect to see a teduction in the extent of opposition "promised" compensation schemes to be implemented in from that group to that particular reform. the future sufficient to placate opposition to institutional Examples of this type of compensation mecha- reforms? This type of question is the so-called "time-incon- nisms include the distribution of shates of pri- vatized fitms to workers in that particular sistency" problems that are common in discussions of firm, and adjustment assistance programs to macroeconomic policy choices (Kydland and Prescott workers who lost their jobs as a consequence of trade liberalization. The increase in take- 1977). Promised compensation schemes can be derailed if home pay following a social security reform is the initial reforms strengthen certain interest groups that another good example of this type of direct- subsequently oppose the implementation of the compensa- compensation scheme. tion mechanisms. Time-inconsistency (i.e., when elements B. Indirect Compensation This mechanismn implies compensating groups affected by a particular reform through the adjust- of the reform package or compensation schemes are not ment of a different polio' that indirectly raises their credible over time) can also produce a status quo bias as revenues or reduces their costs of prodaction. In potential supporters may withdraw from the coalition some cases this type of indirect compensation is "automatic," and is the result of normal eco- when promises of benefits are not credible. Box 2.1 dis- nomic forces at work. In others it is the result cusses how a technically optimal sequence of financial-sec- of specific policy measures. One of the most tro reforms can suffer time-inconsistency problems due to important indirect compensation mechanisms is the real exchange rate. By devaluing the real political factors. exchange rate, import-substituting sectors are Graham and Naim (1998) have also argued that the partially compensated, while exporters experi- timing of the costs and benefits of institutional reforms ence an additional boon. Providing tax exemp- tions to sectors affected by deregulation con- presents special political challenges. For example, they stitutes another common form of indirect argue that a key difference between stabilization programs compensation. and institutional reforms is the timing of the costs and C. Cross Compensation This mechanism entails transferring resources- benefits of the reforms in question. Macroeconomic stabi- either directly or indirectly-to groups not directly affected by the reform. in order to obtain their lization has immediate positive consequences embodied in political support. Transferring shares of priva- the reduction of inflation and perhaps in the acceleration of tized firms to the population at large-as in Bolivia's capitalization program-is a good growth. In the case of institutional reforms, it can be illustration of this mechanism at work. argued that some of the benefits become apparent only D. Exclusionary This mechanism entails excluding certain powerful after an extended gestation period, while the costs of the Compensation groups from the effects of a reform, or implementing reforms (felt mainly by well-organized interest groups) are (i.e., Exemptions) policies that in effect exempt soue sectors from the reform in order to diffuise their political opposition. absorbed up front. In this context, it may be politically dif- By allowing these groups to maintain certain ficult to implement institutional reforms-more difficult, privileges, they are not likely to become active at least, than macroeconomic stabilization programs. antagonists. The special treatment given to the Chilean armed forces regarding that coun- However, it should also be noted that stabilization in try's social security reform is a classic example Latin America and the Caribbean (or "first generation" of this type of compensation mechanism. reforms) has been accompanied by a variety of other E. Political Compensation This mechanism encompassespolitical 'carrots and reforms, including trade liberalization, which also produce sticks'-for example, the appointment of influ- ential representatives of certain groups to up-front costs and delayed benefits. Moreover, some insti- high-level government jobs, which often sends tutional reforms may have immediate positive effects for a (symbolic) signal to interest groups that their main beneficiaries. For example, the introduction of their concerns will be addressed. education vouchers immediately helps the families that can Source: Edwards and Lederman (1998). 29 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER dctonside thc eltonsuechoic ewe ie baopeiinkingbat ser- cost'bankin beitor fofori~nst wituioareforsmy o s~e slome reforms thatepnhanetheabltyo ofat ownedivbdalks ooraiza-e etyo reform+.I strategiv'r~tes incluingth seo opein the onen- tionsto s lieraiztion a"lowine" foraegies'n pownvsipade propr'sati~borin mechnisms iscusse above. egltonai incentiresto thirvestments ian the secorliberallyaio popuar Auevs aenothe motn crdiberanthisonthxt.hsbenrie Ideed,chia (egultio and empoerv ment (vpose)t~' a re good reForernde haechsnd todrk91 pem hat uwncerftoainty ebout i'A'3~~~~~~~~~~~~~~3 obn; e goenen eidst ipeet l iv 0 bnsdreiybyoein te apcs ft2 i }'o utms akadeion aottersqech co 2i'~y'terfr. prces sine eiDPolcy ro tdal rvriain ners aelbeaiainaloigvds:acs ochae oeg financing,w Sorncs- forei~gn oieiinadonrhpi th doesi bank s rmtn foregncpeiin 'hdoisice- Pi eto-a i--.ndtesrnteigoeuatio an spr mIntiseethcaital 'cciiribrlztoii viii ashduld b done rort heiealizatin ofcr s frmo idiettt8 @ copnain ehtim ltcme 'bre~fnna evcst vi rvkn iaca ae h fomerypiieedm;'broesf h crss eoner a elz ta l iereom eei ices niteetrts(n theW trad liheaiaio tha ittrilwefe in thej lon ru e hymyfc oic ayLCcutishveiptei ttesm iime). aan fisc al constrainsithimlettinfthe'fv notnaey,ti early. opnn of th capital aco usme imecoportanto incaseteres grousutatibenefThed fnrom-h hadneidpoflrm thicesubsInized crsedwite stheme aiigainth arctifiian low ratesumof thoie preidu systetitConsequentlyh ercostsnndbheefit of aninstitutiona 'cformpsato mecanipsesome reformes mhay denacidte airsty tof condiducts the orivatiza-o etha ofe rhefr stpprategittes including~ theue patof thesrnen 'tiors domese "exit tries storaptceaper foprevige prnanpia saion otemehnisesehv d scusged te thabowavoeaoi 'marentivs) and toei pospoenth"cae ipolemittionaolth foplr. Angothertm incoritency probiemation thncat has eenro rise Ineied, crhoipe refor) and tempoearmet(oice arefsr unigatoommd to thranezn Roenink o191i theaoetc sctortaint theu laer'dteTe im icosstnd pob~ aiss n hs onex o ntrntina egtitins ore30il INSTITUTIONAL REFORM WHY AND HOW the benefits from reforms may lead voters (or individuals, meet the criteria for effective interest aggregation. or even interest groups) to reject proposals for reform. Such Mainwaring and Scully (1995) suggest four criteria for uncertainty may arise when various reforms are introduced "institutionalized" party systems. First, there must be together (thus making it more difficult for individuals to some stability in the rules and nature of inter-party com- assess the net benefits from reforms) or when there is petition; parties should not simply appear and then just as imperfect information regarding the details and conse- quickly evaporate. Second, the major parties should have quences of the policies. It has already been pointed out that stable roots in society if they are to structure political pref- latent interest groups may be the result of lack of informa- erences over time. Third, the major political parties must tion, especially regarding the technical details and distrib- accord legitimacy to the electoral process, with the expec- utive consequences of policy or institutional reforms. These tation that elections will be the primary route to govern- considerations raise the need to couple reform proposals ing. Fourth, party organizations matter; parties are not with public information campaigns. Nonetheless, reform simply the vehicles of ambitious politicians but acquire an "bundling" may be necessary to offer "something for independent status and value of their own. everybody." This may be needed to raise political support Employing these criteria, the authors note sharp differ- or reduce opposition to specific institutional reforms (Tom- ences among the party systems of countries in the region. masi and Velasco 1996). While Chile, Costa Rica, and Uruguay, for example, appear to have party systems that are relatively well-institutional- The Intermediation of Societal Interests ized, Bolivia, Brazil, Ecuador, and Peru do not. Other So far the discussion has emphasized the nature and interests countries in the region fall somewhere in between. of societal actors, whether and how they are likely to orga- Brazil is often taken as an example of party fragmenta- nize collectively around their interests, whether they are tion. One recent analysis concludes that, in Brazil, "party potential winners or losers from the process of institutional leaders have little control over their members, and many, reform, and how their support can be garnered through var- perhaps most, deputies spend the bulk of their time ious compensation mechanisms and related actions. But arranging jobs and pork-barrel projects for their con- whether pro-change or pro-reform coalitions can be built is stituents. Parties in Brazil rarely organize around national- not only a matter of the societal interests at stake. level questions" (Ames 1998, p. 4). Brazil, however, is cer- Another important factor has to do with the intermedi- tainly not the only country in the region in which political ating role of political leaders and political organizations, parties do not appear to perform the interest aggregation such as political parties. Such intermediaries "aggregate" function particularly well. Why not? the interests collectively articulated by the actors in civil The reasons given in the literature are numerous: society. This is a crucial political function because it facili- * Some observers place at least part of the blame on tates the harmonization of the often contradictory interests electoral systems, particularly on some kinds of pro- of groups in civil society and thus facilitates the adoption portional representation that appear to encourage a and implementation of public policies that command sup- lack of discipline on the part of party politicians. port beyond the narrow political interests of any one group. * Others focus on internal party decision-making struc- In the literature on pluralist political systems, political par- tures and processes that they say contribute to party ties are seen as the essential "interest aggregators." Political oligarchies. (The advent of internal party primaries in parties also provide ideological frameworks and lenses some countries has, however, partially countered this through which their supporters can interpret and assess the argument.) numerous and complex policy issues without having to pos- * Others see the problem more rooted in long-standing sess detailed knowledge on each and every one. cultural and historical persuasions, such as the alleged Political parties have attracted considerable attention tendency toward "personalism" as a basis for political from analysts of politics in Latin America and the affiliation. Caribbean. Although broad, region-wide generalizations * Some political parties have been "captured" by rela- are obviously difficult to make, it is apparent that many tively narrow class or sectoral interests, making them political parties and party systems in the region fail to in effect little more than glorified interest groups. 31 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER * Some focus on the previously discussed history of alone may shorten the other's term, except in extraordinary clientelism in the politics of the region; patron-client circumstances. relationships place a premium on personal ties and Noting that the region is characterized by presidential political kinship, and correspondingly downplay the systems of government is not the same, however, as argu- role of formal party organizations. ing that it is marked by "presidential dominance." In fact, Noting the apparent shortcomings in many political there are wide variations in the powers of the president vis- parties, some have argued that countries in the region are a-vis the legislative branch. There are, for example, signif- facing a "crisis of representation" (Domfnguez 1997). This icant differences in the extent to which the presidents have may be an exaggeration; it is difficult to argue that there is veto power, can promulgate decrees without resorring to less representation in Latin America and the Caribbean the legislature, and retain the exclusive power of introduc- today than existed in the heyday of authoritarian and mil- ing legislation, at least in some key policy domains. To itary regimes. Nevertheless, it is an important question complicate matters, such powers can differ according to the whether the formal democratization of politics in the subject of proposed changes. For example, even presidents region has been accompanied by "effective" advances in with generally strong veto powers may lack such powers representation, particularly of the lower socioeconomic when it comes to budgetary matters, and even presidents straza. The role of political parties is important in this with generally broad decree powers may lack such powers debate; in their absence or their reduced effectiveness, it with regard to particular issues. In addition, there are vari- would become more difficult for the citizenry's interests ations in the procedures for amending the constitution in and demands to be aggregated, processed, and ultimately the different countries of the region. dealt with by governmental leaders. These differences are captured in Table 2.2, which pro- vides a summary of presidential powers over legislation in Governmental Institutions in the Change selected countries. A "proactive" president is one who can Process establish-or attempt to establish-a new status quo. A The formal institutions of government also affect the pos- "reactive" president is one who only can attempt to defend sibilities for institutional change.1 For example, constitu- the status quo against legislative attempts to change it. tions and other formal rules can determine the nature of The relative power of the president, however, is not only checks and balances between the different parts of govern- determined by the formal allocation of authority as spelled ment as well as the structure of the political system in out in the constitution. What also matters greatly is the terms of the number of political parties and the like, and president's partisan support in the legislature. This also they may determine which specific political actors control varies substantially across countries, as shown in Table 2.3. institutional reforms. According to G. Cox and McCub- The mean level of presidential support in the legislature, bins (1996), "political actors' incentives are significantly whether on the part of the president's own party or on the influenced by the rules regulating electoral competition, part of a coalition of parties supporting him, has been low while their capabilities are determined jointly by their historically in Bolivia, Brazil, Chile, and Ecuador. Presi- electoral success and the constitutionally stipulated powers dents have been far from commanding majority support in of the various governmental posts that are at stake (either the legislature in these countries. Moreover, even where directly or indirectly) in elections." support for the president in the legislature hovers near a Chapters 6 and 7 of this volume deal with two impor- majority, such support is frequently vitiated by a lack of tant governmental organizations-the judiciary and the discipline in the president's own party or the coalition of bureaucracy. Here the focus is on the executive and legisla- parties supporting him. While the nature of social cleav- tive branches of government in Latin American and ages and other historical factors undoubtedly plaiys an Caribbean countries that have presidential (not parliamen- important role in determining the extent of party prolifer- tary) political systems. Such systems have two defining ation within the legislature, research has also demonstrated characteristics: (a) The chief executive is popularly elected, the importance of such factors as concurrent elections for and (b) the terms of office of the president and the legisla- the presidency and the legislature, various kinds of propor- ture are fixed. Neither the executive nor the legislature tional representation, and characteristics of the institutions 32 INSTITUTIONAL REFORM- WHY AND HOW TABLE 2.2 nied by at least moderately strong support in the legisla- Summary of Constellations of Presidential Powers over ture is Argentina. In several other cases, presidents with Legislation in Latin American Constitutions strong constitutional powers have had relatively low sup- PRESIDENT'S port in the legislature, rendering decisive governmental CONSTITUTIONAL action difficult. Interestingly, many countries' presidents LEGISLATIVE CONFIGURATION actuall have relativel low levels of formal constitutional AUTHORITY OF POWERS EXAMPLES y y Potentially dominant Decree, strong veto, Chile 1980-89' powers and their strength derives principally from the par- exclusive introduction Colombia 1968-91 tisan support they enjoy in the legislature. This is the case, Decree, strong veto Argentinab for example, with the president of Mexico, traditionally Ecuador"b considered perhaps the strongest chief executive in the Proactive Decree, weak veto, Brazil 1988b region. From a comparative perspective, the Mexican pres- exclusive introduction Colombia 1991" ident has relatively limited formal powers but has been Peru 1993a immensely fortified by his leadership of the Partido Rev- Decree, weak veto Peru 1979' olucionario Institucional (PRI), the formerly hegemonic Reactive Strong veto, exclusive Brazil 1946 political party in the Mexican system. If future political introduction Chile pre-1973 developments in Mexico lead to more power-sharing with Uruguay other political parties, then it is likely that Mexican presi- Strong veto Bolivia dential power will diminish due to the formal constitu- Dominican Republic El Salvador tional provisions. More broadly, Table 2.4 demonstrates Panama that in many countries of the region the vaunted power of Potentially marginal No veto (on annual Costa Ricab the president is closely tied to electoral outcomes and par- appropriation bills) Honduras' tisan configurations in the legislature. Mexicob Attention to executive-legislative relations is arguably Nicaragua Paraguay even more important in the current institutional context in Venezuela the region. The democratization of regional political sys- Notes: Decree-the president may establish new law without prior congressional autho- rtems has, in theory at least, increased the importance of the rization (therefore not including decrees of a regulatory nature). String sete-override legislative function. Thus, issues that have long been requires more than a majority of all members. Excl/sive ixtrodactiaoncertain important bills in addition to the budget must be initiated by the president, or congress may not prominent in the study of executive-legislative relations in increase items of expenditure in budget proposed by the president, a. Decree restricted primarily to fiscal matters. the advanced industrial democracies-such as the internal b. Different veto provisions apply on different types of bills. The Colombian president decision-making structures and processes of the legislature, has strong veto powers over the budget but weak power over other forms of legisla- tion. No other presidents have veto power over budgets. Veto powers over other forms the adequacy of legislative staffs and of legislative access to of legislation are strong in Costa Rica, Hondums, and Mexico, and almost absolute in Ecuadormation sources, and the nature of legisla- c. The Ecuadorian president's veto may not be overridden if he or she vetoes the entire tive oversight of executive decision-making-may take on text, although Congress may request a referendum on the bill; if the president objects only to specified parts of a bill, the veto (of the whole bill) may be overridden by a added significance in Latin America and the Caribbean as two-thirds veto. well. Soarre: Mainwaring and Shugart (1997, Chapter 1). Chapter 7 suggests that the nature of executive-legisla- tive relations may be an important element of achieving (rules) of the legislative branch (Mainwaring and Scully viable bureaucratic reform. Other studies have demon- 1995; Mainwaring and Shugart 1997). strated their importance for a considerable array of policy A recognition of both the formal constitutional and par- domains. A fruitful line of inquiry has been one that tisan powers of the president leads to a more complete emphasizes the alignment of incentive structures in such a understanding of the president's ability to effect change, way that politicians' interests in short-term political sur- including institutional change. Table 2.4 combines a con- vival can be harmonized with their desire to promote more sideration of each kind of power and produces some encompassing social or political objectives (such as institu- instructive results. For example, the only country in which tional reform). Geddes (1994), for example, has demon- the president has strong constitutional powers accompa- strated that the adoption of more meritocratic presidential 33 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE 2.3 Presidents' Parties' Mean Share of Congressional Seats in Latin America (percentages) PRESIDENT'S PARTY PRESIDENT'S COALITION COUNTRY AND PERIOD YEARS NO. OF ELECTIONS LOWER CHAMBER UPPER CHAMBER LOWER CHAMBER UPPER CHIAMBER Argentina 1983-93 6 48.3 52 49.1 52 Bolivia 1980-93 4 33.9 47.2 Brazil la 1945-50 3 34.8 44 52.2 48.6 lb 1954-62 4 26 26 44.8 52 Ila 1985-90 4a 26.9 25.6 37 31.8 lIb 1994 1 12.1 13.6 35.4 42 Chile 1 1932-73 18 23.3 20.8 41.6 41.2 11 1989-93 2 31.7 28.3b 58.3 46.3b Colombia 1945-49, 1974-94 11 55.2 56.3 55.2 56.3 Costa Rica 1953-94 11 49.6 - 51.8 -- Dominican 69.1 Republic 1962,1966-90 8 55.6 - - Ecuador 1978-94 7 22 El Salvador 1 1985-91 4 47.5 - - II 1994 1 46.4 - Honduras 1981-93 4 54.2 - 54.2 -- Mexico 1982 91 4 65.8 95.8 65.8 95.8 Nicaragua 1984-90 2 65.4 - Paraguay 1993 1 47.5 44.4 - - Peru 42.2' 1980 1 54.4 43.3 54.4 43.3 1984-90,1995 3 47.1 40c 47.1 41.7' Uruguay 1942-71, 1984-94 11 45.6 43.8 Venezuela 1958-93 8 41.1 47.4 43.2 47.7 Notes: a. Includes the indirect presidential election of 1985. b. For Chile 11, appointed senators were included in calculating the president's share of Senate sears. If one rakes only elected seats, the percentage increases to 34.2% for the president's party and 56.6% for the coalition. c. Does not apply to 1995; Peru moved to a unicameral legislature with rhe constitution of 1993. Source: Mainwaring and Shugart (1997, Chapter 11). appointment strategies to bureaucratic posts in many change as the factor to be explained (i.e., the "dependent countries of the region has been crucially affected by some variable" in social science terminology), it is clear that fundamental features of the political landscape, including there are various explanatory levels-broadly speaking, the size of the president's party, the discipline of party societal, intermediary, and formal/governmental-that members, and whether the president is attempting to facilitate understanding of why and how such change build a political machine while in office. Such studies occurs. But a systematic appreciation of the precise rela- emphasize that the harmonization of short-term political tionships among these levels, or of their analytically inde- survival goals with the broader "public interest" in institu- pendent contributions to the explanation of varying out- tional reform is likely to be at the heart of reforming insti- comes, remains a daunting task. tutions through the mechanisms of the political process. Nevertheless, this survey of some of the key emphases in the study of the political economy of institutional change Some Guidelines for Reform has highlighted a range of relevant factors that needl to be The discussion in this chapter indicates that a full-blown taken into account by those-be they societal actors, gov- theory of the political economy of institutional change still ernments, or international agencies-who may be inter- remains elusive, despite the contributions of economists ested in promoting reforms aimed at improving the qual- and other social scientists. In focusing upon institutional ity of domestic institutions. Taken together, they 34 INSTITUTIONAL REFORM: WHY AND HOW TABLE 2 4 Relationship Between Presidents' Constitutional and Partisan Powers in Latin America PRESIDENT'S PARTISAN POWERS CONSTITUTIONAL POWERS OVER LEGISLATION VERY LOW MEDIUM LOW MEDIUM HIGH VERY HIGH Potentially dominant Chile, 1989 Colombia, 1968 Argentina Ecuador Proactive Brazil, 1988 Colombia, 1991 Peru Reactive Brazil, 1946 Bolivia El Salvador Dominican Republic Chile, 1925 Uruguay Potentially marginal Costa Rica Honduras Paraguay, 1991 Mexico Venezuela Nicaragua Source: Mainwaring and Shugart (1997, Chapter 11). constitute a few basic guidelines for institutional reformers. international treaties or similar commitments, such as Borrowing from central concepts in institutional economics the signing of summit declarations. and political economy, the guidelines are as follows: * Empowering the beneficiaries is good policy and smart poli- * Pay systematic attention to the nature of prospective winners tics. Graham and Naim (1998) have suggested that and losers from instituitional reform (e.g., trade unions in institutional reforms are more likely to be supported the formal sector with respect to labor market reform, if the potential beneficiaries participate in the design diverse rural interests regarding "market-based" land of the new institutions. One way to do this is through reform, teachers' unions regarding education reform, "voice," in Hirschman's (1970) terminology, which is various subcategories of the financial sector regarding a feedback device by which principals exert control on financial reforms, etc.). Be aware of the main cleavages the decision-making process of their agents and orga- that appear to separate such winners and losers- nizations. As is discussed in greater detail in Chapter sometimes, for example, these may have a regional 5, one way of improving the performance of schools in dimension, sometimes a sectoral base, sometimes a the region is to provide greater participation to par- base in social classes linked directly to their position in ents in the schools' management. Empowerment (or the productive structure. In any event, monitor closely voice) then becomes a means to ensuring that the the nature and intensity of support or opposition as school acts according to the interests of the house- manifested by public opinion polls, proclamations of holds, which is particularly important when the interest groups, public demonstrations, and the like. "exit" or choice strategy is not available. It is also safe * Based on well-informed assessments, attempt to craft to argue that the beneficiaries will welcome such poli- compensation schemes that are politically viable and thus cies. In the context of financial reforms, protecting credible. In addition, it may be necessary to make minority shareholders' rights is also a voice strategy promises about future compensation schemes, which for reform, which should not face severe political may be key ingredients for both the effectiveness of obstacles on the part of public opinion. The decen- the reforms and for their political sustainability. One tralization of state functions to local governments is way to enhance the credibility of compensation another example where local communities can poten- schemes is to raise the costs of "exit" from commit- tially gain voice over public bureaucracies in the con- ments by future governments through adherence to text of democratic (choice) politics. 35 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Providing choice to beneficiaries is also good policy and smart enclaves within the formal structure of government, politics. Some of the assertions stated above about etc.). Such an understanding, simple as it seems but empowerment also apply to the provision of "choice" important as it is, could facilitate ex ante calculations or "exit" strategies for beneficiaries. If principals are of the feasibility of institutional reform-of labor not satisfied with the quality of services provided by market reform in Argentina, to take but one example, their agents or organizations, they can desert and look or of administrative and social security reform in for the services elsewhere. The threat of exit is a com- Brazil, to take another. With a clear understanding of plement to the provision of voice or empowerment the constitutional forces in play, and of their close strategies for institutional reforms, since the threat of relationship to partisan forces particularly in the leg- exit strengthens the voice strategy. In the case of edu- islature, think creatively about how to exploit the cation, for example, the use of education vouchers may opportunities and overcome the constraints. be a good (and politically popular) complement to * Focusing on reforming incentive structures is good policy and empowerment strategies, where the schools have to smart politics. Perhaps the most difficult reforms to respond to the demands of parents, especially if pub- undertake from a political standpoint are those that lic-education subsidies for schools are linked to chil- aim to change or reduce the size of public ernploy- dren's enrollment (as in the case of Chile, for example). ment. This is the case for two reasons: First, the losers •Public-information campaigns should be part and parcel of have human faces that become the symbols of the reform efforts. A frequent issue that emerges out of the costs of such reforms, and second, the public jobs are political-economy perspective is that latent interest often part of the political game by which supporters groups are politically inactive as a consequence, in of certain leaders get rewarded. Consequently, reforms part, of the costs of collecting information about the of the civil service and societal organizations (e.g., potential costs and benefits of particular reforms. If schools) should focus on the incentive structures reforms are viewed as a collective good, however, there rather than on changing personnel or installing the is clear justification for reformers to spend resources latest technology. This approach may not only be explaining the details and likely consequences of pro- more politically viable than wholesale changes in per- posed policy initiatives. This role is particularly sonnel, but may also be the most appropriate important in the context of democratic regimes, approach from a technical standpoint, based on the where voters have political voice and exit strategies emphasis placed on incentives by the new institu- available but may not use them to defend their inter- tional economics. ests effectively. * Pay careful attention to the political support for prospective In sum, the analysis in this chapter indicates that reforms at the intermediate level, particularly among key reforming institutions in the countries of Latin America political leaders and political party organizations. Assess and the Caribbean is far from an impossibility. To the con- the possibilities for "deals" and tradeoffs among them. trary: The increased demand for more effective institutions Be particularly aware of the "political cycle"-i.e., in the region can be matched by commensurate supply. how windows of opportunity for institutional change What is required from would-be institutional reformers is might open (or close), depending upon impending sustained political commitment and carefully tailored elections. strategies for putting such commitment into operation in *Have a clear understanding of the constitutional-i.e., for- specific sectors. Policy analysts, using some of the tools of mal/governmental-facilitators and obstacles to institu- the trade discussed in this chapter, have an important role tional change and reform (e.g., the possibilities of intro- to play in helping reformers craft such strategies. The com- ducing reforms via presidential decree, the realistic bination of historical trends that have raised the demand scope for reform in situations in which a strong leg- for institutional reforms, and the fashioning of astute islative branch shares important powers with the pres- reform strategies that are sound from both technical and ident, the potential for creating autonomous or semi- political points of view, present Latin American and autonomous agencies as relatively non-political Caribbean countries with a historic opportunity to close 36 INSTITUTIONAL REFORM. WHY AND HOW the "institution gap" that currently afflicts the region-a Note gap that both threatens the consolidation of the vital 1. The discussion in this section draws extensively from Main- reforms already undertaken and impedes the implementa- waring and Shugart (1997), especially Chapters I and 11. tion of the "second generation" reforms that are crucial. 37 PART 2 Institutional Reform in MIarkets: The Case of the Financial Sector CHAPTER 3 Institutions, Governance, and Incentives in Banking: Safety Net Arrangements A CONSENSUS HAS DEVELOPED IN THE LAST FEW DECADES THAT WHEN IT COMES TO THE performance of the financial sector, institutions play an essential role-especially because of the relevance for this sector of information issues, agency problems, transaction costs, and property rights. While there is general agreement on the importance of having an ade- quate institutional framework for well-functioning financial markets, how to achieve it remains a challenge. The general purpose of this chapter is to illustrate incentives and insti- tutional issues in banking systems by reference to broadly defined safety-net arrangements-including lender-of-last-resort facilities, deposit insurance, capital requirements, supervision, and exit policies- because safety nets tend to be, de facto or de jure, an important component of financial systems.1 We endeavor to highlight those aspects that are of greater relevance to LAC countries. A useful framework to tackle the key issues systematically is that provided by agency theory and the theory of incomplete contracts. Information and incentives in banking, the first section in this chapter, deals with the conceptual aspects of asymmetric information and agency issues, illustrating them principally by reference to loan and deposit contracts. Banking regulation and safety nets starts reviewing the information and incentives issues that motivate regulation. Safety nets have evolved over time, but not in a linear way; in any one country the existing safety-net arrangement has ing argument is that even if there is no ex ante provision been shaped by that country's history, including the nature for safety nets, these tend to surface ex post in one form or of financial crises and regulatory pressures. Thus the rest of another-and frequently in an improvised manner. Given the section takes a historical perspective, first focusing on that safety-net arrangements seem to be here to stay, and instances of prudential regulation without ex ante safety because of the complexities they involve, the careful design nets, and then considering illustrations of pressures for of ex ante safety nets is crucial. Although extensive safety nets even before the introduction of central banks. A research has been done on capital requirements and moni- discussion of central banks and the provision of safety nets toring (including prudential regulation and official super- (both ex ante and ex post) follows, with special attention vision as well as market monitoring), closure mechanisms given to the shift of risks to the government that results have had relatively less attention in the academic litera- from safety nets and to the trade-offs involved between this ture. The safety-net tripod: capital, mnonitoring, and closure shift of risks and financial deepening. reviews these three crucial aspects of safety nets, emphasiz- Throughout this chapter, the structure of incentives ing the complexities involved in closure. that emerge from safety nets in banking systems is consid- Fire protection and fire fighting: designing exit policies dis- ered to depend principally on the strength of a tripod con- cusses elements of effective exit policies, drawing on an sisting of capital, monitoring, and closure. The underly- analogy with good fire protection and fire-fighting policies. 41 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER By this analogy, effective bank exit policies need to be sup- agency relationships: lender-borrower and bank-depositor.3 ported by preventive regulation, monitoring, and capital Banks act as delegated monitors of firms when they engage requirements (building codes and inspections), early detec- in their lending activities: Numerous depositors, instead of tion mechanisms and prompt corrective rules (smoke monitoring each one of the borrowing firms, delegate such alarms, fire extinguishers). They should include carefully monitoring to a bank, whose ability to stay in business designed bank intervention and bank failure resolution largely depends on its capacity to exploit economies of capacities (fire-escape routes, fire drills, and fire depart- scale because of the natural monopoly aspect of informa- ments) so as to avoid disorderly evacuation and unnecessary tion-generation. "Know your client"-the basic rule of panic if the worst comes to pass. Insolvent banks and unfit good banking-presupposes that banks are able to gather bankers should not be allowed to remain in operation and accumulate sufficient information, which, on the one ("condemned" buildings are not suitable to be inhabited). hand, enables them to assess the risk of debtors and, on the Bank resolution is a broad term that goes beyond plain other, limits debtors' ability to shift easily from one bank vanilla liquidation and closure mechanisms. We describe to another. alternatives and explicitly discuss deposit insurance, focus- On the liability side, banks must provide liquidity for ing not only on its role in curbing panics and contagion depositors. Depositors do not always know when they will risks, but mainly on its role as a potentially important tool need cash, although not all depositors have liquidity needs for bank-failure resolution. at the same time. The bank must provide cash when it is The adequate design for safety nets needs to deal with needed, while channeling funds from a pool of typically the day-to-day issues related to individual banks' idiosyn- short-term and small deposits to longer-term and larger- cratic risks as well as with systemic risk. A commonly scale investments. observed occurrence in banking systems is that the rules of the game regarding ex ante risk-shifting toward taxpayers Loan Contracts tend to change or be ignored once the going gets tough. To lend money, banks write a loan contract, which, based Safety nets, systemic risk, and catastrophe insurance explores sys- as it is on information possessed only by the bank, is not temic risk and the difficult decision of rules versus discre- easy to trade in markets.4 Writing a loan contract is the tion that a government faces in the event of a potential sys- formal equivalent of writing a put option on the assets of temic crisis. If difficult theoretically, this issue appears all the debtor firm: The firm pays if it can and defaults (deliv- the more complex in practice. We review a few lessons ers, or puts the assets to the creditor) if it cannot pay. Given from experience and discuss alternative ways to address this asymmetric information, the problem for a bank is to problem. The chapter ends with a summary and conclud- determine the ex ante profitability of the projects to be car- ing remarks. ried out by borrowers. To the extent that the bank (princi- pal) cannot distinguish between projects, it would charge a Information and Incentives in Banking "lemons" premium across all projects, taking into account Information problems are particularly important in finan- the probability of mistakenly funding a bad project.5 As a cial markets because financial transactions take place over result, borrowers (agents) with good projects havve an time. Financial markets deal with a promise to pay in the incentive to look elsewhere for funding, leaving the bank future, which involves uncertainty and risks. Information with a worse pool of borrowers from which to choose. If problems arise when the agents involved (investors or this adverse selection or hidden information problem is depositors, financial institutions, and borrowers) have severe enough, lending may not take place. An increase in asymmetries in information about the likelihood of com- the lending interest rate tends to increase adverse selection pliance with the promises kept-that is, when one of the (Stiglitz and Weiss 1981). parties has less, or less accurate, information than the other In addition to, and independently of, the adverse selec- party.2 tion problem, asymmetric information creates moral haz- Banks have a dual function, as delegated monitors on ard, as borrowers have an incentive to take risks aimed at the asset side and liquidity providers on the liability side capturing potential gains under good states of the world, (Dewatripont and Tirole 1994), which gives rise to two while not losing more than the capital they have invested 42 INSTITUTIONAL REFORM IN MARKETS. THE CASE OF THE FINANCIAL SECTOR in the project in bad states of the world. To this end bor- tract is a restriction on the amount of debt that a firm can rowers could engage in hidden actions that increase the have, the collateral required, and an injunction that loans probability of default after the loan has been made. from other lenders be reported to the bank. These restric- tions are a form of capital requirement, since they limit the Deposit Contracts leverage of the firm.7 Loan contracts normally also require If a loan contract is the formal equivalent of a put option some form of monitoring-for instance, to ensure that the on a debtor's assets, a deposit contract is the formal equiv- debtor firm is audited by an independent auditor, that the alent of a put option on a bank's assets. Banks pay a return bank has a representative on the firm's board of directors, on deposits that includes an implicit premium to cover the or that debtors that share liabilities within a group moni- probability of default (the premium on the implicit put tor each other's behavior. Finally, many loan contracts are option). Asymmetric information between depositors and short term but rolled over so that the bank can force the the bank's owners and managers once again gives rise to firm to default and seize assets (closure) if it becomes clear selection and moral hazard problems. If depositors cannot that the firm will not be able to pay back the loan. This distinguish between good and bad bankers, the adverse provision limits the bank's losses from actions the firm selection problem will create a "lemons" premium that could take to run down the value of its assets once it ceases penalizes good banks and favors the entry and expansion of to be a viable enterprise. bad banks. Once depositors give their money to a bank, Capital requirements, monitoring provisions, and moral hazard arises, as the bank has an incentive to closure mechanisms in loan contracts lower the cost of undertake hidden actions that help bank owners at the intermediation by allowing the bank to sort out prudent expense of depositors. borrowers from risky borrowers, and by influencing the Moral hazard implies that bankers have incentives to incentives of prudent borrowers once loans have been undertake riskier lending with the expectation of captur- made. Successful use of these measures results in financial ing the upside potential, while keeping the downside lim- deepening: more projects are undertaken, and their adop- ited to the amount of bank owners' net worth invested in tion does not depend on self-financing. At the same time, the bank. In addition, they may even misrepresent earnings bank financing of these projects results in financial risk and engage in old-fashioned looting. Akerlof and Romer shifting between private agents: Firms with limited lia- (1993) argue that moral hazard may induce bank managers bility bear only part of the downside risk for their projects, and owners of undercapitalized and inadequately moni- with banks assuming the remainder of the risk. tored banks not only to "go for broke" (i.e., to bet for the Even if introduced spontaneously by private agents resurrection of the bank by undertaking riskier activities, without inducement from the government, capital, moni- which may not hurt depositors or taxpayers if the low-prob- toring, and closure would mitigate selection and moral ability bet happens to pay off) but also to "go broke" (i.e., hazard problems in loan contracts only to the extent that a to defraud and loot the bank, with depositors and/or tax- society and its government are capable of providing a min- payers always losing in the process). imally reliable set of public goods and services that facili- tate contract enforcement at reasonable costs. This Mitigating Agency Problems would imply that property rights are well-defined, bank- As discussed, there are two common ways in which a ruptcy rules are clearly specified in the law and custom, debtor firm can hurt the creditor bank: misrepresentation and judicial systems can be counted upon to facilitate col- of assets (leading to a problem of adverse selection) and lateral repossession and execution of guarantees, as well as misuse of those assets once the loan is approved (the result to deter and punish breaches of contract.8 Governments of moral hazard). Both banks and firms find it advanta- have a significant role to play in strengthening the institu- geous to create non-price solutions-that is, institutions- tional framework needed for loan-contract enforcement, to mitigate these problems and, thus, enable the credit sys- although the required effort to succeed in this is not inde- tem to operate.6 Screening devices, debtor information pendent of a country's legal heritage. La Porta et al. centers (credit bureaus) and loan covenants play a central (1997a, 1998a) show that covenants establishing the rights role in this connection. A standard covenant in a loan con- of creditors over debtors are weaker and less enforceable in 43 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER countries whose legal heritage can be traced to the civil- banks in loan contracts, and each depositor has an incentive law tradition, particularly the Napoleonic Code, than in to free ride on the information-generation efforts of other countries with common-law tradition (see Chapter 4).9 depositors, enjoying the benefits without incurring the Although a parallel can be made between the informa- costs. A market failure thus arises that provides a basis for tion asymmetries involved in loan contracts and deposit official regulation. contracts, the fact is that the associated selection and moral hazard problems are much more difficult to mitigate Banking Regulation and Safety Nets through voluntary actions by private agents in the case of The rationale for banking regulation is the need to address deposit contracts. To be sure, unregulated banks have an externalities that the market does not adequately deal incentive to create mechanisms to mitigate the effects of with, either through the price mechanism or via non-price asymmetric information, given that the "lemons" problem arrangements set up voluntarily by private agents.'1 In may lead to equilibrium where there are no bank particular, regulation addresses the market failures that deposits.'0 In effect, as a partial voluntary solution, a bank arise from information asymmetries. As with all market may be willing to submit to an external audit, in order to failures, those stemming from asymmetric information in create publicly available and credible information about the banking system create a wedge between private and assets, reserves, loans to related parties, and other data that social interests: The "invisible hand" of the market fails to signal the quality of the bank. In addition, banks may issue produce a convergence of the two. Because the key market demandable debt, such as bank notes or demand deposits, failure to be dealt with is the one that constrains monitor- rather than time deposits. In the absence of lender-of-last- ing of banks by depositors, there is a case for a govern- resort facilities (see below) or credible deposit insurance, mental supervisory agency to monitor banks as the repre- demandable debt creates a sequential service constraint of sentative of all depositors. first-come, first-served (Mishkin 1991) that makes it This is the representation hypothesis introduced by worthwhile for some depositors to invest resources in mon- Dewatripont and Tirole (1994). The main argument is that itoring the bank. Demandable debt reduces the problem of to ensure adequate monitoring, control over banks should free riders, since depositors that rely on others to monitor shift from soft claim holders (equity-holders) to tough ones are less apt to recover their funds if there is a run on the (debt-holders) in case of mediocre performance. This is bank. Depositors would vote with their feet if the bank is because stockholders will tend to "gamble for resurrection" not viable, forcing its closure. during difficult times and will thus not be tough monitors These voluntary non-price solutions tend to be insuffi- when it is needed the most. Nevertheless, debt-holders cient in the case of deposit contracts because the free-rider (depositors), especially small ones, typically do not moni- problem is more severe, as implied by the theory of collec- tor banks, as explained above. Thus, governments can tive action (Olson 1965). This theory suggests that, due to monitor banks as representatives of depositors, with signif- the free-rider problem, larger groups, such as those made icant gains to society's welfare. up of numerous depositors, tend to be less effective in The type of government intervention called for is cen- achieving a common objective than smaller and more con- tered on prudential regulation and supervision2-that is, centrated groups, such as those made up of relatively few on the set of norms and monitoring policies and procedures banks. Once a depositor invests in the costly activity of that encourage greater "prudence" in risk-assessment, risk- gathering information on a bank, such information cannot taking, and risk-management by banks and their clients. be hidden from other depositors in a sustainable manner. The goal of prudential regulation is to promote financial The very actions that would result from such information deepening while complementing market-originated moni- (e.g., a run on the bank) immediately discloses the infor- toring and enhancing overall governance in the system. mation. Also, individual depositors usually have relatively The appropriate design and implementation of prudential small claims on banks, so it normally does not pay for them regulation and supervision is a complex subject, especially to engage in costly information-gathering activities. because of the challenges posed by rapid financial ininova- Hence, no depositor has an incentive to enforce covenants tion and other changing market realities. Sound prudential in deposit contracts that mimic the covenants imposed by regulation places emphasis on complementing and sup- 44 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR porting market forces (Caprio and Klingenbiel 1996). bates moral hazard problems, which, if sufficiently severe, Optimal government regulation should mimic the mecha- increase systemic instability, thereby undermining the nisms of a perfect market (Klein and Leffler 1981) with a safety net's raison d'&tre (Calomiris 1997). view to creating incentive compatibility, i.e., such align- The design of safety nets is also complicated by the ment of incentives facing the various agents that removes recurrent emergence of ex post safety nets in the midst of the wedge between private and social benefits. As indicated crises, the "too big to fail" phenomenon, and the relevance in the discussion on loan contracts above, appropriate pru- of aggregate systemic risk-rather than idiosyncratic, dential regulation in the banking system would need to loan-specific risk-for banks in volatile economies such as cover well the three dimensions of capital, monitoring, and those of Latin America. In effect, there is the historical closure (see the safety-net tripod: capital, monitoring, and clo- observation that, to the extent that a formal safety-net sure for more details). arrangement fails to anticipate political and economic As we will illustrate subsequently, prudential regula- pressures during a crisis, an ex post safety net will emerge tion in banking systems needs not provide for an explicit in which risk-shifting is driven by governmental discretion safety net and may not even entail an implicit one. How- rather than rules. Also, even if risk-shifting to taxpayers is ever, the existence of a safety net (explicit or implicit) not excluded or rigorously minimized by ex ante rules of the only presupposes prudential regulation but also heightens game, behavior does tend to be influenced by the rather the need for it. This is because, by definition, safety nets in pervasive perception that an implicit safety net exists to banking are sets of organizations and institutions that cover 100 percent of the liabilities of at least those banks imply risk-shifting toward taxpayers, which gives rise to a considered to be "too big to fail." Finally, with much of the form of implicit social contract whereby the official super- risk faced by banks in Latin America being of the aggre- visor undertakes to maintain tight prudential oversight gate kind (e.g., terms-of-trade shocks), the issue arises as to over banks to control taxpayers' exposure to losses. This the conditions under which deposit insurance can be risk-shifting is nonetheless taken as the price to be paid for thought of as a form of catastrophic insurance, with the enhancing systemic stability and curbing socially costly government having a comparative advantage in handling externalities associated with bank runs, payments system aggregate risk. failure, generalized credit collapse, etc. Although it is fre- The issues and themes mentioned above will be quently argued that the best safety net is one that results explained further in the next subsections through a in market participants' behaving as if the safety net did not selected historical recounting of how safety nets emerged, exist, well-designed bank safety nets should alter behavior even when they were not explicitly stated ex ante and even and deepen financial intermediation by shifting some risk prior to the creation of central banks. to the government.i3 A well-functioning safety net must balance its institu- Prudential Regulation without a Safety Net tional components-lender-of-last-resort facilities, deposit Prior to the development of a bank safety net, banks have insurance, capital requirements, prudential oversight, and more often been subject to a "light" amount of prudential bank-closure (exit) policies-so as to carefully control the regulation rather than no regulation at all. The key similar- risk borne by taxpayers, while achieving the financial deep- ity between no regulation and light regulation is that the ening that results from the strengthened ability of the sys- closure mechanism is generally triggered by bank runs, and tem to withstand such disturbances as contagious bank the government remains outside of the compensation runs. Ultimately, a well-designed safety net would improve process for bank creditors. The key difference is that under social welfare. The proper balance between the safety-net light regulation the government creates laws and norms components becomes more feasible to the extent that the that commit banks to more stringent prudential practices safety-net architecture does not blunt, but rather supports than banks could insure depositors without such regulation. the monitoring, discipline, and governance provided by The main features of "light" regulation can be seen by the market. By the same token, when a safety net's institu- examining free banking legislation in the United States in tional components are poorly designed or inadequately the nineteenth century. After enactment of the 1838 New implemented, then it becomes dysfunctional and exacer- York State Free Bank Law, free bank legislation in most 45 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER states created a form of "narrow banking." States would holdings. Asymmetric information potentially leading to give bank notes to free banks in exchange for U.S. govern- adverse selection, coupled with the lack of a legal mecha- ment bonds or approved state bonds. Free banks would cir- nism to enforce contracts, resulted in a "lemons" p)roblem culate bank notes and, equally importantly, take deposits where no long-term lending took place. to make loans. Free banks were subject to minimum capi- This is a clear example where, even in the lending side, tal requirements and sometimes "double liability" require- there is room for government intervention to strengthen ments (shareholders were personally responsible for an the broader institutional framework affecting the defini- additional amount up to the par value of bank capital in tion of property rights and the enforcement of contracts. In case of the bank's liquidation). The comptroller of a state effect, the Chilean Congress responded in 1856 by creating could, upon petition by depositors, order the detailed a "special" mortgage that gave clear rights to the lender. inspection of a free bank and have the results published. Property registries were set up to make information on Between 1838 and 1863 many free banks failed in the mortgages, sales, and censuses readily available to the pub- United States. Only a small portion of note holders lost lic. In addition, the congress also created a state-sponsored money, but deposit holders lost much more. Despite each mortgage bank, the Caja de Credito Hipotecario, to accom- state's role in the regulation of free banks, state govern- pany the new mortgage law.1" The property registries ments avoided becoming insurers of bank notes or helped to mitigate the asymmetric information problem deposits.14 Thus, free banking and other forms of narrow facing lenders. Equally important, the Caja became a del- banking worked successfully in the nineteenth century egated monitor that could reduce the costs of lending by United States in the sense that they protected the means of adhering to legally mandated collateral requirements, by payment (bank notes) and did not result in government holding a diversified portfolio of loans, and by economies bank rescues in crisis times. Despite the fact that deposi- of scale in monitoring. The new mortgage law gave clear tors did lose deposits during crisis times and banks failed, authority to the Caja to enforce bankruptcy proceedings there appears to have been no effective pressure for a safety (i.e., a closure rule) if a landholder fell sufficiently behind net to be provided by the government. Part of the lack of in making mortgage payments. The reform of mortgage government response was tied to successive U.S. govern- laws and the creation of the Caja solved the "lemons" prob- ments commitment to the gold standard. Part of the lack lem, thereby permitting some risk shifting from landown- of response may also have been associated with the decen- ers to the Caja and to purchasers of the Caja's securities. tralized federal form of government. As these characteris- The accompanying financial deepening propelled the tics are unusual from the perspective of the small open expansion of irrigation and other land improvements by economies of Latin America, a historical example from landholders. Chile in the nineteenth century will help illustrate how the Although the rapid five-year credit expansion permitted borrowers of financial institutions with no explicit safety landowners to undertake capital improvements to their net could generate enough pressure on the government to land, it also exposed them to macroeconomic shocks. Many create an ex post safety net. landowners became unable to make their mortgage pay- ments at the end of the 1850s, when the Californian and Pressures for Bank Safety Nets before Central Banks Australian export markets collapsed. Responding to The earliest documented ex post governmental safety net intense pressure to prevent a generalized foreclosure of in Latin America was created in Chile about 140 years ago. landholdings, the government in 1858 and 1859 clandes- During the early 1850s foreign gold rushes had created tinely channeled to landholders about 2 million pesos of a pressures to increase grain exports from the Central Valley 7 million peso railroad loan that had been financed in the of Chile to California and Australia. But large Chilean London bond market by the government. This governament landowners who wanted to expand their production had no action caused risk-shifting to taxpayers that was not: envi- access to long-term credit, largely because existing mort- sioned by the Caja Hipotecaria's institutional structure. gage laws were poorly defined in legal terms. In addition, Pressures by foreign investors in the railroad loan, as well potential lenders were faced with the lack of reliable infor- as the installation of a new government, led to an attempt mation on the holdings, quality, and legal status of land- to recover the clandestine loans in 1860. The resulting eco- 46 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR nomic contraction of 1861 and 1862, which produced the While excluding an explicit safety net in the ex ante liquidation of a large number of landholdings, was Chile's rules of the game, the new institutional framework pro- first financial crisis and one of its most severe economic vided for another form of guarantee, one that was incon- contractions of the nineteenth century.16 sistent with a safety net based on unchecked access to cen- The example highlights features common to the tral bank liquidity. The legislation created a guarantee dynamics of many implicit financial safety nets. The Caja that domestic currency would be convertible into foreign Hipotecaria and the accompanying legal reforms were cre- exchange at a fixed rate. The sustainability of such guar- ated to promote financial deepening. Prudential lending antee hinged on a strict money-issuance rule-namely, practices were legally mandated, so that risk-shifting to that high-powered money (domestic currency plus a the Caja was carefully controlled in theory. In practice, the bank's deposits in the central bank) could not exceed a cer- Caja expanded its mortgage lending so quickly that it tain multiple (normally less than two) of the central became overly exposed to the risk of a mass default by bank's holdings of gold and foreign exchange. The combi- landowners in response to an external shock. The threat of nation of a fixed exchange rate with the newly instituted massive foreclosures of landholdings initially created pres- environment of oversight of the banking system favored sure for an ex post safety net financed by the government capital inflows and, hence, further fostered financial deep- and then contributed to a severe economic downturn, as ening. However, this combination also entailed risk-shift- liquidation of properties finally took place. ing to the government inasmuch as foreign exchange lia- bilities of banks enjoyed an implicit guarantee on account The Creation of Central Banks without Ex Ante of the fixed peg. Safety Nets The introduction of central banks into Latin America in Ad Hoc Safety Nets and Externalities the mid-1920s had far-reaching consequences for financial Safety nets emerged in Latin America just as in the United deepening and for incentives to shift risk onto the govern- States-as an ad hoc response to the Great Depression. The ments. The Central Bank of Chile-like most other central strain this event put on the newly established institutional banks in Pacific-Rim countries-was the outcome of a mis- arrangements mentioned in the previous section eventually sion led by Edwin Kemmerer, invited to establish a set of led to the abandonment of the exchange-rate peg. As in organizations that would allow Chile to return to the gold current episodes of financial crises, the process followed a standard and eliminate the inflation that began in 1878. fairly typical sequence. Orthodox authorities initially reaf- The Kemmerer Commission made recommendations to firmed the commitment to the peg in the face of capital establish a central bank and a superintendency of banks outflows and massive deposit withdrawals. But at the same that would jointly watch over the financial system. time, the central bank was forced into a progressive depar- The new legislation and institutional structure did not ture from the strict money-issuance rules, as its liquidity directly create an explicit safety net for banks. It did set up assistance to distressed banks grew. Such departure increas- capital requirements, provisions for monitoring, and ingly eroded the viability of the convertible fixed peg, authority to close banks. Thus, the Chilean system in the which eventually was abandoned. Thus, in much of Latin late 1920s operated with capital/asset ratios of 20 to 30 America, ad hoc safety nets emerged ex post, as govern- percent (very high by current standards), monitoring by ments could not guarantee the functioning of the banking the Superindencia de Bancos was rigorous, and closure was system under the rules of the gold-linked fixed exchange enforced-in fact, Chile closed its second largest bank in rate system. The ad hoc bank safety nets created in response the first year of the operation of the new institutions. By to the shock of the Great Depression saved the banks, typ- normal standards, the financial apparatus erected in the ically at the cost of emergency loans from the central bank, mid-1920s was sufficient to render moral hazard and inflation/devaluation spirals that reduced the real value of adverse selection unimportant. At the same time, it facili- deposits, and moratoriums on foreign debt. tated financial deepening by lowering the cost of financial The pressure for ex post safety nets is related to the neg- intermediation, as depositors required a lower premium ative externalities that could result from a failing bank.'" and borrowers paid lower loan rates.17 Bank failures (especially in the case of big banks) can have 47 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER spillover, or "domino" effects in the form of a price race bonds, and low-risk directed credit to highly protected (high interest rates on deposits initiated by weak banks and import-substituting firms. Bank capital was allcwed to followed by the rest in a fight for market share) and a gen- erode.20 Government institutional capacity in monitoring eralized run on deposits. Depositors will rush to withdraw eroded. Thus, the stability of the financial system came at to avoid being the last one in line.'9 Depositors may find a price: From the 1930s to the 1970s financial deepening themselves increasingly unable to distinguish between did not take place in Latin America. individual bank problems and system-wide problems, Such stable but atrophied systems implied a sort of which would induce them to withdraw their funds from safety net consisting of low-yielding loans and invest- otherwise sound banks, leading to failure by self-fulfilling ments, and negative real deposit rates within a context of prophecy. This would be exacerbated by a fire sale of assets capital controls and high trade barriers. A wider, but by distressed banks, as these would further reduce the latent, safety net was also maintained in the form of a banks' net worth. The distribution of losses among bank broader role for the central bank as the lender of last resort. creditors, governed only by the sequential-service con- In the aftermath of the economic disorder caused by the straint on deposit withdrawals, may be socially inefficient. Great Depression, legal frameworks were amended in sev- Individual incentives to seek immediate liquidity would eral Latin American countries to relax or eliminate the be incompatible with social preferences for systemic stabil- money-issuance rules typical of the gold-linked fixed ity. Failures to service deposit withdrawals can cause a dis- exchange rates, thereby enabling central banks to provide ruptive chain reaction on the payments system, with liquidity more broadly and for longer maturities, not only potentially high costs to the real economy. If massive against the collateral of government securities and com- deposit withdrawals are associated with capital outflows- mercial paper but also against loan portfolios. In some as they are likely to be in the midst of a sharp erosion of Latin American countries, in the 1960s and 1970s, central confidence in the domestic banking system-an excessive banks became the main source of funds to banks. As a contraction in credit would ensue, with prolonged adverse counterpart to these developments, exchange rates gener- consequences on output and employment. ally became fixed-but-adjustable, within the broader Bret- These externalities are a serious concern that governs ton Woods architecture. Hence, as governments reduced much of the authorities' actions in banking system turbu- exposure to the risk embedded in a hard commitrment to lence. Ensuring safety at all costs could lead to a sharp the exchange-rate peg, their exposure to risk increased via reduction in financial deepening, as was the case under the latent safety net implied by the greater lending powers financial repression, or lead the state to stand behind the for central banks. This latent safety net was to surface in entirety of the banking system's liabilities, as has often the midst of the subsequent liberalization-driven crisis. been the case of ex post safety nets unveiled under a crisis. Financial Liberalization Financial Repression Attempts at financial liberalization in Latin America- Financial repression is an interesting historical episode that beginning in the Southern Cone in the 1970s and spread- illustrates the trade-off involved between risk-shifting and ing through much of the rest of the region during the financial deepening (or the lack of thereof). With the clos- 1980s and 1990s-have not been painless. In most coun- ing of Latin American economies and the advent of import- tries financial regulatory schemes had not changed appre- substitution industrialization, many financial systems ciably since the 1940s. But information problems during became instruments of government policy. This led to liberalization became more severe than in the preceding 40 severe distortions in resource allocation and "repressed" the years as banks' portfolios switched from low-risk, low-yield development of financial systems (McKinnon 1973). government paper to much higher-yield and higher-risk Financial repression, however, created banking-system sta- loans to companies, construction, and consumers. bility in the sense that under that regime banks rarely In the rapidly changing circumstances fueled by finan- failed. Asymmetric information became less of a problem cial liberalization and the associated rapid credit growth, for banks and bank supervisors because much of banks' monitoring of borrowers by banks was difficult, and there portfolio assets were held as central bank reserves, treasury was much incompetence among bankers regarding both 48 INSTITUTIONAL REFORM IN MARKETS THE CASE OF THE FINANCIAL SECTOR initial and ongoing evaluations. There was equally great orthodox on the surface, the apparent initial success of the incompetence by bank examiners whose skills had been liberalization process was held together by the strength of blunted by the rather boring situation of atrophied systems an implicit government guarantee for bank liabilities. This under financial repression. Modernization of supervisory was the case in Argentina, Uruguay, and Chile at the agencies often got off to a slow start and, given the natural beginning of 1980s, Colombia in 1985, Venezuela in complexities and typical delays of institutional reform, 1994, and Mexico in 1995. In each case the true bank could not keep up with the fast-acting, stroke-of-the-pen safety net was only unveiled ex post and in an ad hoc man- liberalization reforms.21 Additionally, banks hired away ner as the financial crisis began.23 many of the best bank examiners at high salaries, and the remaining examiners were too few and too powerless to Crises engage in prudential supervision. During liberalization, The improvised unveiling of ex post safety nets has bank capital was inadequate vis-a-vis the new and more reflected authorities' attempts to prevent a financial melt- complex risks, and even published capital/asset ratios were down, characterized by failures in the payments system and frequently overstated by concealing, or simply not report- a collapse in credit with deleterious and long-lasting effects ing, bad loans and by double gearing within an economic on economic activity. Typically, ex post safety nets have group.22 taken the form of an implicit or explicit blanket protection Compared with the Chile of the 1920s, the Chile of the of bank liabilities. The implicit ex post case has tended to 1970s felt that it could not afford to follow orthodox bank- apply more to weak or near-crisis situations, with agents ing rules for fear that the liberalization process would be perceiving an increasing willingness of governments not to derailed. This and an apparent growing concern among allow bank failures, at least for banks considered to be "too authorities about systemic and contagion risks in an econ- big to fail," and acting in accordance with such perception. omy open to capital flows eroded the political will to Often this type of implicit ex post safety net also has implement bank closures when needed. The relaxation of included widespread regulatory forbearance and abuse of rules governing central bank provision of liquidity con- access to central bank liquidity facilities, which have dis- tributed to that erosion, as deep-seated bank problems guised the large and growing undercapitalization or insol- could be temporarily masked by the use and abuse of cen- vency in banks, with the illusory expectation that an tral bank lender-of-last-resort facilities. upturn in the economic cycle would restore viability to When banking crises erupted in Latin American coun- troubled banks and their debtors.24 If and when a major tries, beginning with Chile in the late 1970s, governments banking crisis has begun, and as part of the emergency con- tended to bail out banks and implicitly guarantee bank lia- tainment efforts, ex post safety nets covering all banking bilities. In Chile, for instance, the rescue of Banco Osorno system liabilities have in a number of cases been made in early 1977 saved foreign creditors from losses that would explicit by governments (e.g., Mexico in 1995 and South- have put an end to capital inflows that were helping to fuel east Asian countries over the past year). the economic recovery. The implicit government guarantee Regardless of their specific form, such ex post and often meant that after 1977 bank spreads-the difference universal safety nets for bank liabilities, while preventing between loan interest rates and deposit interest rates-were a financial meltdown, have implied massive risk-shifting much lower than they would have been. The lower bank toward the government. They also have introduced sub- spreads encouraged further financial deepening accompa- stantial social costs through severe incentive distortions in nied by excessive risk-taking. The ratio of private sector financial markets, particularly in the form of runaway domestic credit to GDP rose from 8.8 percent in 1977 to moral hazard. As ex post safety net protection in a number 39.3 percent in 1981. But the implicit insurance-in the of cases extended beyond bank depositors toward bank context of severe problems of asymmetric information, poor debtors and bank owners and managers (e.g., Mexico in monitoring capacity, and low bank capital-also caused 1995), perverse incentives have tended to proliferate even excessive risk-shifting to the government. more rapidly, eroding efficiency in financial intermedia- Thus, although the Chilean financial liberalization and tion. Experience has widely shown that the only way to many other liberalizations in Latin America appeared extract the good out of these major distortions is for gov- 49 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER ernments boldly to seize the occasion to restore banking cant segments of the banking system remain undercapital- system soundness through well-designed and executed cri- ized or insolvent, the formalisms of an apparently well- sis management and resolution programs. The Chilean sys- designed safety net will not reduce risk-taking, because the temic bank recapitalization and restructuring process in behavior of banks without net worth would be dorninated the early 1980s and the Argentine post-Tequila program in by incentives to "gamble for resurrection" or to loot the the mid-1990s provide regional illustrations of certain bank at depositors' or taxpayers' expense. In those circum- "best practice" elements in this connection.25 stances, the first order of business, which must be accom- As "best practices" in resolving banking crises are being plished prior to recasting a new and functional safety net, identified in an increasingly systematic Way,26 one impor- would be to root out insolvency through recapitalization, tant challenge is the need to design well-functioning reg- merger, or liquidation (Garber 1997). ulatory and safety-net arrangements. This need arises in part from stylized facts that show how ex post safety nets Capital emerge regardless of precommitments, which sometimes Capital is the first leg of the tripod of a well-functioning involves excessive risk-shifting to the government. The safety net. Capital is the difference between assets and lia- design of well-functioning safety nets is clearly a very com- bilities; it represents the ownership interest in a firm. plicated task, not the least because experience has shown Because bank owners have a residual claim on the bank cash that the excessive risk-shifting to the government and the flows, capital helps reduce the problems associated with incentive distortions created by ex post safety nets are very information asymmetry by reducing incentives for excessive difficult to dismantle. risk-taking: Other things being equal, as more of their own wealth is at stake, bank owners would have greater incen- The Safety-Net Tripod: Capital, Monitoring, and tive to monitor the activities of bank managers, pressuring Closure for improved internal controls, and less incentive to "milk" One recurrent theme of the foregoing discussion is the ex the bank for their benefit at the expense of depositors. Cap- post emergence of ad hoc safety nets, which causes ex ante ital aligns incentives much as collateral does in the loan rules to be abandoned. This suggests that credible efforts side. In addition, capital provides a cushion for losses: the to "outlaw" risk-shifting to governments are unrealistic in greater the amount of capital, the greater the amount of the case of banking systems, except perhaps in a very nar- assets that can default before the bank is technically insol- row set of cases.27 Hence, it becomes crucial to concentrate vent, lowering the bank's risk (Koch 1992). on the design of ex ante regulatory and safety-net arrange- While from the point of view of regulators more capital ments that could better align the incentives of agents. is sure to align incentives better than less capital, it also Well-functioning safety nets would promote sustainable increases lending rates and reduces financial deepening.28 financial deepening by strengthening market-originated So, beyond a certain threshold, there can be "too much" discipline, while carefully minimizing and controlling capital in light of other financial-sector policy objectives.29 risk-shifting to the government. This would ensure greater At the same time, competitive forces in increasingly inter- resiliency vis-a-vis systemic disturbances and, thus, mini- nationalized markets create pressure on regulators to make mize the probability of abandoning ex ante rules under bad capital requirements lower, but this can erode incentive states of the world. To this end, the design of functional compatibility. In this context, the best international prac- safety nets should ensure that they reinforce rather than tice has been to establish minimum standards, leaving supplant the tripod of private capital, monitoring, and clo- countries free to set their individual requirements above sure mechanisms. As will be explained later, an explicit such minimums. deposit-insurance scheme may play a useful role in insur- There is no simple formula to ascertain the appropriate ing safety-net functionality, particularly regarding the minimum level of capital for a bank. This depends on a host most difficult to enforce element of the tripod: closure. of factors, including the degree of macroeconomic volatility, The feasibility of establishing functional safety nets is the quality and frequency of bank examinations, the relia- not independent of initial conditions and raises important bility of published information, the quality of private mon- sequencing issues. For instance, in countries where signifi- itoring (through internal controls, external auditors, risk- 50 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR rating agencies, large depositors, etc.), and the riskiness of interlinked levels of monitoring in banking systems: inter- bank activities, whether they are on or off the balance nal governance, external governance, and interna- sheets. Liberalization of financial markets confronts banks tional governance.32 The functionality of these forms of with more varied and increasingly complex risks to manage. governance requires a sound broader legal and institutional In addition to credit risk, banks must increasingly manage environment, including, crucially, a reliable judiciary and liquidity risk, interest rates, exchange rates, and other mar- adequate and enforceable corporate, bankruptcy, contract, ket risks, as well as risks associated with conglomerate and private-property laws. Also, these levels of monitoring structures, derivative products, and other off-balance sheet work to the extent that bank shareholders and at least the items (Koch 1992). In this context, the simple Basle ratios large and sophisticated bank creditors have their funds of capital to risk-weighted assets, which initially focused on truly at risk. Transparency-particularly through reliable credit risk, are being complemented by more sophisticated (timely, consistent, and accurate) information and adequate risk-assessment and risk-management models. disclosure standards-is a necessary condition for effective In a number of Latin American countries the regulatory monitoring at any level. Conversely, good governance authorities have set capital-to-risky-asset requirements at would itself improve the reliability, depth, and coverage of levels higher than the 8 percent Basle minimum, princi- information and disclosure. pally in recognition that the relatively more volatile envi- The entry of reputable foreign banks, particularly in the ronment calls for greater capital cushions. For instance, form of branches, into the domestic system could signifi- recent changes in supervision have set them at 11.5 percent cantly improve overall governance in banking. Foreign in Argentina, 9 percent in Ecuador, and 11 percent in Peru. banks accelerate the strengthening of developing countries' However, relative to industrial countries, much effort is still financial systems by "importing" good banking prac- needed in many Latin American countries to improve both tices-including high-quality internal controls to monitor, the measure and quality of capital. Reliable capital mea- evaluate, and manage risks; access to extra liquidity in cri- sures hinge on appropriate accounting standards (including sis times; and the services of home-country regulation and rules for asset-classification, loan-loss provisions, and supervision (Gavin and Hausmann 1997). However, in the income recognition on non-performing loans), information absence of effective exit policies, foreign bank entry would and disclosure requirements, limits on loan-concentration put downward pressures on the franchise value of weak and loans to related parties, and enforcement.30 domestic banks, distorting incentives and leading to exces- Even where accounting and disclosure standards appear sive risk-taking in an environment of unhealthy competi- adequate, the quality of capital can be undermined by a tion-e.g., distressed borrowing at high interest rates that combination of shallow markets for bank stock and the forces general rates up in a disorderly fight for market share concentration of wealth (Rojas-Suarez 1997). This is (Garber 1997). because interconnected balance sheets, which are typical of wealth-concentration, make it easier to offset capital Internal Governance increases with increased debt contracted directly or indi- Internal governance succeeds in reducing informational rectly through related parties. These conditions also facili- asymmetries and conflicts of interest to the extent that tate "creative accounting," making it difficult for supervi- bank shareholders, with their own money at risk, have sors to verify that the capital is real, that it does not consist incentives (a) to appoint able directors and managers and of borrowed funds, and that it is actually paid up in liquid (b) to ensure, through those directors, that managers strive form. Also, the absence of a liquid market for bank stock to increase the value of the bank and do not divert net earn- leaves private and official supervisors without a key price ings away from shareholders through excessive salaries, signal to ascertain the true value of a bank's capital.31 purchase of overvalued assets, and other forms of wasteful expenditures or imprudent investments. The reduction of Monitoring agency problems via sound internal governance requires Monitoring is the second leg of the tripod that sustains a that bank licensing be based on appropriate "fit and well-functioning safety net. Lindgren, Garcia, and Saal proper" tests for bank owners, directors, and managers.33 (1996) usefully distinguish three mutually reinforcing and Internal governance would fail from the outset if the regu- 51 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER latory environment does not normally prevent unscrupu- ing appropriate rules for loan-loss provisioning, asset-clas- lous bank owners and administrators, whose main goal is sification, and income-recognition; and promoting risk to loot the bank, from getting into the business. diversification and prudence in risk-taking (not only as regards credit risk, but also liquidity, exchange-rate, inter- External Governance: Market Discipline est-rate, and concentration risks, as well as risks associated External governance encompasses market discipline as well with conglomerate structures). To reinforce market disci- as official regulation and supervision.34 To the extent that pline, official regulation and supervision should emrphasize relevant bank creditors (large depositors, subordinated information disclosure of consolidated financial statements, debt holders, and interbank creditors) have their funds at and it should enhance the role of credit bureaus ancd credit- risk, they have incentives to monitor the bank's data-and rating agencies, and provide for smooth exit of non-viable they can respond to problems by withdrawing their banks (see below). resources, which could force the bank to close unless it Bank supervision, through off-site monitoring and on- promptly corrects its deficiencies. These creditors also can site inspection, seeks to ensure compliance with prudential require a higher interest rate, although only up to a point, regulations, enhance the quantity and quality of informa- because the adverse selection problem accentuates incen- tion available to the market, and provide a seconcL line of tive incompatibilities between agents. Responses to the defense to complement internal governance and market perception of weaknesses in a bank are quickest in well- discipline. Official prudential regulation without supervi- functioning interbank markets, where weak banks often sion will be worthless. The rapid pace of financial innova- are forced out of the market-hence the relevance of inter- tion and globalization, and the presence of complex finan- bank market signals for supervisory systems aimed at early cial and mixed conglomerate structures give rise to detection of problems (Rojas-Suarez 1998). Appropriately difficult challenges for official supervisory capacity. defined responsibilities for external auditors can substan- Supervisory agencies in many Latin American countries tially enhance overall governance, while credit-rating are still in a process of transition from the rule-oriented, agencies can play an important role in strengthening mar- mechanistic practices brewed under the era of financial ket discipline. Recent reforms in Latin America have repression, toward more sophisticated and forward-looking started to implement these practices. methodologies to evaluate a bank's risks and prospects in an integrated manner. This transition is also toward a more External Governance: Official Regulation and dynamic assessment of the quality of management and the Supervision adequacy of the internal systems used to evaluate, control, Official regulation and supervision is a crucial form of and report on risk. External auditors-with their back- external monitoring. While there are no substitutes for ward-looking assessment of the quality and consistency of internal governance and market discipline, there are also accounting practices, accuracy of financial statements, and limits to these two forms of monitoring. The fundamental adequacy of internal risk-management systems-can com- rule of a well-functioning system of official prudential plement official supervision. Indeed, the trend is toward oversight is that it should be complementary to private- increasing reliance on external auditors by supervisory sector monitoring and market discipline. To reinforce the authorities. operating environment for banks, official regulation should To perform its functions properly, the bank supervisory include procedures for the granting and revocation of agency must have sufficient human and financial resources, licenses; a definition of the scope of banks' permissible which implies budgetary autonomy and the capacity to activities; and a clear layout of the industrial structure of attract and retain high-quality professionals through ade- banking, which includes specifying the nature of its con- quate salary policies and merit-based career paths. Supervi- nection to other financial businesses (particularly securities sion must also be empowered with clear legal authority to and insurance) and to non-financial activities. act without delay in dealing with troubled banks and To reinforce internal governance, the focus of official against managerial incompetence and non-compliance regulation and supervision should be on fostering adequate with prudential norms. Supervisors need to be legally pro- capitalization of banks; restricting insider lending; ensur- tected against personal lawsuits for actions they take as 52 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR Mad*et Dlscipil%1Leuder-o14Ast-Resort Facilites,, and Exchange Rate Sydstem tent r-bank lender-fx st-resrt filities: could under- jreverse repo nd to be in the: 1-day range. LOLR tfacil- mie mmarket discipline.un ess t ly short-term . ities tend to be structuetd as overcollateralized l'oans, and inc rrpoate prons tomnmiz rs selection wre accetable collateral! nrally consists of top an nmoral hazar -an n jiorti. - . maketdiscipline, but theyca sob de lTh feaiity of L R cilities is i er Tey rt to -enail- ocialIy txternalities (contagi the degre commitmet to a aminal exchange rate, - ranted panic, payments systemdissprion, excessive the extreme 'of hi ct c-edit compression, err.). To curb undesirable externalities where the ixed nominal exchange rate is baked by a rgid - while not blunting market disciplit, "best pti" pr oney issuance rule, acdng- to which the chainge i isis for ende t re ot (LO ) fci-ities include hih redbe t hd b strict tatnards for. aes. (e.g., repsentations that the- e uivaent change in centra bakio S . ban: is solvet) oveca ization and hihi h-,quality Und a strict currency board aOciities are ruled- collatera req i t restrictions onS th uses of the liq- our.A curre rules out tOtE. faicili and idiyprodedbthe cn tral bank (e.gg-:prohibition of M reina approaches this ca buit di icresig risky assts, so sto avoid .a de facto subordi- three important dertures froi it. First, the converribil- -nation ofdepositors i-ncase the bankiturns out to be inso- ityr law allrws for uiof central ba v-ent) covenants allowing for intensified monitoring of internationa resees to be constituted by d1ollar-eomi- the bank and its kaities; and well-specified penaltie i naed government securities, which del.nks the au aic cse f nc iance with such covenants. This tpe of tonnection between changes in true foreign exchange provision would effectively support market dkisiline if reserv ai-powered money; it o the lender oId a e threat th it can eaves-room for a centra b to g in opera precipitate, rather than dela, te losure of a non-viable formall equivlent to LOLR faclties, Second the con- bank kwherewar-ranted..vertihiliry law enables a delinkingo e connection. tOtE. facilities are often consed hopenma oner aggregates and ca nges in ..opetatio,ns ( s), but they are nceptually bandop ritby a l ing t e central bank tchange,at its tio df t. s -normally icl de shor-ter di cretion a r p rrve reposad reverse with gover 'ment or central bank rwhih a l investe paper e aim -t arolngh rowth ofmonetay -ex tera -d i ird, aggregates. A a dervaie effect, oMOs give access to the Argetine authrities hve 'orchestrated the invoive Iminiediate liquidity -t banks holding governmetor ten- ment of the foreign banks as4lnder.s of last r b se-! tral bank par. LtOt &cilities, do not aimat ring u d commitment f6e f, a co conialing gr,fi in -oir att -a) ft lie iX t1 controllig growth nmonetay aggregates (lhuh eplieequivalent to about 10 percent0 ofth sysem's he afect it); but rather seek to avoid eter alities sits. scon nt eks ime- .most atin. ettcan counties the maturities. of LiR ace dollat liquidity to goverLnent bonds held by banksI f--Eacilities are in the 300dyirang whreas repos and 0drin t de r l a cait flight. part of the proper discharge of their official duties. All of International Governance this requires that the supervisory process be insulated from International governance is a level of monitoring that is political interference and from the pressures of powerful receiving increasing attention as financial markets become bank lobbies. Many Latin American countries have a long more international. Such a process can complicate informa- way to go in achieving the type of institutional capacity, tion asymmetries and weaken monitoring, promoting authority, and independence described here. excessive risk-taking on an international level. Typical con- 53 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER cerns have to do with the cross-jurisdictional nature of provided safety net so they can "gamble" in emerging mar- operations; the potential for international regulatory arbi- kets whose risks they do not sufficiently understand. trage; and the growing linkages between banking systems Furthermore, the very process of increasing the presence across borders, which magnifies spillover effects of what of foreign banks in domestic markets stimulates the were initially localized crises. In addition, the recent finan- improvement of national supervisory systems and their con- cial crisis in Southeast Asia and the associated international vergence toward international standards, quickening the rescue packages have accentuated concerns about interna- pace of cross-country coordination. Given asymmetric infor- tional moral hazard in at least two dimensions. First, emer- mation, if depositors have a choice between domestic and gency lending by the IMF and other multilateral agencies, foreign banks they would tend to choose the bank whose aimed primarily at stabilizing capital accounts and supervisor is considered to be more effective. This will cre- exchange rates in order to prevent large and undesirable ate an incentive for national banks and supervisory authori- externalities, also has the unintended consequence of bail- ties to improve supervision to attract or keep depositors. ing out large portfolio investors, thereby eroding monitor- Up to now, the main vehicle to improve international ing incentives and market discipline. Second, banks in governance in banking has been international coordina- industrial countries, as they experience competitive pres- tion. In general, arrangements and mechanisms to enhance sures in their home markets, may ride on their home- international governance are at an embryonic stage. Fail- BOX? .2 .. 5.... *-*- ' ' * ....................... .. .. lfleSfziona harmoiWatio dfeults a'i:..t niiy a 4c& 4&y- l-^.ow0the ro?dtoWa$l0tY.'''Y.. Yss msssidno u searoa syt.SandardYs in'g: sdpsrvisi0rr.q: gin thes (.-iO M|g ,ore reenl the. JoiSntou of the9 flasl dvelbpizi wol Thes obtces inltd 'rsrCost 'Cm ite opsdo ak nuac n euiis inomto-hrn soite ihlglcnietat .suprvS333isorst3yt ha pf.romul 3gate rectOommenda3333323333 3Y1tionY3,s f33 th3e,,*33Y proisions, Yy,* adiibn'y,j*{ ful*l hanoniaatio of, prdnil '.Suevso of Fiaca Cogoeates~I h ee omsi nt- always aprprae facossc as th7ge 'dre wrl te os abiios ndcoprhesie fmarocoomc oltiit te uait o aco54h INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR ures in cross-border coordination between supervisors non-financial firms. Both conditions tend to magnify the abound, as illustrated by such well known bank failures as real or perceived systemic risk implications of badly han- that of the BCCI, Barings, and Daiwa, and the lesser- dled bank failures. The "too big to fail" phenomenon not known Latin American cases, such as Banco Latino in only implies erosion of monitoring but may also deter Venezuela. Additionally, little has been done to avoid supervisory authorities from applying stiff penalties, weak- spillovers and systemic risks arising from bank failures. In ening enforcement. Political economy constraints abound, particular, there has been no international harmonization not the least because bank owners typically have substan- of exit policies and standards for early intervention and tial political connections and lobbying powers-indeed, as bank-failure resolution. And the broader issue of a "new Garber (1997; p. 184) puts it, "Closing down or strin- international financial architecture" looms large in the gently disciplining a bank is inherently a political act in all agenda for the future.35 countries." Supervisors may fall prey to "regulatory cap- ture," partly due to corruption or to a perception that the Closure and the Complexities of Exit Policies supervisors' main role lies in preventing bank failures; as a Compared with capital and monitoring, the third leg in result, regulation may be manipulated so as to favor those the tripod that supports a functional safety net, closure, has that it was intended to constrain (Stigler 1971). Finally, been much less analyzed in the academic literature and has the absence of adequate legal and institutional framework not been systematically dealt with in ongoing interna- for the orderly resolution of failing banks undermines the tional coordination efforts. As a result, there is much less authorities' willingness and capacity to take non-viable consensus as to what constitutes best international prac- banks and unfit bankers out of the market. tices in this area. However, it is logical that capital and Ineffective bank exit policies lead to a frequently monitoring alone, no matter how well designed, can fail to observed bad equilibrium that, in the absence of shocks, sufficiently resolve the problems associated with informa- can be fairly durable. This equilibrium may be labeled "the tional asymmetries without closure, which is the disciplin- walking dead bank syndrome," and is characterized by ing event par excellence. Severe misalignment of incentives insolvency and undercapitalization in important segments between participants in the banking system occurs if bank of the system, disguised by some form of regulatory for- shareholders and bank creditors (other than small deposi- bearance that enables window-dressing accounting and tors) do not really have resources at risk. Exit policies give postpones corrective action. In this type of situation, full meaning to "having resources at risk."36 authorities and bankers tend to join in a speech of denial However, bank closures caused by pure market forces- while hoping that an upturn in the business cycle will cure essentially by a run on the bank-and liquidation carried the problem over time. In reality, however, the insolvency out without an appropriate framework can produce sizable "hole" only grows bigger, and the viability prospects for negative externalities (see ad hoc safety nets and externalities, troubled banks only deteriorates, as a result of the associ- above). Consequently, countries typically do not rely on the ated dramatic exacerbation of perverse incentives. Reckless unfettered operation of market forces when it comes to risk-taking by weak and unsound banks force sound banks bank closure. Regulated exit is the norm. Exit policies and into short-term strategies that are unsustainable in the procedures are typically put in place with a view to han- longer term and end up weakening sound banks, adding to dling failures of individual banks while minimizing nega- systemic unsoundness. Bank owners and managers of trou- tive externalities. The efficacy of these policies and proce- bled banks tend to be driven by a "heads I win, tails you dures, however, varies widely across countries. lose" attitude, and face an escalating temptation to loot the Exit in banking systems is a more complicated affair bank, as they move from "management by window dress- than closures in the non-financial sector. This is due to a ing" to "desperate management" (de Juan 1987). Bank host of factors. Problems of asymmetric information are creditors tend to act under the perception of a universal relatively more pronounced in the case of banking, and safety net, with the attendant erosion of monitoring incen- even more acute as banks approach insolvency. Banks are at tives. Bank debtors tend to organize themselves to lobby the heart of the complex web of payments-clearance and for protection (debt forgiveness) from the government.37 In settlement systems, and they are highly levered relative to these circumstances, efficiency in the intermediation 55 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER process suffers severe erosion and systemic vulnerability to Early Warnings, Prompt Correction, and Intervention shocks rises dramatically. Early-warning systems suitable for Latin American coun- tries have to integrate macroeconomic, sectoral, and micro- Fire Protection and Fire Fighting: Designing Exit economic aspects. While systems developed in industrial Policies countries typically emphasize CAMEL-type microeconomic The key elements of effective bank exit policies are analo- variables (an acronym for quantifiable indicators of Capital, gous to good fire protection and fire fighting: There is a Asset quality, Management, Earnings, and Liquidity), a premium on prevention, but there also is the wherewithal recent trend in the literature on early-warning systems for to deal with fires that do occur. In order to ensure that developing countries has sought to combine macro and buildings are safe against fire hazards, fire-resistant mate- micro aspects, while endeavoring to identify microeconomic rials are promoted and building codes and other regula- variables that could have greater predictive power than the tions are enforced. Building inspectors ensure compliance traditional CAMEL variables. For instance, Rojas-Suarez with these regulations and minimum building standards. (1998) shows the usefulness of alternative indicators such as Smoke alarms and fire-detection devices are installed. Fire fast growth in loan portfolio, relatively higher deposit rates, extinguishers prevent small fires from getting out of hand. low financial spreads, and decreased access to interbank mar- There are carefully designed fire escape routes and regu- kets in predicting banking crises in some Latin American larly conducted fire drills to avoid disorderly evacuation countries (Mexico, Venezuela, and Colombia). and unnecessary panic if the worst comes to pass. Fire The approach based on early-warning indicators should departments and firefighters can be quickly summoned be complemented by forward looking, dynamic simulations through emergency phone numbers or automatic fire or "stress tests" of the performance of banks under dlifferent alarms. They have well-rehearsed procedures to deal with scenarios. Early-warning systems are off-site tools cf super- large fires and, when in action, seek to contain the fire vision that help identify banks that should be subject to from spreading to neighboring homes or buildings. Poorly special, intensified on-site monitoring. Such closer scrutiny constructed buildings or those severely affected by fires is justified given that, as has been already mentioned, can be "condemned" and demolished if they are not fit to incompatibility of the incentives facing bank owners, man- be occupied. agers, bank creditors, and debtors increases exponentially as For a city to be safe from fires, it must rely not only on the economic value of the bank decreases toward insolvency. good building codes and their enforcement through Prompt corrective actions aim at automatically curb- inspections but also on its fire department to fight those ing risk-taking by owners and managers of a bank that falls occasional fires that do occur. Similarly, for a banking sys- into increasing capital shortages relative to the required tem to be safe and sound, it not only needs risk-based cap- level, and at reversing the capital deficiency quickly. The ital standards and integrated internal, external, ahd inter- key purpose is to leave little or no room for discretion, national monitoring; it also needs effective exit policies. thereby avoiding the rarely effective and usually detrimen- Exit policies include some basic components to function tal delays in addressing risky practices. Under these rules, well. Early-warning systems (smoke alarms) help detect dete- as soon as a capital shortage is detected, restrictions are rioration of a bank and identify banks that require intensi- automatically triggered, while pre-specified deadlines fied monitoring. Rules for prompt corrective action (fire extin- come into effect to secure the needed capital injection. Ini- guishers) try to reverse bank deterioration through tial restrictions tend to prohibit the payment of dividends increasingly more stringent enforcement measures. Inter- or any increase in risky assets, together with the obligation vention and resolution processes (fire escapes, firefighters) deal to invest new deposits in highly liquid and safe instru- with, and dispose of, problem banks in a manner that is ments. As capital shortages increase, so do the restrictions, expeditious and least disruptive to the system. Basic exit which tend to move from limits on growth toward forced policies, if well-designed, should deal well with a broad shrinkage, as supervisory authorities become increasingly range of bank troubles and failures, while credibly ensur- proactive in selecting assets to sell.38 The main object of ing that non-viable banks, even if large, would not remain prompt corrective actions is to try to secure recapitalization in operation. of the troubled bank by the private sector, without- expos- 56 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR ing public funds. To this end, certain laws, such as Chile's, address typical complexities of "exit" in banking systems, have an explicit range of recapitalization options.39 including the potentially destabilizing effects of disorderly Enforcement of restrictions within the prompt correc- failures. Resolution is the most complicated component of tive action process normally requires a stepped-up form of exit policies-legally and politically. As with any bank- on-site monitoring that, under certain pre-specified condi- ruptcy process, a key element of the needed legal infra- tions, may reach its strongest version: intervention. Inter- structure for bank-failure resolution consists of clear and vention often implies the removal of the bank's manage- enforceable priority of claim rules. These establish the ment, even for reasons other than criminal activities if position in the queue of the various classes of claims over deemed necessary, and the appointment of a "conservator" the assets of the failed bank, with depositors' claims at or or a "temporary administrator" to manage the bank. very near the beginning of the queue, and shareholders' In principle, intervention may serve a number of useful claims always at the end. purposes. The threat of intervention, if credible, will A failing bank is considered to have been "resolved" encourage compliance with prompt corrective actions when the following conditions have been met: (a) its required from bank owners and managers. Once in effect, insured deposits have been paid off in cash or transferred to intervention can allow unfettered room for the supervisors another, sound financial institution and (b) its assets have to determine the true condition of the bank. It can also been disposed of and its non-insured creditors have been help "conserve" the franchise value of the bank, protecting paid according to well-defined rules that set the priority of depositors from the harm that could be caused by bank claims. From the point of view of systemic stability, owners and managers bent on "gambling for resurrection" accomplishing the first condition in an orderly and rapid or "looting." And it can provide unhindered room for spe- manner is of utmost importance, with some analysts and cialized personnel to evaluate alternative ways to resolve practitioners limiting the definition of resolution to it. the bank and make recommendations on the best course of While the "plain vanilla" form of resolving a failed bank is action. To achieve these purposes, the intervention author- closure cum liquidation, resolution encompasses a broad ity must be firmly and unambiguously established in the range of alternatives. In fact, traditional liquidation (where law, and it should provide for the suspension of the rights all creditors are paid off according to the priority of their of shareholders to interfere with the process. claims but only to the extent that the assets of the failed In practice, however, the execution of an intervention bank are collected or sold) tends to be a suboptimal form faces a number of difficulties, particularly in emerging of resolution.41 Well-designed exit policies should leave economies. Due to legal constraints or human resource flexibility for alternative resolution structures, so long as shortages, often the "temporary administrator" is not the the chosen alternative meets the least-cost criterion and senior banker who is needed under the circumstances, but a there is adequate governance and transparency in its execu- bureaucrat with little managerial and banking experience tion. In principle alternative resolution schemes are possi- who stifles the bank's operations and loses, rather than con- ble in the absence of explicit deposit insurance, although in serves, franchise value. This situation tends to be accentu- practice they tend to presuppose its existence. ated where the legal ground for intervention is weak, expos- Basic alternatives in bank-failure resolution42 include ing the authorities to the risk of being personally sued by the following: First, the immediate cash payment of disgruntled bank owners.40 Also, intervention may fuel, insured deposits, leaving the rest of the claims in the liq- rather than curb, the run on the bank, if depositors inter- uidation ("receivership"), including those of the deposit pret that it has just been placed on a faster track toward liq- insurer, to be paid off over time as assets are disposed of. uidation, which can be a serious problem where confidence Second, the transfer of insured deposits to an existing in the banking system is fragile and contagion risk high. sound bank, with uninsured creditors paid off over time out of the liquidation of all the assets of the failed bank. Bank-Failure Resolution This transfer can be accomplished, for instance, through a Bank-failure resolution procedures are analogous to bank- "negative bid," where the deposit insurer pays cash to the ruptcy proceedings for non-financial firms, but to be effec- bank that asks for the least amount of compensation in tive they must contain elements specifically designed to order to take the insured deposits. And third, the transfer 57 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER of insured deposits and other liabilities, together with Equity considerations normally make the protection of enough "good" assets, to an existing sound bank, with liq- small depositors a politically noncontroversial issue.49 As uidation of the remaining assets and payment of remaining small depositors are not likely to monitor a bank systemat- liabilities over time.43 If existing banks are unable to ically due to the free-rider problem, explicit deposit insur- absorb good assets and liabilities (e.g., because the failed ance for them is also noncontroversial from the point of bank is large and existing banks do not have sufficient view of incentive compatibility. Furthermore, uninsured excess capital), a "bridge bank" could be created (with creditors will indeed have greater incentives to monitor "good" assets at least equal to liabilities), which could be their bank-and to discipline it by withdrawing their sold or merged with another institution later. This third deposits or requiring a premium in interest rates--if well- alternative can be efficient given that the liquidation value designed deposit insurance enhances exit policies, because of assets is significantly lower that their going-concern these creditors will perceive a higher probability oft having value. Its complications lie, however, in the need for the to share in the losses in the event of the bank's failure.50 acquiring bank to examine the assets of the failed bank, And bank owners and managers are sure to behave more which takes time, leads to strategic negotiating, and often prudently to the extent that they perceive that the proba- requires complex contractual contingency clauses (e.g., bility of being bailed out has been virtually banished.' stop-loss, loss-sharing, and profit-sharing clauses).44 While consistent with monitoring and market discipline, The resolution phase may follow intervention if the lat- a limited and explicit deposit insurance is unlikely to sig- ter does not solve the situation of a troubled bank. But nificantly deter contagious runs: the "hot" money is in this sequence is neither necessary nor always advisable, large deposits and non-deposit bank liabilities, and it is given the problems in intervention mentioned earlier. their withdrawal that can bring a bank quickly to its What is essential, however, is for the legal and regulatory knees, posing a risk of systemic instability. framework to specify clearly the threshold at which reso- If the goal is not so much to avoid large deposit runs but lution is formally activated, i.e., the point at which the mainly to facilitate the prompt resolution of unviable shareholders are deemed to have lost their property rights banks, then a case can be made in favor of a deposit-insur- over the troubled bank.45 In addition, it is crucial to the ance scheme that is explicit, limited in coverage, funded by effectiveness of resolution to substantially minimize or premiums paid by the industry, and organized into an eliminate the capacity of shareholders to interfere through agency that would be publicly managed, endowed with legal means with the resolution process, circumscribing intervention and failure resolution authority, ancd given their rights only to pecuniary compensation if they were access to a contingent line of credit from the ministry of to prevail in a lawsuit.46 Without this, the authorities finance.52 In addition, to ensure compatibility between the would not have full control of the resolution process and various components of the safety net, the functionality of would be hesitant to act for fear of personal entanglement deposit insurance with the mentioned traits requires that in legal proceedings. every precaution be taken to circumscribe the lender-of- last-resort facilities of the central bank to liquidity (rather The Role of Deposit Insurance in Bank-Failure than solvency) problems. Otherwise, there would be incen- Resolution tives to abuse those facilities to artificially delay bank clo- The literature has emphasized a dual role of deposit insur- sures, undermining overall governance in the system. ance-protecting depositors from losses in the event of a What follows in this section discusses the mentioned char- bank failure and avoiding the undesirable externalities of acteristics in some detail. panic deposit withdrawals (Chandavarkar 1996)-while Some of the arguments in favor of an explicit scheme focusing concerns on its moral hazard implications (e.g., that would confine its coverage to small depositors have Calomiris 1996, 1997).47 Only recently has there been an been advanced in previous paragraphs. In addition, com- increasing recognition that well-designed, limited, and pared with implicit schemes or to ex post safety nets, explicit deposit insurance can significantly improve market explicit and limited deposit insurance would imply clearer discipline by facilitating early closure without adding moral ex ante rules. These would provide greater insulation vis-a- hazard, provided that capital and monitoring are adequate.48 vis political pressures to bail out all bank creditors and/or 58 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR shareholders in difficult times and would also increase the be less likely to close an insolvent bank, because they-and likelihood of consistent results over the long run. More- the public-tend to perceive their role as one of prevent- over, such a scheme would limit and help control risk- ing banks from going insolvent in the first place, which shifting to the government, making it a cheaper alternative makes the closing of a bank appear as an admission of fail- for taxpayers, particularly if capital and monitoring are ure (Lindgren, Garcia, and Saal 1996). The organizational well designed and consistently enforced. setup described here, which allows for appropriate division Funding via premiums assessed on the banking of labor and improving incentives,53 may be more compat- industry reduces the adverse selection problem to the ible with well-functioning exit policies than alternative extent that membership is compulsory; if membership were organizational arrangements, and may more likely reduce voluntary, there would be an adverse selection ("lemons") the probability of the "walking dead bank syndrome" men- problem as weaker banks would be more willing to pay the tioned earlier.54 premiums than prudent banks (Mishkin 1996; Stiglitz Finally, the access to a contingent line of credit from 1993). However, when both risky and prudent banks pay the ministry of finance is needed for the scheme to be the same premium, there would be an incentive to become credible. Otherwise, market participants would anticipate a risky bank. And if the established premium agrees with the abandonment of the rules of the game under difficult the risk of the riskier banks, prudent banks may engage in circumstances-e.g., if a large bank or a significant num- risky practices since they would already be paying to incur ber of small banks were to become insolvent. In fact, purely such risk. Hence, in addition to compulsory membership, "private deposit insurance schemes have a long and fairly the deposit-insurance premiums should be allowed to vary uniform history-they eventually fail" (Garber 1997; p. according to the riskiness of individual banks-rewarding 198); therefore, purely private schemes are unlikely to help prudent banks. Variable premiums make it easier to avoid in exorcising the ghost of ex post, improvised safety nets. underpricing the safety net. Some form of coercive power is It would be convenient for the ministry of finance to back needed to enforce compulsory membership and variable its contingent obligation to the deposit insurance with a risk-based premiums and this suggests that, in one way or contingent liability to international creditors exclusively another, an explicit deposit-insurance scheme would need dedicated to this purpose. This would help avoid inflation- to be an instrumentality of the state. ary pressures and strengthen the exchange rate, thereby There appear to be other advantages to organizing preventing the deterioration of the balance sheets of deposit insurance into a publicly administered agency, debtors who hold foreign-denominated liabilities. different from the supervisory agency, provided that such Furthermore, if a contingency line to foreign creditors is agency is independent from political interference, has bud- involved-particularly if they are multilateral organiza- getary autonomy and high professional standards, and pos- tions-it could enhance governance if it is accompanied by sesses well-designed governance structures to enhance appropriate conditionality. In order to reduce the attendant accountability and transparency. It could be advantageous moral hazard, this conditionality should focus on improve- to empower the deposit-insurance agency with authority to ments in banking laws, the legal environment, and gover- execute early intervention and to carry out the bank-failure nance for the corporate sector, as well as on strengthening resolution processes. An agency of this type may be more the supervisory and regulatory environment for the finan- likely to provide the right incentives to perform interven- cial sector.55 tion and resolution responsibilities efficiently because, by The risk-shifting implied in the contingent access to legal statute, it would have the mandate to protect the public-sector resources, however, must be counterbalanced interests of insured depositors and the integrity of the by giving "hard control rights" to official supervision (to accumulated insurance fund. In the absence of a publicly ensure appropriate prudential oversight over banks) and to managed deposit-insurance agency, both bank supervision the deposit-insurance agency (by endowing it with power- and closure would have to be in the hands of the supervi- ful tools for early intervention and effective bank-failure sory authorities. But bank supervisors and inspectors do resolution). not typically have the skills needed for bank intervention If activated, the contingent line should be payable and failure resolution. Furthermore, bank supervisors may through future premiums; this would help clarify the way 59 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER in which the burden is to be distributed over time and ance against catastrophes, risk-shifting to taxpayers among different groups in society.56 However, a sudden becomes unavoidable, and tougher monitoring and large use of the contingent line would likely run into the enforcement rules become necessary to limit moral haz- constraints posed by shallow domestic capital markets for ard.59 In the case of banking systems, historical experience long-term government debt typical of many Latin Ameri- has repeatedly shown that risk-shifting to the government can countries. This suggests two possible solutions: First, cannot be avoided in the face of systemic crises, and mar- to allow room for the central bank to purchase government ket participants know it.60 While the complexity of this debt (which should carry market-related interest rates) matter goes beyond the scope of this chapter, it seems use- from the deposit-insurance agency-this would be feasible ful, for illustrative purposes, to classify into three groups as long as a flexible exchange-rate system is in place. Sec- the stylized government approaches to deal with the prob- ond, to enable the government to contract a contingent lem of safety nets under catastrophe-type circumstances. long-term line of credit from international sources-this The "never-ever" approach. In this case, govern- would be the only choice if there is a hard commitment to ments are prevented by legislation from bailing out bank a nominal exchange-rate peg that rules out the central creditors, owners, or debtors, except for the limited cover- bank's purchases of government debt securities. age of an explicit deposit protection, if any. Any bailout beyond explicitly insured deposits would require the Safety Nets, Systemic Risk, and Catastrophe approval of legislative changes. The credibility of this pre- Insurance commitment not to abandon the ex ante rules of the game A well-designed safety net-one that carefully controls would be a positive function of the soundness of the tripod risk-shifting to the government while promoting sustain- of capital, monitoring, and closure, and a negative function able financial deepening-requires a sound tripod of capi- of the degree of the economy's exposure to macroeconomic tal, monitoring, and closure. The desirable design features shocks. If this tripod is poorly designed and enforced, the that have emerged throughout the previous sections, legal, regulatory, and policy framework will burst sooner including, importantly, the elements of functional exit poli- rather than later, even under the pressures of idiosyncratic cies described in fire protection and fire fighting: designing exit risk, let alone the pressures of macroeconomic shocks. In policies, should lead to an ex ante safety net that limits fact, the costs associated with not knowing when to relax incentive incompatibility substantially, thereby enhancing ex ante rules and not knowing how to implement rescue banking system safety and soundness. These features should packages are large (Caprio and Klingebiel 1996). In prac- also lead to a resilient safety net-that is, one that is able to tice, some form of bailout beyond the limits of explicit withstand a broad range of states of the world, without deposit protection can be carried out in most countries altering ex ante rules and without exposing taxpayers to within their existing legal frameworks, which puts us into losses. However, under extreme circumstances involving the second category of stylized approaches. aggregate (systemic) risk rather than idiosyncratic (loan or The "maybe-maybe not" approach. In this case, bank bank-specific) risk, even the best designed ex ante rules may bailouts can occur in some fashion without legislative be overwhelmed and abandoned in favor of ad hoc safety-net changes. Approaches in this group allow for "constructive rules, which under certain circumstances could be the opti- ambiguity" because the authorities can maintain a "no mal thing to do.57 This raises the complex issue of when and bailout" discourse (whose credibility depends on the same how to override ex ante rules and whether the ex ante factors mentioned for the first case) but are known to have design should make provision for catastrophe-type states of the discretion to decide when to unfold a wider safety net the world, so as to minimize improvisation and enhance the than that contained in the discourse. Often discretion is effectiveness of emergency responses. allowed in the use of instruments that are inappropriate to True systemic risk in banking can be likened to cata- the task; for instance, the law may allow the use of lender- strophe situations, and catastrophes tend to be uninsurable of-last-resort facilities without appropriate safeguards, by the private sector.58 Governments have a comparative opening the room for its misuse in dealing with insolvency. advantage in handling catastrophe risks, but to the extent By contrast, if lender-of-last-resort facilities are well that they make explicit their willingness to provide insur- designed and if capital, monitoring, and closure are sound, 60 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR then "constructive ambiguity" may have the advantage of result from disorderly bank failures. But bankers, borrow- keeping market participants guessing while giving flexi- ers, and depositors each have their own reasons for wanting bility to the authorities. In particular, if closure contains to be covered by a safety net that have nothing to do with effective early intervention and bank-failure resolution negative externalities. Bank owners, especially, may be policies and procedures, then a banking crisis probably powerful politically. In addition, if the government relies would be managed much better than otherwise.6" At the upon the banking system as a substantial holder of govern- other extreme, however, the "maybe-maybe not" approach ment debt, then the government may be reluctant to let would lead to acute incentive incompatibility if the discre- banks go bankrupt. Efforts to "outlaw" safety nets are nor- tion it provides lives in the context of an undercapitalized mally unrealistic in the case of banking, and market par- banking system, with weak governance and poorly ticipants know it. Hence, it becomes crucial to concentrate designed exit policies. on the design of ex ante regulatory and safety-net arrange- The "only if" approach. Under this approach, the ments that could better align incentives. legal framework would contain explicit contingency A well-functioning safety net must balance its institu- clauses, according to which, if a relevant set of authorities tional components-lender-of-last-resort facilities, deposit were to determine the existence of systemic risk, then the insurance, capital requirements, prudential oversight, and coverage of bank liabilities could extend beyond the explic- exit (closure) policies-so as to control risk-shifting to the itly insured amount. This approach has the advantage of government, while promoting sustainable financial deep- reducing discretion while not ruling out the need to extend ening by supporting, rather than blunting, market-origi- the safety net if systemic considerations warrant it. This nated discipline. A well-designed safety net would ensure "only if' approach also allows for "constructive ambigu- greater resiliency vis-a-vis systemic disturbances, thereby ity": Market participants do not have certainty as to when minimizing the probability of abandonment of ex ante the authorities would determine that there is systemic risk. rules under bad states of the world, and ultimately improv- As in the previous case, the potential superiority of the ing social welfare. When a safety net's institutional com- "only if' approach hinges on the quality and strength of ponents are poorly designed or inadequately implemented, the tripod of capital, monitoring, and closure, and it pre- it becomes dysfunctional and exacerbates moral hazard supposes that lender-of-last-resort facilities are appropri- problems, which, if sufficiently severe, increase systemic arely designed and implemented.62 instability, thereby undermining the safety net's raison The "only if' and "maybe-maybe not" approaches imply d'etre. that in managing a systemic crisis, there may be com- It has been argued throughout the chapter that well- pelling reasons to protect depositors and other bank credi- functioning safety nets need to be adequately supported by tors beyond the explicitly insured amounts. However, even a tripod of capital, monitoring, and closure. All of these under these extraordinary circumstances, there does not three legs of the tripod are equally important and mutually seem to be a compelling reason for bank owners not to be reinforcing: the effectiveness of each one depends on the the first to absorb the losses. In other words, systemic crises effectiveness of the other two. However, closure (exit poli- may call for bailing out depositors and other bank creditors cies), arguably the most difficult to implement of the three, but not for bailing out bank owners and administrators. legally and politically, has also been much less analyzed in the academic literature and has not been systematically Summary and Conclusions dealt with in ongoing international coordination efforts. Drawing from historical experience as well from the analy- As a result, there is much less consensus as to what consti- sis of incentive problems under asymmetric information, tutes best international practices in exit policies. this chapter has explored issues in banking system safety Well-designed exit policies need to involve effective nets. Even if not envisioned in regulatory arrangements, early-warning systems, prompt corrective actions, inter- safety nets, which imply risk-shifting to the government, vention, and bank-failure resolution. There is a good argu- have emerged ex post and in improvised fashion under ment that a very useful tool in facilitating prompt and financial crisis situations. The pressures for safety nets are orderly bank-failure resolution is explicit deposit insur- related to the substantial negative externalities that could ance-insurance with limited coverage that is funded 61 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER through variable premium assessments on the industry, Notes backed by a contingent line of credit from the fiscal author- 1. Much as been written in the last decade on financial safety nets, ity, and endowed with intervention and failure-resolution including Benston et al. (1986); Brock (1992a); Dewatripont and powers. While thitpeofepsi-inuTirole (1993); Hausmann and Rojas-Suarez (1996); Rojas-Sudrez (1997); Alexander, Davis, Ebrill, and Lindgren (1996); Lindgren, to be consistent with the needed incentive structure, the Garcia, and Saal (1996); Garber and Weisbrod (1992); Sheng (1996); jury is still out on this issue, and other institutional and and Chandavarkar (1996). Papers by Calomiris (1997), Mishkin organizational arrangements may serve equally well in (1996), and Garber (1997) are particularly valuable contribotions to complementing effective exit policies. the safety-net literature for developing countries. The most The design of a well-functioning safety net should not researched aspect related to safety nets has been deposit insurance (Talley and Mas 1992; Cull 1998). Against the background of recent financial crises in Latin America and in Southeast Asia, the literature ing crisis resolution and bank restructuring programs. In has started to deal more systematically with such difficult issues as countries where significant segments of the banking sys- resolution of bank insolvency (Caprio and Klingebiel 1996; Caprio tem remain undercapitalized or insolvent, the formalisms and Keefer 1998; drafts for World Bank 1998b). of an apparently well-designed safety net will not reduce 2. Although many of the concepts described below apply to risk-taking, because the behavior of banks without net broadly defined financial markets, the discussion concentrates on the banking sector, because the focus of the chapter is to analyze issues worth would be dominated by incentives to "gamble for for the provision of safety nets for depository institutions (banks). resurrection" or to loot the bank at depositors' or taxpay- 3. The agency literature deals with a principal and an agent. Typ- ers' expense. In those circumstances, the first order of busi- ically, there is asymmetry in the information that each one holds; the ness, which must be accomplished prior to recasting a new agent has access to information not available to the principal, and and functional safety net, would be to root out insolvency information-gathering and analysis is costly. Even if information is equally shared between the principal and the agent, their incentives thence,uhe dpesign ofa funkrstrctional usafetynetgispnotr imay not align if contracts are too costly to enforce. Hence, the design of a functional safety net is not inde- 4. Tradability of loan contracts is enhanced by such things as the pendent of initial conditions and gives rise to important bank's reputation and the degree of standardization of loan-under- sequencing issues. writing practices. A resilient safety net should be able to withstand a 5. The "lemons problem" was first introduced by Akerlof (1970). broad range of states of the world, without altering ex ante The main argument is that in the market for used cars, sellers and buyers have different knowledge about the quality of a given used rules and without exposing taxpayers to losses. However, car. Since the buyer does nor know with certainty if a used car is of under extreme circumstances involving aggregate (sys- low quality (a "lemon"), she will be unwilling to pay beyond the temic) risk rather than idiosyncratic (loan- or bank-spe- price of a "lemon." The seller of a good-quality used car, in turn, cific) risk, even the best designed ex ante rules may be over- would be unwilling to accept only a lemon's price. This gives rise to whelmed and abandoned in favor of ad hoc safety-net rules. conditions where the market for used cars tends to be a market for lemons, and good-quality cars are not t raded. This raises the complex subject of when and how to over- leosangodqliycraenttae. This raies nthe complexs subject of whenherthx antedew thoud o6. These solutions tend to arise even in the absence of govern- ride ex ante rules, and of whether the ex ante design should ment intervention. Information problems hurt firms by limriing the make provision for catastrophes, so as to minimize impro- willingness of banks to lend and by raising the cost of loans that are visation and enhance the effectiveness of emergency made. Banks, in turn, are hurt by a loss of profits. Thus, both debtor responses. The issues involved can be in large part analyzed firms and creditor banks have an incentive to promote institutions from the viewpoint of catastrophe insurance and involve that reduce selection and moral hazard problems, facilitating the the discussion on rules versus discretion. "Constructive flwocrdt 7. A leverage index, which can be created by finding the ratio of ambiguity" and contingent clauses in the legal framework a firm's debt financing to its equity financing, also is a measure of are two interesting ways of handling the eventuality of sys- potential earnings volatility. The greater the leverage, the more temic-that is, catastrophe-risk. volatile the firm's net profit or loss (Koch 1992). A necessary word of caution: Mechanically exporting 8. North (1990) defines property rights as the exclusive rights for safety-net arrangements from one country to another is very individuals to appropriate their own labor and the goods and services safety-net arrangements from one country to another is very they posses. likely to run into problems. The cultural, historical, politi- 9. Levine (1997b) shows that financial market development is cal, legal, and institutional setup of each country has to be positively and significantly related to legal heritage, again with taken into account if a safety-net design is to be viable. mon law performing better than civil law. 62 INSTITUTIONAL REFORM IN MARKETS- THE CASE OF THE FINANCIAL SECTOR 10. As in the case of loan contracts, the price mechanism alone is very small, the cost is smaller still if she withdraws her own may not work, because the increase in interest rate required by deposits. depositors would tend to exacerbate the adverse selection problem. 20. For example, in Chile, bank capital as a fraction of assets fell 11. Broadly speaking, an externality entails that the cost or ben- from 27 percent in 1932 to 20 percent in 1940, 14 percent in 1948, efit of a certain activity is borne by someone other than the person 9 percent in 1955, and 6 percent in 1962. undertaking the activity (Carlton and Perloff 1990). Externalities 21. This is related to "regulatory dialectic" (Kane 1977). The arise when property rights are not clearly defined. Positive external- underlying argument is that regulation affects the behavior of ities occur when someone undertakes an action that benefits others agents, leading to a different outcome than otherwise. This process who do not pay for these benefits. The externality is negative if induces innovation on the part of the financial agents to avoid regu- agents do not bear the full cost of their actions. lation. These innovations tend to spread rapidly, while authorities 12. In addition to prudential regulation, government interven- tend to lag in adapting regulation to the innovations. Eventually, tion can result in financial repression, which occurs when the gov- though, new regulation takes place and the process starts again. ernment distorts incentives in the financial system, typically with 22. Double gearing means using the same capital to support the hope of achieving certain social objectives. Interest-rate ceilings additional assets. (frequently resulting in negative real interest rates), high non-inter- 23. Recent research has shown some evidence that financial liber- est beating reserve requirements, and targeted credit are examples of alization has contributed to the eruption of financial crises the instruments of financial repression widely used in developing (Demirgu,c-Kunt and Detragiache 1998). countries before the 1980s, with perverse consequences for growth 24. A bank is insolvent if the economic value of its assets is not and development (McKinnon 1973). enough to pay its debtors. An insolvent bank has negative equity in 13. The argument that a good safety net enables agents to act as economic terms. In principle, a bank may be insolvent but liquid, or if there were no net could be interpreted as pointing to the need for solvent but illiquid. Contrary to an insolvent bank, a solvent but prudential regulation to mimic the market to the extent possible, illiquid bank has the assets to meet its debt obligations over time, thus aligning agents' incentives to reduce divergence between the but not immediately due to the illiquidity of the assets (i.e., the pursuit of private and social benefits. assets cannot be transformed into cash fast enough). In the absence of 14. New York's 1838 Free Bank Law stated: "...nothing in this central bank lender-of-last-resort facilities, this raises the possibility act contained shall be considered as implying any pledge on the part of runs on banks, even it they are solvent. The disposal of assets at of the state for the payment of said bills or notes beyond the proper "fire sale" prices could turn a solvent into an insolvent bank. In fact, application of the securities pledged to the comptroller for their illiquidity and insolvency interact and may reinforce each other redemption." See Brock (1992a, p. 431). (Gavin and Hausmann 1997). 15. The Caja was authorized to lend up to 50 percent of the 25. Important elements of best practices outside the Latin Amer- assessed value of the real estate and was given legal precedence in the ican region include the phase of the U.S. savings-and-loan crisis that collection of its loans. In exchange for the mortgages the Caja issued started after 1988, once Congress appropriated needed funds for the letras de crdito with maturities of 21-25 years and coupon rates of deposit insurance agency and created the Resolution Trust Corpora- 5-8 percent. Borrowers could then sell the letras on a secondary mar- tion, and the resolution of Spain's banking crisis during the 1980s. ket for cash. When landholders went to sell the letras on the sec- 26. See, for instance, the book edited by W. E. Alexander et al. ondary market, they received a higher price because purchasers (1997) and Rojas-Suarez (1997). demanded a smaller insurance premium knowing that the Caja, via 27. One possible case is where the entire domestic banking system is its capital and reserves, was the primary insurer of the mortgages. small and fully dominated by first-rate foreign banks. However, during 16. See Brock (1992b) and Fetter (1931). the Latin American debt crisis of the 1980s, banks were bailed out in 17. Deepening, however, was not automatic in all countries with Uruguay, a small open economy dominated by first-rate foreign banks, new central banks. For instance, in Mexico the public continued to because foreign banks threatened to leave the system (Brock 1992a). rely heavily on the Banco de Mexico as a depository institution as a 28. While regulators prefer more capital because of their concern result of continued distrust of commercial banks. for systemic stability and safety, bankers generally prefer less capital 18. While from the point of view of regulators the pressure mainly and higher leverage ratios (Koch 1992; Jensen and Meckling 1976). arises from externalities, there are internal pressures, too, because 29. Narrow banking involves the same trade-off of higher stabil- bankers, borrowers, and depositors each have their own reasons for ity but lower financial deepening. wanting to be covered by a safety net that have nothing to do with 30. Lending to related parties and underprovision of reserves externalities. In particular, bank owners may be powerful politically, against loan losses have been common causes of vulnerability in the which enables them to exert significant pressure over regulators. banking system in Latin America. The former affects the true stake 19. Depositors may run even when there is a deposit-insurance of owners in the bank, and the latter overstates the book value of cap- scheme, if they fear that in the event of a failure the payment from ital relative to its true economic value. the insurer would take a long time. For an individual depositor, 3 1. This adds an argument in favor of promoting bank monitor- while the risk from the withdrawal of deposits by another depositor ing through price revelation in the markets for bank liabilities, 63 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER which are well-developed markets relative to equity markets in Latin assumes its liabilities. It is also a variant of the so-called "good America. This approach can work to the extent that there is no bank/bad bank approach," where good assets and as much bank lia- explicit or implicit government backing of all bank liabilities and bilities as possible are carved out to be transferred to an existing could be accomplished by requiring banks to issue subordinated debt bank, or bad assets are taken out to permit the sale of the "good" to comply with part of their regulatory capital (Rojas-Suarez 1997 bank. and Calomiris 1997), as is done currently in Argentina. 44. Supervisory and intervention authorities can also pressure 32. This subsection draws heavily on Lindgren, Garcia, and Saal owners into a merger or a sale of their troubled bank-typically dur- (1996). ing the prompt corrective action and intervention stages. "Induced" 33. In addition, the criminal history of owners and managers can mergers and acquisitions are also a form of "resolution," with the dif- be taken into account in the licensing process of banks, as has been ference that shareholders still have some net worth at stake. "Assisted the case in a number of Latin American countries (e.g., Argentina mergers" (where public funds are used to enable a merger) are not and Peru). uncommon, especially in crisis situations, but they run the risk of 34. Marker discipline can be undermined by poorly designed upsetting basic priority-of-claim principles by benefiting sharehold- lender-of-last-resort facilities (see Box 3.1). ers ahead of depositors. 35. For a recent discussion of international lenders of last resort 45. This often involves constitutionally defined property rights. for emerging markets, see Mishkin (1998). To ensure that no property is "taken," the U.S. failure resolution 36. The term "exit policies" is better than "closure," as it conveys process (which formally starts when a troubled bank is placed in better the complexity and breath of this key third leg of the safety- "receivership") is triggered when the bank's regulatory capital is net tripod. non-positive. In some Latin American countries, resolution (fre- 37. For instance, the Barz6n in Mexico, which is a debtors' orga- quently limited to traditional liquidation) is triggered before regu- nization. latory capital reaches zero, in order to maximize the probability that 38. Prompt corrective action rules in the FDIC Improvement Act shareholders still have something at risk at time of closure. For of 1991 (section 131) of the United States are considered to be an instance, in Peru a bank is considered "insolvent" when its ratio of example of "best practice." Lindgren, Garcia, and Saal (1996, Table capital to risk-weighted assets is less than 5 percent. Under most 24, p. 198) summarize these rules. A significant number of banking banking legislation in Latin America, deficiencies in loan loss provi- laws in Latin America contain specific clauses in this area, although sions are subtracted from capital in order to arrive at the measure of enforcement varies widely. regulatory capital. 39. About one-fourth of the 1986 Chilean bank law is devoted to 46. The receivership figure in the United States seems to achieve the precise specification of alternative closure and recapitalization this objective well. Labor and corporate laws, as well as the broader mechanisms for troubled banks. The recapitalization mechanisms civil code, seem to complicate matters in this regard for Latin Amer- include recapitalization by the bank's owners or outside capital injec- ican countries. In this connection, the case of Spain (with a similar tions, interbank loans that turn into preferred stock if the borrowing legal heritage) may be a useful reference. It provides for an expedi- bank does not regain solvency, and a recapitalization process that tious route to translate the loss of regulatory capital of a failing bank turns a fraction of a bank's noninsured liabilities into preferred stock. into a legal extinction of capital, thereby enabling the resolution 40. In comparison with the legal protection afforded in the agency (the Spanish Fondo de Garantia de Dep6sitos) to subscribe com- United States for supervisory and FDIC officials (no private right of mon stock and become the sole owner in full control of the bank. action can be leveled against them for actions performed in the dis- 47. Based on econometric results, Demirguc-Kunt and Detra- charge of their official duties), virtually all Latin American countries giache (1997) argued that although the presence of explicit insurance fail to offer sufficient immunity, a situation that is aggravated by may have reduced the incidence of self-fulfilling bank runs, it unreliable judicial systems. This is a key factor in explaining why appears to have worsened banking sector fragility. They suggested there are often delays in taking decisive action on bank closures in that this could be attributable to moral hazard or to weaknesses in various countries in the region. the design and implementation of deposit insurance schemes. Their 41. In most Latin American countries, bank liquidation processes sample of countries included 31 systemic crises (six in Latin Amer- tend to be legally cumbersome and potentially destabilizing, with ica) between 1980 and 1994. depositors typically getting paid several months after the liquidation 48. This is suggested, for instance, by Lindgren, Garcia, and Saal has been decreed. This is in large part because deposit insurance (1996). schemes do not exist, or, if they do, they lack well-designed policies 49. However, from a historical perspective, explicit deposit insur- and procedures, are underfunded, or are poorly managed. ance does not seem to have been politically popular; rather, it is a rel- 42. For a more complete description of alternative resolution atively recent and untested institutional innovation. Since 1981, the structures see Talley and Mas (1992) and Federal Deposit Insurance number of countries in the world with explicit deposit insurance Corporation (1997). programs has risen from 15 to 45 (Lindgren, Garcia, and Saal 1996). 43. This alternative is known as a "purchase and assumption"- Until the last decade, depositors were nor generally included in i.e., the existing bank purchases assets of the failed institutions and explicit safety nets in Latin America; indeed, depositors' interests 64 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR tended to be sacrificed to save bank shareholders through implicit or claims are paid and the time premium are collected could be solved ex post safety nets-a bad habit that lingers on in several present-day by accumulating large pools of liquid capital, but these are difficult Latin American countries. to maintain due to tax disincentives or because they attract hostile 50. Benston et al. (1986) and Calomiris (1996, 1997) have per- takeovers, among other reasons. Reinsurance is limited in availabil- suasively argued in favor of obligating banks to issue subordinated ity: The international reinsurance catastrophe capacity is only $15 debt as a way to improve monitoring. billion. Catastrophe futures and options are only lightly traded on 51. De Juan (1995) notes that discipline increased in the Spanish the Chicago Board of Trade, partly due to the lack of any options banking industry after 1980, as bankers realized that the govern- pricing formula, like the Black-Scholes, for catastrophe losses. ment had come to posses powerful and effective instruments to root 59. For instance, where governments are committed to assist out bad bankers, particularly as regards intervenrtion and failure homeowners whose homes in oceanfront areas have been damaged by resolution. hurricanes, strict rules must be enforced to prevent home builders 52. Talley and Mas (1992) discuss extensively the different design from building too close to the ocean or building with too flimsy con- alternatives for deposit insurance, although they would not necessar- struction materials. ily endorse the combination suggested here. 60. Efforts to avoid the unavoidable can ofren lead to higher costs 53. To avoid the risk of stifling bureaucratization, the deposit than otherwise. For instance, Chile's decision in the early 1930s to insurance agency could, by statute, be forced to rely heavily on sub- stand by the gold standard and to enforce prudential regulation and contracting of professional expertise as needed. bank closures was eventually followed by a military coup and the 54. Having intervention and failure-resolution powers in the adoption of a series of ad hoc economic policies that later formed the deposit insurance agency may create problems of coordination and basis for the import substitution strategy in that country. information-flow with the supervisory agency that would need to be 61. The framework in place in Spain since the crisis management appropriately solved. period of the 1980s would seem to fall under the second stylized 55. This line of reasoning, albeit limited to a deposit-insurance case: "Constructive ambiguity" is allowed in the context of strong scheme, is consistent with Mishkin's (1998) thoughts on potential capital, monitoring, and closure, and the Fondo de Garant/a de Dep&si- international lender-of-last-resort arrangements. tos has remarkably effective early intervention and bank-failure reso- 56. Variable, risk-based deposit insurance premiums pose an lution capacities. added burden on already fragile individual banks, exacerbating their 62. The framework in place in the United States since the FDIC rate of deterioration. This is not an argument against variable pre- Improvement Act of 1991 may be considered as a "best practice" ref- miums; it rather is an argument in favor of intensified monitoring erent in regard to the "only if' approach. Under this Act, when han- and, if necessary, of early intervention and resolution. Onerous pre- dling a troubled bank, the FDIC can extend coverage to bank liabil- miums, however, would be counterproductive if the entire banking ities beyond the insured amounts if there is a joint determination by system or a large segment of it is fragile. the board of governors of the Federal Reserve, the FDIC board, and 57. As discussed in Sheng (1996), even the most sophisticated the treasury secretary (after consulting with the president of the bank supervisors in advanced OECD countries have not been able to United States) that the failure of such bank would entail systemic prevent bank failures completely. Besides adequate supervision, risk. To curb the attendant adverse selection, limit risk-shifting to greater attention is needed to sector and national imbalances that the government, and avoid underpricing of the safety net, the FDIC may destabilize the banking sector and to the creation of an overall is authorized not only to assess risk-based premiums, but also to national risk management system. assess premiums on the basis of total bank liabilities, and not just 58. For instance, because of the Northridge earthquake in 1994 deposits, as banks whose failure could pose systemic risk normally and Hurricane Andrew in the Miami area in 1995, private catastro- have a relatively low ratio of deposits to total liabilities. phe insurance came close to drying up in the United States. In the 63. Currently, the so-called Group of Ten (G-10) actually has 11 case of catastrophe insurance, the law of large numbers does not members: Belgium, Canada, France, Germany, Italy, Japan, the apply. Hence, insurance companies cannot predict accurately the Netherlands, Sweden, Switzerland, the United Kingdom, and the level of claims and cannot adjust insurance premiums so as ro reduce United States. the mismatch between premiums collected and claims paid out in a 64. The relevant publications are listed in the Bank for Interna- given year. Catastrophe insurance requires a large point-in-time pay- tional Settlements (BIS) website (http://www.bis.org/publ/pub- out after a catastrophe occurs, but can be funded only by collecting list.htm). premiums over a long time period. This mismatch between the time 65 CHAPTER 4 Capital Markets and Legal Institutions I N THIS CHAPTER WE FOCUS ON THE INSTITUTIONS REQUIRED TO SUPPORT LARGE CAPITAL MAR- kets and survey the empirical evidence on the link between legal institutions and financial markets. Specifically, we are interested in providing an answer to why we observe such large differences in the size, breadth, and valuation of capital markets. Why, for example, are equity markets so much larger in South Africa than in Mexico or Peru? Why did many companies go public in India and Hong Kong ^,in 1995, while no company went public in Brazil or Uruguay or Venezuela in the same year? Why do countries like New Zealand have large credit markets while Argentina and Colombia do not have them? In a simple Modigliani-Miller (M&M) framework (Modigliani and Miller 1958) the size of capital mar- kets is determined only by the cash flows that accrue to investors. Therefore, roughly speaking, the size of capital markets should be proportional to GNP. To explain the large discrepancies in the size of financial markets across countries with similar GNP, we need to recognize that securities are more than the cash flows they represent, since they entitle investors to exercise certain rights. Shares not only entitle investors to dividend payments, but also to exercise control over management through the voting process. Similarly, debt not only entitles creditors to receive interest payments, but also to regain their collateral in the event of bankruptcy of the firm. The separation between ownership and control can have (Grossman and Hart 1986). Management discretion, a large effect on the size of capital markets once we depart although a cost-effective way of dealing with the separation from the M&M assumptions and allow for the existence of of ownership and control, can unfortunately be used to agency costs. To take an extreme view, outside equity expropriate financiers through outright expropriation, would have no value if shareholders did not have control transfer-pricing, or asset-stripping. rights to force managers to pay out dividends. In the same The agency model could, in principle, explain why vein, creditors would be unwilling to lend money at any some countries have much larger capital markets than oth- interest rate if their control rights did not allow them to ers, since it is apparent that countries differ enormously in punish debtors who default on their financial obligations. the extent to which they afford legal protection to Both financiers and management would benefit from the investors. Not only does a shareholder in Mexico, for exam- elimination of the agency conflict if they could write a ple, have a very different bundle of rights from one in the complete contract that specified what the manager should United States, but his recourse to redress is likely to be do with the funds and how he would give them back to significantly weaker. The agency model predicts larger cap- investors in all states of the world. Of course, a complete ital markets in countries where agency costs are reined in contract cannot be implemented in practice, making it by the law and the institutions built to support their necessary for management to have a level of discretion enforcement. La Porta et al. (1998a) systematically assess 67 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER the rights of investors as well as the quality of their modeled their initial corporate laws on the laws of Eng- enforcement for 49 countries. La Porta et al. (1997a, land. There are 18 common-law countries in the sample. 1998a, and 1998b) relate legal institutions to the size and The French legal family includes France, Spain, Portugal, breadth of external capital markets as well as to corporate and their colonies. There are 21 French legal origin coun- ownership concentration around the world. tries in our sample, including nine in Latin America. The In this chapter, we review and summarize the cross- German tradition has had less influence, and we have only country evidence on the influence of institutions on capital six in this family: Austria, Germany, Japan, the Republic markets' development. The chapter is divided into four of Korea, Switzerland, and Taiwan, China. Finally, the sections. Legal protection to investors, the first section, Scandinavian family includes the four Nordic countries of describes the differences in legal protection for sharehold- Denmark, Finland, Norway, and Sweden. ers and creditors in a cross-section of 49 countries. Because There are numerous potentially measurable differences investor rights are not only determined by laws, Enforcement among countries in their company and bankruptcy laws. of laws compares the quality of the legal enforcement and WVe focus exclusively on those basic rules that scholars (e.g., accounting standards across nations. The ultimate question Paul Vishny 1994, White 1993, American Bar Association is whether countries with poor investor protections actu- 1989 and 1993) and observers (e.g., Investor Responsibil- ally do suffer. If laws and their enforcement matter, then ity Research Center 1994, Institutional Shareholcler Ser- countries that offer entrepreneurs better terms of external vices 1994) believe to be essential to corporate governance. finance would have both higher-valued and broader capital Furthermore, we restrict our attention to those basic rules markets. We also predict that countries that offer entre- that easily can be interpretable as either pro-investor or preneurs better terms would have widely held corpora- pro-management. tions. Consequently, External finance and legal institutions compares external finance and ownership concentration Shareholder Rights across countries as a function of the origin of their laws, the Shareholders have residual rights over the cash flows of the quality of legal investor protections, and the quality of law firm. The right to vote is the shareholders' main source of enforcement. Conclusion andpolicy implications concludes the power. This right to vote in the general assembly for the chapter and discusses policy implications of the results. election of directors and on major corporate decisions guar- antees shareholders that management will disgorge the Legal Protection to Investors firm's cash flows to shareholders through the payment of La Porta et al. (1998a) assembled a data set covering legal dividends rather than divert the funds to pay themselves rules pertaining to the rights of investors, and to the quality higher compensation or undertake poor acquisitions, for of enforcement of these rules, for 49 countries with publicly example. Therefore, voting rights and the rights that sup- traded companies. Naturally, laws in different countries are port voting mechanisms are the defining features of equity. typically not written from scratch, but rather transplanted- Table 4.1 provides a detailed description of all the vari- voluntarily or coincidentally-from a few legal families or ables that we use in this chapter, and Table 4.2 presents the traditions. In general, commercial laws come from two broad evidence on shareholder rights for the cross-section of 49 traditions: common law and civil law. Most English-speak- countries. A useful way to begin the discussion of share- ing countries belong to the common-law tradition based on holder rights is by first assuming the role of an investor in the British Company Act. The rest of the world belongs to a U.K. firm and then switching identity to beccme an the civil-law tradition, derivative of Roman law, which has investor in a Mexican corporation. (We do this not to praise three main families: the French family based on the the United Kingdom, nor to single out Mexico for criti- Napoleonic Code of 1804, the German family based on Bis- cism, but rather merely to illustrate differences between a marck's Code of 1896, and the Scandinavian family, which legal system based on English common law and one based legal scholars describe as less derivative of Roman law but on French civil law.) "distinct" from the other two civil families. The first column of Table 4.2 shows that not all U.K. The common-law family includes former British shareholders have the right to vote. That is probably a bad colonies and other nations like Thailand and Israel who thing, because when votes are tightly linked to dividends, 68 INSTITUTIONAL REFORM IN MARKETS. THE CASE OF THE FINANCIAL SECTOR TABLE 4.1 The Variables This table describes the variables collected for the 49 countries included in our study. The first column gives the name of the variable. The second column describes the variable and gives the range of possible values. The third column provides the sources from which the variable was collected. VARIABLE DESCRIPTION SOURCES Origin Identifies the legal origin of the Company Law or Commercial Code of each country. Foreign Law Encyclopedia Commercial Laws of the World One share-one vote Equals I if the Company Law or Commercial Code of the country requires that ordinary Company Law or Commercial shares carry one vote per share, and zero otherwise. Equivalently, this variable equals 1 when Code the law prohibits the existence of both multiple-voting and non-voting ordinary shares and does not allow firms to set a maximum number of votes per shareholder irrespective of the number of shares owned, and 0 otherwise. Proxy by mail Equals I if the Company Law or Commercial Code allows shareholders to mail their proxy Company Law or Commercial vote to the firm, and 0 otherwise. Code Shares not blocked before Equals 1 if the Company Law or Commercial Code does not allow firms to require that Company Law or Commercial meeting shareholders deposit their shares prior to a General Shareholders Meeting thus preventing Code them from selling those shares for a number of days, and 0 otherwise. Cumulative voting or Equals 1 if the Company Law or Commercial Code allows shareholders to cast all of their Company Law or Commercial proportional votes for one candidate standing for election to the board of directors (cumulative voting) or Code representation if the Company Law or Commercial Code allows a mechanism of proportional representation in the board by which minority interests may name a proportional number of directors to the board, and 0 otherwise. Oppressed minorities Equals 1 if the Company Law or Commercial Code grants minority shareholders either a Company Law or Commercial mechanism judicial venue to challenge the decisions of management or of the assembly or the right to Code step out of the company by requiring the company to purchase their shares when they object to certain fundamental changes, such as mergers, assets dispositions, and changes in the articles of incorporation. The variable equals 0 otherwise. Minority shareholders are defined as those shareholders who own 10 percent of share capital or less. Preemptive rights Equals I when the Company Law or Commercial Code grants shareholders the first opportu- Company Law or Commercial nity to buy new issues of stock and this right can only be waved by a shareholders' vote, and Code 0 otherwise. Percentage of share capital It is the minimum percentage of ownership of share capital that entitles a shareholder to Company Law or Commercial to call an extraordinary call for an Extraordinary Shareholders' Meeting. It ranges from 1 to 33 percent. Code shareholders' meeting Anti-director rights An index aggregating the shareholder rights, which we labeled as "anti-director rights." Company Law or Commercial The index is formed by adding 1 when: (1) the country allows shareholders to mail their Code proxy vote to the firm; (2) shareholders are not required to deposit their shares prior to the General Shareholders' Meeting; (3) cumulative voting or proportional representation of minorities in the board of directors is allowed; (4) an oppressed minorities mechanism is in place; (5) the minimum percentage of share capital that entitles a shareholder to call for an Extraordinary Shareholders' Meeting is less than or equal to 10 percent (the sample median); or (6) shareholders have preemptive rights that can only be waived by a share- holders' vote. The index ranges from 0 to 6. Restrictions for going Equals 1 if the reorganization procedure imposes restrictions, such as creditors' consent, to Bankruptcy and into reorganization. file for reorganization. It equals 0 if there are no such restrictions. Reorganization Laws No automatic stay on Equals I if the reorganization procedure does not impose an automatic stay on the assets of Bankruptcy and secured assets the firm upon filing the reorganization petition. Automatic stay prevents secured creditors to Reorganization Laws gain possession of their security. It equals 0 if such restriction does exist in the law. Secured creditors first Equals 1 if secured creditors are ranked first in the distribution of the proceeds that result Bankruptcy and from the disposition of the assets of a bankrupt firm. Equals 0 if non-secured creditors, such Reorganization Laws as the government and workers, are given absolute priority. Management does nor Equals 1 when an official appointed by the court, or by the creditors, is responsible for the Bankruptcy and stay operation of the business during reorganization. Equivalently, this variable equals 1 if the Reorganization Laws debtor does not keep the administration of its property pending the resolution of the reor- ganization process, and 0 otherwise. (Table continues on the following page.) 69 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE 4.1 (Continued) VARIABLE DESCRIPTION SOURCE'S Creditor rights An index aggregating different creditor rights. The index is formed by adding 1 when: (1) Bankruptcy and the country imposes restrictions, such as creditors' consent or minimum dividends to file for Reorganization Laws reorganization; (2) secured creditors are able to gain possession of their security once the reorganization petition has been approved (no automatic stay); (3) secured creditors are ranked first in the distribution of che proceeds that result from the disposition of the assets of a bankrupt firm; and (4) the debtor does not retain the administration of its property pending the resolution of the reorganization. The index ranges from 0 to 4. Efficiency of judicial Assessment of the "efficiency and integrity of the legal environment as it affects business, partic- Business Inrernarional system ularly foreign firms" produced by the country-risk rating agency Business International Corpora- Corporation tion. It "may be raken to represent investors' assessments of conditions in the country in ques- tion." Average between 1980-83. Scale from 0 to 10, with lower scores lower efficiency levels. Rule of law Assessment of the law and order tradition in the country produced by the country-risk rating International Country Risk agency International Country Risk (ICR). Average of the months of April and October of the Guide monthly index between 1982 and 1995. Scale from 0 to 10, with lower scores for less tradi- tion for law and order. (We changed the scale from its original range of 0 to 6.) Corruption ICR's assessment of corruption in government. Lower scores indicate "high government offi- International Country Risk cials are likely to demand special payments" and "illegal payments are generally expected Guide throughout lower levels of government" in the form of "bribes connected with import and export licenses, exchange controls, tax assessment, policy protection, or loans." Average of the months of April and October of the monthly index between 1982 and 1995. Scale from 0 to 10, with lower scores for higher levels of corruption. (We changed the scale from its original range of 0 to 6.) Accounting standards Index created by examining and rating companies' 1990 annual reports on their inclusion or International Accouncing and omission of 90 items. These items fall into seven categories (general information, income Auditing Trends, Center for statements, balance sheets, funds flow statement, accounting standards, stock data and spe- International Financial cial items). A minimum of 3 companies in each country were studied. The companies repre- Analysis & Research, Inc. sent a cross-section of various industry groups where industrial companies numbered 70 percent while financial companies represented the remaining 30 percent. Ownership, 10 largest The average percentage of common shares owned by the three largest shareholders in the 10 Moodys International, CIFAR, private firms largest non-financial, privately owned domestic firms in a given country. A firm is consid- EXTEL, WorldScope, 20-Fs, ered privately owned if the state is not a known shareholder in it. Price-Waterhouse, ancI various country sources External cap / GNP The ratio of the stock market capitalization held by minorities to gross national product for Moodys Inrernational,CIFAR, 1994. The stock market capitalization held by minorities is computed as the product of the EXTEL, WorldScope, 20-Fs, aggregate stock market capitalization and the average percentage of common shares not owned Price Waterhouse, ancI various by the top three shareholders in the 10 largest non-financial, privately owned domestic firms country sources in a given country. A firm is considered privately owned if the state is not a known share- holder in it. Domestic firms l Pop Ratio of the number of domestic firms Listed in a given country to its population (in milLions) Emerging Market Factbook in 1994. and World Development Report 1996 IPOs / Pop Ratio of the number of initial public offerings of equity in a given country to irs population Securities Data Corpor-ation, (in millions) for the period July 1995 to June 1996. AsiaMoney, LatinFinanace, GT Guide to World Equity Markets, and World Developmenr Report 1 996 Debt / GNP Ratio of the sum of bank debt of the private sector and outstanding non-financial bonds to International Financial Sta- GNP in 1994, or last available. tistics, World Bondmtarket Factbook GDP growth Average annual percent growth of per capita gross domestic product for the period World Development Report 1970-93. 1995 Log GNP Logarithm of the Gross National Product in 1994. World Development Report 1996 70 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR it is more difficult to control a company by having a small Suppose that the shareholders' meeting took place not fraction of the equity. Yet, as it turns out, one share-one in London but in Mexico City. As in the United Kingdom, vote rules are uncommon everywhere-including Mexico. not all shares are endowed with the same right to vote. The next six columns of Table 4.2 provide different However, unlike in the United Kingdom, investors in measures of how strongly the corporate law protects minor- Mexico will be notified of the forthcoming shareholders' ity shareholders against expropriation of managers or dom- meeting but will not typically receive detailed information inant shareholders. We label these rights as "anti-director" on the items to be discussed. Only by going to the meet- rights. The first four anti-director rights measure how easy ing will they know what is discussed. In fact, attending the it is for an investor to exercise any voting rights that he meeting-or designating someone to do so in their place- may have. Shareholders in the United Kingdom will is the only way in which they can vote; proxy by mail is not receive proxy statements two weeks in advance of the allowed. Furthermore, announcing that they intend to vote shareholders' meeting with detailed information on the their shares will cause them to be blocked, making it items that are going to be discussed at the meeting. Should impossible for them to trade the shares in the days sur- they wish to vote, they do not need to show up in person rounding the meeting. At the meeting, shareholders vote at the meeting-they can mail their proxy vote instead. on the slate of directors proposed by management and are The shares of investors who have indicated that they will not allowed proportional representation on the board. participate in the shareholders' meeting will not be Investors in Mexican firms must have at least 33 percent of blocked in the days that surround the meeting; the free- share capital to have a resolution considered by the ESM. dom to trade shares around shareholders' meetings is an Fortunately, investors in Mexico do have a preemptive important right for people who may want to form alliances right that prevents dilution. Regrettably, this is the only to challenge management proposals. Directors are chosen right (of those that we collect) that shareholders in Mexico one at a time through a majority vote, and thus sharehold- have, since they do not have any legal recourse against the ers are not entitled to have proportional representation or decisions of the majority. To summarize, Table 4.2 paints a cumulative voting for directors. Our hypothetical investor very bleak picture of shareholder rights in Mexico. may have a resolution that he would like to be considered A convenient way of summarizing shareholder rights is by an extraordinary shareholders' meeting (ESM). If that is to aggregate anti-director rights into an index adding 1 if the case, he has the right to call an ESM if he owns 10 per- the corporate law protects minority shareholders, and a cent of the share capital. zero otherwise. For the case of the percentage of share cap- The next right in Table 4.2 measures the protection of ital needed to call an ESM, we give a I to those countries minority shareholders against a particular type of expro- where this percentage is at or below the world median of priation: issuing shares at favorable prices to, for example, 10 percent. When we add up these six anti-director rights associates of the controlling shareholders. In the United scores, the United Kingdom has a score of 5 while Mexico's Kingdom, shareholders have a preemptive right to buy score is only 1. new issues of stock of their holdings, and that right can be The comparison between Mexico and the United King- waived only by a shareholder vote. Finally, U.K. investors dom illustrates the broad findings of Table 4.2: Share- who feel they have been hurt by the decisions of the holder protection in common-law countries is significantly majority can seek redress through the courts. When the better than in French civil-law countries. While the inci- court believes that oppression has indeed taken place, it dence of one-share-one-vote rules, cumulative voting for may order that the oppressed members' shares be bought directors, and preemptive rights are not statistically differ- out at a fair price or that the firm remedy the matters at ent across English and French legal origins, the remaining issue. More generally, best-practice countries provide legal four measures show marked differences. Common-law mechanisms for the protection of oppressed minorities. To countries more frequently allow shareholders to exercise give just another example, a dissenting investor in Chile their vote by mail than French-origin countries (39 percent has the right to request-at the meeting-that the firm vs. 5 percent). No common-law country blocks shares buy back his shares at the market price prevailing before before shareholders' meetings, while 57 percent of French the meeting. civil-law countries do. On average, 9 percent of the share 71 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE 4.2 Shareholder Rights around the World This table classifies countries by legal origin. Definitions for each of the variables can be found in Table 4.1. Panel B reports the test of means for the different legal origins. PANEL A: SHAREHOLDER RIGHTS (I = INVESTOR PROTECTION IS IN THE LAW) SHARES NOT CUMULATIVE %, OF SHARE ONE PROXY BY BLOCKED VOTING/ CAPITAL PREEMPTIVE A.NTI- SHARE - MAIL BEFORE PROPORTIONAL TO CALL RIGHT OPPRESSED DIRECTOR COUNTRY ONE VOTE ALLOWED MEETING REPTN. AN ESM TO NEW ISSUES MINORITY RIGHTS Australia 0 1 1 0 0.05d 0 1 4 Canada 0 1 1 1 0.05 0 1 5 Hong Kong, China 0 1 1 0 0.10 1 1 5 India 0 0 1 1 0.10 1 1 5 Ireland 0 0 1 0 0.10 1 1 4 Israel 0 0 1 0 0.10 0 1 3 Kenya 0 0 1 0 0.10 0 1 3 Malaysia I 0 1 0 0.10 1 1 4 New Zealand 0 1 1 0 0.05 0 1 4 Nigeria 0 0 1 0 0.10 0 1 3 Pakistan 1 0 1 1 0.10 1 1 5 Singapore 1 0 1 0 0.10 1 1 4 South Africa 0 1 1 0 0.05 1 1 5 SriLanka 0 0 1 0 0.10 0 1 3 Thailand 0 0 1 1 0.20e 0 0 2 United Kingdom 0 1 1 0 0.10 1 1 5 United States 0 1 1 1 0.10 0 1 5 Zimbabwe 0 0 1 0 0.05 0 1 3 English origin avg. 0.17 0.39 1.00 0.28 0.09 0.44 0.94 4.00 Argentina 0 0 0 1 0.05 1 1 4 Belgium 0 0 0 0 0.20 0 0 0 Brazil 1 0 1 0 0.05 0 1 3 Chile 1 0 1 1 0.10 1 1 5 Colombia 0 0 1 1 0.25 1 0 3 Ecuador 0 0 1 0 0.25 1 0 2 Egypt, Arab Rep. 0 0 1 0 0.10 0 0 2 France 0 1 0 0 0.10 1 0 3 Greece 1 0 0 0 0.05 1 0 2 Indonesia 0 0 1 0 0.10 0 0 2 Italy 0 0 0 0 0.20 1 0 1 Jordan 1 0 1 0 0.25 0 0 1 Mexico 0 0 0 0 0.33 1 0 1 Netherlands 0 0 0 0 0.10 1 0 2 Perni 1 0 1 1 0.20 1 0 3 Philippines 0 0 1 1 open 0 1 3 Portugal 0 0 1 0 0.05 1 0 3 Spain 0 0 0 1 0.05 1 1 4 Turkey 0 0 1 0 0.10 0 0 2 Uruguay 1 0 0 0 0.20 1 1 2 Venezuela 0 0 1 0 0.20 0 0 1 Latin American avg. 0.44 0.00 0.67 0.44 0.18 0.78 0.44 2.67 Rest of French origin avg. 0.17 0.08 0.50 0.17 O.12 0.50 0.17 2.08 French origin avg. 0.29 0.05 0.57 0.29 0.15 0.62 0.29 2.33 72 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR TABLE 4.2 (Continued) PANEL A: SHAREHOLDER RIGHTS (I = INVESTOR PROTECTION IS IN THE LAW) SHARES NOT CUMULATIVE % OF SHARE ONE PROXY BY BLOCKED VOTING/ CAPITAL PREEMPTIVE ANTI- SHARE- MAIL BEFORE PROPORTIONAL TO CALL RIGHT OPPRESSED DIRECTOR COUNTRY ONE VOTE ALLOWED MEETING REPTN. AN ESM TO NEW ISSUES MINORITY RIGHTS Austria 0 0 0 0 0.05 1 0 2 Germany 0 0 0 0 0.05 0 01 Japan 1 0 1 1 0.03 0 1 4 Korea, Rep. of 1 0 0 0 0.05 0 1 2 Switzerland 0 0 0 0 0.10 1 0 2 Taiwan, China 0 0 0 1 0.03 0 1 3 German origin avg. 0.33 0.00 0.17 0.33 0.05 0.33 0-50 2.33 Denmark 0 0 1 0 0.10 0 0 2 Finland 0 0 1 0 0.10 1 0 3 Norway 0 1 1 0 0.10 1 0 4 Sweden 0 0 1 0 0.10e 1 0 3 Scandinavian origin avg. 0.00 0.25 1.00 0.00 0.10 0.75 0.00 3.00 Sample average 0.22 0.18 0.71 0.27 0.11 0.53 0.53 3.00 PANEL B: TESTS OF MEANS (T-STATISTICS) Common vs. civil law -0.72 3.03 a 497 0.15 1.48 -0.91 5 59 5.00 a English vs. French origin -0.87 2.82 a 3.87 a -0.05 -2.53 -1.08 5.45 a 473 a English origin vs. Latin America -1.57 3.29 a 2.00 -0.85 -3.56 a -1.67 3.44 2.98 French vs. German origin -0.22 1.00 -1.78' -0.22 2.64 b 1.23 -0.96 0.00 French vs. Scandinavian origin 2.83 b -1.37 -3.87 2.82 b 2.43 -0.48 2.83 -1.06 Rest of French origin vs. Latin America -1.39 1.00 -0.74 -1.39 -1.71 -1.29 -1.39 -1.11 a = Significant at 1% level; b = Significant at 59 level; c r Significant at 10% level; d = as a percentage ofvores; e = as a percentage ofthe nunber ofshares capital is sufficient to call an ESM in common-law coun- one-share-one-vote (44 percent vs. 17 percent), is less tries, whereas 15 percent of share capital is required in likely to block shares (67 percent vs. 50 percent), has a French civil-law nations. Finally, 94 percent of common- higher incidence of proportional representation (44 percent law countries have an oppressed minority mechanism in vs. 17 percent), is more likely to grant preemptive rights place, while only 29 percent of French-origin countries do. (78 percent vs. 50 percent), and has a higher incidence of The differences between English- and French-origin coun- oppressed minority remedies (44 percent vs. 17 percent). tries are captured in the anti-director's index, which has an On the other hand, Latin America never allows proxy by average of 4.00 for common-law countries and only 2.33 mail (vs. 8 percent for the rest of French origin) and for French civil-law nations (t-statistic of 4.73). requires a higher fraction of the share capital to call for an It is important to note that Mexico actually scores lower ESM (18 percent vs. 12 percent). With the exception of the than the rest of Latin America when it comes to share- percentage needed to call an ESM, these differences are not holder rights, so our choice of Mexico for this example does statistically significant when taken in isolation. Although not mean that Mexico is typical of Latin America. In fact, not statistically significant, differences add up to margin- Latin America generally scores a little higher than the aver- ally better shareholder rights in Latin America than in the age of the rest of the French-origin countries in many rest of the French civil origin when rights are aggregated shareholder rights. Latin America has a higher incidence of in the anti-director index (2.67 vs. 2.08). However, Latin 73 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER America's anti-director rights index is statistically signifi- Suppose that a debtor to whom the creditor has lent cantly lower than that of common-law countries. money files a petition for reorganization in a London court. German civil-law countries share the French origin lack The court will then notify the creditor, who will have two of protection of shareholder rights. Although German-ori- weeks to oppose reorganization. A secured creditor who gin countries have a significantly higher incidence of chooses to oppose a reorganization petition has the right to oppressed minority mechanisms, they block shares more appoint a so-called trustee to decide what will happen to often than French countries do. The average anti-director the firm. The important thing is that the debtor does not scores for the German and French families are the same have the right to unilaterally file for reorganization. Even (2.33). Finally, Scandinavian-origin countries, although if the borrower has his petition for reorganization accepted, clearly inferior to common-law countries in shareholder pro- there is not an "automatic stay" that prevents secured cred- tection, are the best within the civil-law tradition. The aver- itors from gaining access to their collateral. In addition, age Scandinavian anti-director rights score is 3. In short, rel- secured creditors who choose not to withdraw their collat- ative to the rest of the world, common-law countries have eral are paid first in the event that reorganization fails and the package of laws most protective of shareholders. liquidation ensues. Finally, the bargaining position of cred- itors is strengthened by the fact that pending the resolu- Creditor Rights tion of the bankruptcy procedure, the old management In principle we would like to measure the ability of credi- team will not continue to run the firm. Rather, a trustee tors to use the law to force companies to meet their credit appointed by the creditors would be in charge of the firm's commitments. In practice, creditor rights are difficult to day-to-day operations. assess for two main reasons. First, most countries have in Now suppose that a debtor to whom the creditor has place both reorganization and liquidation procedures that lent money files a petition for reorganization in a Mexico are used with varying frequency and may confer different City court. Creditors have no say in whether the firm's reor- levels of protections to creditors. Therefore, a country may ganization petition is accepted or declined. But if the peti- be very protective of creditors if it offers, for example, tion is accepted, secured creditors are not able to pull their strong rights in liquidation and weak protection in reorga- collateral out of the firm; an "automatic stay" is triggered nization, provided that the reorganization procedure is sel- by the acceptance of the reorganization petition. Secured dom used. Second, creditors, unlike shareholders, do not creditors have additional worries, because if liquidation have a homogeneous claim against the firm-i.e., they dif- takes place, they are not paid first. Rather the state and the fer in the priority of their claim. As a result, it is possible firm's employees take priority. The creditors' predicament that measures that favor some creditors (e.g., unsecured is aggravated by the fact that the debtor not only will write creditors) may hurt others (e.g., secured creditors). the reorganization proposal, but will continue to run the To undertake a cross-country analysis of creditor rights, firm pending the resolution of the bankruptcy procedure, we score creditor rights in both reorganization and liqui- which may take several years. dation, and add up the scores to create a creditor rights As with shareholders, one way to summarize the differ- index, in part because almost all countries rely to some ence in creditor rights across countries is to create an index extent on both procedures. In assessing creditor rights that adds 1 when the pro-investor right is granted by law, below, we also take the perspective of senior secured credi- and zero otherwise. This index is shown in the last column tors, in part for concreteness, and in part because much of of Table 4.3 and takes a value of 4 for the United Kingdom the debt in the world has that character. Creditor rights for and zero for Mexico. Again, as with shareholder rights, the the cross-section of 49 countries in the sample can be found picture for creditor rights in Mexico is substantially in Table 4.3. Once again, to illustrate differences between bleaker than in the United Kingdom. English common-law countries and French civil-law coun- Although the Mexico-U.K. comparison is extreme, com- tries, we describe the data by comparing the rights of an mon-law countries in general offer creditors stronger legal investor who has a credit against a firm incorporated in the protections against managers. Table 4.3 shows that all four United Kingdom versus the rights of an investor with a measures of creditor rights are weaker for countries of credit against a firm incorporated in Mexico. French legal origin by an amount that is statistically signif- 74 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR icant. A total of 72 percent of common-law countries place least. German and Scandinavian civil-law countries are in restrictions on managers seeking court protection from the middle. The one exception is the strong protections creditors, while only 42 percent of French civil-law nations that German civil-law countries afford secured creditors. do. The incidence of having no automatic stay on assets is 72 percent in common-law countries versus only 26 percent Enforcement of Laws in French civil-law nations. Relatively fewer countries of Legal rules are only one element of investor protection; the French legal origin (65 percent) assure that secured credi- enforcement of these rules may be equally or even more tors are paid first than do countries of English legal origin important. If good laws are not enforced, they cannot be (89 percent). Finally, only 26 percent of French civil-law effective. Likewise, investors may enjoy high levels of pro- countries remove managers in reorganization, compared tection despite bad laws if an efficient judiciary system can with 78 percent of countries of the common-law family. In redress expropriations by management. In this way, strong brief, the average aggregate creditor rights score is 3.11 for legal enforcement may serve as a substitute for weak rules. English origin and a mere 1.58 for French origin. Table 4.4 presents several proxies for the quality of Unlike the case of shareholder rights, Latin America enforcement of laws in different countries. These measures offers considerably less legal protection to creditors than the are collected by private credit-risk agencies for the use of rest of the French civil-law countries. When compared foreign investors interested in doing business in the respec- with the other countries of French civil-law origin, coun- tive countries (Business International Corporation, Politi- tries in Latin America are less likely to place restrictions for cal Risk Services). We use three measures: efficiency of the going into reorganization (38 percent vs. 45 percent), have judicial system, rule of law, and corruption. The first two no-automatic-stay policies (13 percent vs. 36 percent), pay of these proxies pertain to law enforcement, while the last secured creditors first (56 percent vs. 73 percent), and pre- one captures the government's general attitude toward vent management from remaining in office (25 percent vs. business. In addition to these measures, we also collected 27 percent). However, these differences are not statistically data on the quality of accounting standards of publicly significant even when aggregated in the creditor-rights traded firms in different countries. Accounting is central to index (1.25 vs. 1.81). corporate governance, as it may be difficult to assess man- Latin America also scores lower than the other two civil- agement performance without reliable accounting stan- law families. German legal origin countries are relatively dards. More broadly, cash flows may be very difficult to more pro-creditor than Latin American and French civil- verify in countries with poor accounting standards; conse- law countries, averaging an aggregate score of 2.33. The quently, the menu of financial contracts available to differences between German- and French-origin countries investors may be substantially narrower in such countries. are particularly significant in liquidation measures: 67 per- The index of accounting standards in Table 4.4 is provided cent of German civil-law countries have no automatic stay by the Center for International Financial Analysis and and always pay secured creditors first. Research based on examination of company reports of firms Finally, countries of Scandinavian origin always pay in each country. It is available for 41 of the 49 countries in secured creditors first, but always allow management to stay our sample. pending reorganization. In three out of four cases they We can begin the discussion of this data by focusing on impose an automatic stay on assets and place restrictions to the Latin American average. Compared with the English- go into reorganization. As a result, the aggregate creditor- origin average, Latin America has very weak legal institu- rights index for countries of Scandinavian legal origin has a tions and accounting standards. A corrupt or inefficient value of 2.00-a difference that is not statistically signifi- legal system coupled with poor disclosure standards could cant from the 1.58 value for countries of French legal origin. render legal rules ineffective. To summarize the results thus far, bankruptcy laws dif- While the Latin American average across all enforce- fer a great deal across countries. In particular, they differ ment variables is below the French-origin average, it turns because they come from different legal families. Relatively out that the French civil-law family shares Latin America's speaking, common-law countries protect creditors the weak legal-enforcement mechanisms. The French family most, and French civil-law countries protect them the has the weakest quality of legal enforcement and account- 75 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE 4.3 Creditor Rights around the World This table classifies countries by legal origin. Definitions for each variable can be found in Table 4.1. Panel B reports tests of means for the differenr legal origins. PANEL A: CREDITOR RIGHTS (I = CREDITOR PROTECTION IS IN THE LAW) RESTRICTIONS SECURED MANAGEMENT FOR GOING INTO NO AUTOMATIC CREDITORS DOES NOT STAY CREDITOR COUNTRY REORGANIZATION STAY ON ASSETS FIRST PAID IN REORGANIZATION RIGFITS Australia 0 0 1 0 1 Canada 0 0 1 0 1 Hong Kong, China 1 1 1 1 4 India 1 1 1 1 4 Ireland 0 0 1 0 1 Israel I I I 1 4 Kenya 1 1 1 1 4 Malaysia I 1 1 1 4 New Zealand 1 1 0 1 3 Nigeria 1 1 1 1 4 Pakistan 1 1 1 1 4 Singapore 1 1 1 1 4 South Africa 1 0 1 1 3 Sri Lanka 1 1 0 1 3 Thailand 0 1 1 1 3 United Kingdom 1 1 1 1 4 United States 0 0 1 0 1 Zimbabwe 1 1 1 1 4 English origin avg. 0.72 0.72 0.89 0.78 3.11 Argentina 0 0 1 0 1 Belgium 0 1 1 0 2 Brazil 1 0 0 0 1 Chile 1 0 1 0 2 Colombia 0 0 0 0 0 Ecuador 1 1 1 1 4 Egypt, Arab Rep. 1 1 1 1 4 France 0 0 0 0 0 Greece 0 0 0 1 1 Indonesia I 1 1 1 4 Italy 1 0 1 0 2 Jordan n.a. n.a. n.a. n.a. n.a. Mexico 0 0 0 0 0 Netherlands 1 0 1 0 2 Peru 0 0 0 0 0 Philippines 0 0 0 0 0 Porrugal 0 0 1 0 1 Spain 0 1 1 0 2 Turkey 1 0 1 0 2 Uruguay 0 0 1 1 2 Venezuela n.a. n.a. 1 n.a. n.a. Latin American avg. 0.38 0.13 0.56 0.25 1.25 Rest of French origin avg. 0.45 0.36 0.73 0.27 1.81 French origin avg. 0.42 0.26 0.65 0.26 1.58 76 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR TABLE 4.3 (Continued) PANEL A: CREDITOR RIGHTS (1 = CREDITOR PROTECTION IS IN THE LAW) RESTRICTIONS SECURED MANAGEMENT FOR GOING INTO NO AUTOMATIC CREDITORS DOES NOT STAY CREDITOR COUNTRY REORGANIZATION STAY ON ASSETS FIRST PAID IN REORGANIZATION RIGHTS Austria I 1 1 0 3 Germany 1 1 1 0 3 Japan 0 0 1 1 2 Korea, Rep. of 0 1 1 1 3 Switzerland 0 0 1 0 1 Taiwan, China 0 1 1 0 2 German origin avg. 0.33 0.67 1.00 0.33 2.33 Denmark 1 1 1 0 3 Finland 0 0 1 0 1 Norway 1 0 1 0 2 Sweden 1 0 1 0 2 Scandinavian origin avg. 0.75 0.25 1.00 0.00 2.00 Sample average 0.55 0.49 0.81 0.45 2.30 PANEL B: TESTS OF MEANS (T-STATISTICS) Common vs. civil law 1.86 C 2.65 a 1.04 4.13 a 3.61 a English vs. French origin 1.89 C 3.06 a 1.75b 3. 55C 3.61 a English origin vs. Latin America 1.71 C 3.25 a 2.04 b 2.83 a 3.42 a French vs. German origin 0.37 -1.85 ' -3.20 a -0.32 -1.29 French vs. Scandinavian origin -1.18 0.05 -3.20 a 2.54 -0.60 Rest of French origin vs. Larin America 0.33 1.14 0.77 0.11 0.90 a = Significant ar 1% level; b = Significant at 5% level; c = Significant at 10%, level. ing standards. Scandinavian countries have the strongest itor rights may actually understate the severity of the cor- enforcement mechanisms, with German civil-law and com- porate governance problem in the region. mon-law countries close behind. Common-law countries, although behind Scandinavian nations, are still ahead of External Finance and Legal Institutions the French civil-law countries. Note that rule of law is the There are at least two reasons why legal institutions may only measure where differences in means between common have no effect on the pattern of external financing of firms. law and French legal origin are not statistically significant. First, laws may not be necessary to support external financ- These results do not support the conclusion that the ing if, for example, companies deliver on their promises quality of law enforcement substitutes or compensates for not because they are forced to but because they want to the quality of laws. An investor in Latin America-and build a good reputation to facilitate their access to capital more generally in a French civil-law country-is poorly markets (Diamond 1989, 1991; Gomes 1996). Reputation protected by both the laws and the system that enforces unravels if there is ever a time when the gains from cheat- them. The converse is true for an investor in a common-law ing exceed the value of keeping external financing open, country, on average. Poor enforcement and accounting since investors, through backward induction, would never standards aggravate, rather than cure, the difficulties faced extend financing to such a firm to begin with. by investors in French civil-law countries.' The weak Second, poor laws and their enforcement may have no scores obtained by Latin America in shareholder and cred- real consequences if firms can easily opt out of the laws of 77 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE 4.4 Enforcement of Laws This table classifies countries by legal origin. Definitions for each of the variables can be found in Table 4.1. Panel B reports the tests of means for the differ- ent legal origins. PANEL A ENFORCEMENT VARIABLES ACCOUNTING GNP PER CAPITA RATING ON EFFICIENCY OF ACCOUNTING COUNTRY JUDICIAL SYSTEM RULE OF LAW CORRUPTION STANDARDS (U.S.$) English origin avg.' 8.15 6.46 7.06 69.62 9,353 French origin avg.b 6.56 6.05 5.84 51.17 7,102 Latin American avg. 6.47 5.18 5.22 46.25 3,1077 Rest of French origin avg. 6.62 6.70 6.30 55.10 10,121 German origin avg.' 8.54 8.68 8.03 62.67 22,0)67 Scandinavian origin avg.d 10.00 10.00 10.00 74.00 24,185 Sample average 7.67 6.85 6.90 60.93 11,156 a=Counrries surveyed: Australia, Canada, Hong Kong (China), India, Ireland, Israel, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Sri Lanka, Thailand, United Kingdom, Unired States, Zimbabwe. b=Countries surveyed: Argentina, Belgium, Brazil, Chile, Colombia, Ecuador, Arab Republic of Egypt, France, Greece, Indonesia, Italy, Jordan, Mexico, Netherlands, Peru, Philippines, Portugal, Spain, Turkey, Umguay, Venezuela. c=Countries surveyed: Austria, Germany, Japan, Republic of Korea, Switzerland, Taiwan (China). d=Countries surveyed: De-nark, Finland, Norway, Sweden. PANEL B: TESTS OF MEANS (T-STATISTICS) Common vs. civil law 1.27 -0.77 0.39 3.12 a -0.94 English vs. French origin 2.65 a 0.51 1.79 c 4.66 a 0.°5 English origin vs. Latin America 2.37 1.25 2.33 b 6.53 a 2.()8 French vs. German origin -2.53 a -2.55 r -2.49 a -2.10 -3.79 French vs. Scandinavian origin -9.34a -20.80 a9.77 a -3.32 a -4.28' Rest of French origin vs. Latin America 0.19 1.52 1.28 1.49 2.27 a = Significanr at 1% level; b = Significant at 5% level; c = Significant at 10% level. their legal jurisdictions. Easterbrook and Fischel (1991) are same time, entrepreneurs are going to be more reluctant to skeptical that legal rules are binding in most instances, offer shares at discounted prices, thus resulting in higher since entrepreneurs can offer better investor rights, when it ownership concentration as well as smaller and narrower is optimal to do so, through corporate charters that effec- markets for external equity.2 Similarly, bad creditor rights tively serve as contracts between entrepreneurs and may have analogous price and quantity effects on debt mar- investors. In practice, however, opting out may be costly kets. In other words, if laws do not protect the rights of both for firms that need to write non-standard contracts creditors, debt markets may be small since creditors may and for investors who need to study them. In addition, demand high interest rates and firms may be reluctant to courts may be unwilling or unable to enforce non-standard borrow from arm's-length sources in such conditions. contracts, further limiting the scope for opting out. Ultimately, the question of whether legal institutions Alternatively, if legal institutions matter, ownership matter is fundamentally empirical: If opting out were concentration should be higher in countries with poor cheap and simple, the patterns of ownership and external investor protection than in countries with strong protec- finance of firms would not be affected by differences in tions for investors for at least two reasons: First, agency legal institutions across countries.3 Accordingly, in this problems may call for large shareholders to monitor man- section, we examine two types of evidence regarding the agers and thus prevent or minimize expropriation. Second, influence of legal institutions on external finance: owner- minority shareholders may be unwilling to pay high prices ship concentration, and the size and breadth of capital mar- for securities in countries with weak legal protection. At the kets. Table 4.5 summarizes the results. 78 INSTITUTIONAL REFORM IN MARKETS THE CASE OF THE FINANCIAL SECTOR Data the stock market data that we use to measure external We describe sequentially our measures of ownership con- equity financing does not exist for debt markets. However, centration, external equity financing, and debt markets. the fact that our debt measure includes not only corpora- First, to measure ownership concentration, La Porta et al. tions but the whole private sector may actually be an (1998a) assembled data for the 10 largest publicly traded, advantage, because in many countries entrepreneurs raise non-financial private domestic firms in each of 45 coun- money on their personal accounts to finance their firms (for tries. For each country we measure ownership concentra- example, by mortgaging their properties). tion as the median percentage owned by the three largest shareholders in each of these 10 firms. Results Second, as in La Porta et al. (1997a), we also use three The first striking result of Table 4.5 is that in the world as measures of equity finance. The first measure is the 1994 a whole, dispersed ownership is a myth: In an average ratio of external equity finance to GNP in each country. To median firm 45 percent of the common shares are held by compute a rough proxy of external equity finance, we mul- the largest three shareholders. The second result is that tiply the total market value of common stock of all pub- those countries with weaker investor protections have licly traded firms by the average fraction of the equity not larger share ownership concentration. In particular, coun- held by the largest three investors (i.e., the complement of tries of the French legal family have an average ownership the ownership variable just described). We scale the total concentration of 55 percent. Statistically this number is market value of common stock by the fraction of equity significantly higher than the mean of the rest of the world held by minority shareholders to avoid overestimating the and of each of the other three legal families individually. availability of external financing. For example, when 90 Like the rest of the French origin, Latin America has percent of a firm's equity is held by insiders, looking at the highly concentrated ownership. With the exception of market capitalization of the whole firm gives a tenfold Chile, which has strong shareholder rights, all Latin Amer- overestimate of how much has actually been raised exter- ican countries in the sample have higher ownership concen- nally. The procedure we follow may still overestimate the tration than the world mean. After Greece (68 percent), the level of external financing, because our ownership concen- three largest ownership concentration levels in the world tration figures are based on the largest firms and because are found in Colombia (68 percent), Mexico (67 percent), they ignore cross-holdings. Still, this procedure is concep- and Brazil (63 percent). In sum, these data indicate that tually better than looking at the ratio of market capitaliza- Latin American countries-and the French civil-law coun- tion to GNP.4 tries in general-have unusually high ownership concentra- The remaining two measures of external equity finance tion, possibly as an adaptation to weak legal protection. capture market breadth. The first is the number of domes- Several interesting patterns emerge from looking at our tic firms listed in the stock exchange of each country rela- proxies for external equity finance on Table 4.5. First, tive to its population. The second is the number of initial access to external equity financing is most limited in public offerings of shares in each country between mid- French civil-law countries. Specifically, both the ratio of 1995 and mid-1996 (the period for which we have been external capital to GNP and the ratio of domestic firms to able to obtain the data), also relative to the population. We population are roughly half the world mean, whereas the look at both the stock and flow of new companies obtain- ratio of IPOs to population is roughly one-fifth of the ing equity financing since the development of financial world mean. Equity markets are particularly narrow in markets has accelerated greatly in the last decade, and Latin America; the ratio of the number of firms to popula- hence the IPO data provides a more recent picture of exter- tion is roughly one-third of the world mean, whereas the nal equity financing. ratio of the number of IPOs to population is more than 10 Finally, also as in La Porta et al. (1997a), we measure the times smaller than the world mean. In contrast, all three availability of debt financing in each country as the ratio of equity measures indicate that, on average, access to exter- the sum of private-sector bank debt and corporate bonds nal equity is easiest in common-law countries: The ratio of outstanding to GNP. Our choice of debt variable is partly outsider-held stock market to GNP is 60 percent, vs. 40 determined by data availability, because the analogue of percent for the world mean; the number of listed firms per 79 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE 4.5 Extermal Finance and Legal Insfitutions This table classifies countries by legal origin. Definitions for each of the variables can be found in Table 4.1. Panel B reports tests of means for the diFferent legal origins. PANEL A OWNERSHIP CONCENTRATION EXTERNAL (10 LARGE FIRMS) CAP / GNP DOMESTIC GDP LOG COUNTRY (MEDIAN) (MEAN) FIRMS / POP IPOS / POP DEBT / GNP GROWTH GNP Australia 0.28 0.49 63.55 - 0.76 3.06 12.64 Canada 0.24 0.39 40.86 4.93 0.72 3.36 13.26 Hong Kong, China 0.54 1.18 88.16 5.16 - 7.57 11.56 India 0.43 0.31 7.79 1.24 0.29 4.34 12.50 Ireland 0.36 0.27 20.00 0.75 0.38 4.25 10.73 Israel 0.55 0.25 127.60 1.80 o.66 4.39 11.19 Kenya - - 2.24 - - 4.79 8.83 Malaysia 0.52 1.48 25.15 2.89 0.84 6.90 11.00 New Zealand 0.51 0.28 69.00 0.66 0.90 1.67 10.69 Nigeria 0.45 0.27 1.68 - - 3.43 10.36 Pakistan 0.41 0.18 5.88 - 0.27 5.50 10.88 Singapore 0.53 1.18 80.00 5.67 0.60 1.68 11.68 South Africa 0.52 1.45 16.00 0.05 0.93 7.48 10.92 SriLanka 0.61 0.11 11.94 0.11 0.25 4.04 9.28 Thailand 0.48 0.56 6.70 0.56 0.93 7.70 11.72 United Kingdom 0.15 1.00 35.68 2.01 1.13 2.27 13.86 United States 0.12 0.58 30.J1 3.11 0.81 2.74 15.67 Zimbabwe 0.51 0.18 5.81 - - 2.17 8.63 English origin avg. 0.42 0.60 35.45 2.23 0.68 4.30 11.41 French origin Argentina 0.55 0.07 4.58 0.20 0.19 1.40 12.40 Belgium 0.62 0.17 15.50 0.30 0.38 2.46 12.29 Brazil 0.63 0.18 3.48 0.00 0.39 3.95 13.03 Chile 0.38 0.80 19.92 0.35 0.63 3.35 10.69 Colombia 0.68 0.14 3.13 0.05 0.19 4.38 10.82 Ecuador - - 13.18 0.09 - 4.55 9.49 Egypt, Arab Rep. 0.62 0.08 3.48 - - 6.13 10.53 France 0.24 0.23 8.05 0.17 0.96 2.54 14.07 Greece 0.68 0.07 21.60 0.30 0.23 2.46 11.25 Indonesia 0.62 0.15 1.15 0.10 o.42 6.38 11.84 Italy 0.60 0.08 3.91 0.31 0.55 2.82 13.94 Jordan - - 23.75 - 0.70 1.20 8.49 Mexico 0.67 0.22 2.28 0.03 0.47 3.07 12.69 Netherlands 0.31 0.52 21.13 0.66 1.08 2.55 12.68 Peru 0.57 0.40 9.47 0.13 0.27 2.82 10.92 Philippines 0.51 0.10 2.90 0.27 0.10 0.30 10.44 Portugal 0.59 0.08 19.50 0.50 0.64 3.52 11.41 Spain 0.50 0.17 9.71 0.07 0.75 3.27 13.19 Turkey 0.58 0.18 2.93 0.05 0.15 5.05 12.08 Uruguay - - 7.00 0.00 0.26 1.96 9.40 Venezuela 0.49 0.08 4.28 0.00 0.10 2.65 [0.99 Latin American avg. 0.57 0.23 7.49 0.10 0.29 2.84 :11.11 Rest of French origin avg. 0.53 0.19 11.89 0.28 0.56 3.43 111.89 French origin avg. 0.55 0.21 10.00 0.19 0.45 3.18 11.55 80 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR TABLE 4.5 (Continued) PANEL A OWNERSHIP CONCENTRATION EXTERNAL (10 LARGE FIRMS) CAP / GNP DOMESTIC GDP LOG COUNTRY (MEDIAN) (MEAN) FIRMS / POP IPOS / POP DEBT I GNP GROWTH GNP Austria 0.51 0.06 13.87 0.25 0.79 2.74 12.13 Germany 0.50 0.13 5.14 0.08 1.12 2.60 14.46 Japan 0.13 0.62 17.78 0.26 1.22 4.13 15.18 Korea, Rep. of 0.20 0.44 15.88 0.02 0.74 9.52 12.73 Switzerland 0.48 0.62 33.85 - - 1.18 12.44 Taiwan, China 0.14 0.88 14.22 0.00 - 11.56 12.34 German origin avg. 0.33 0.46 16.79 0.12 0.97 5.29 13.21 Scandinavian origin Denmark 0.40 0.21 50.40 1.80 0.34 2.09 11.84 Finland 0.34 0.25 13.00 0.60 0.75 2.40 11.49 Norway 0.31 0.22 33.00 4.50 0.64 3.43 11.62 Sweden 0.28 0.51 12.66 1.66 0.55 1.79 12.28 Scandinavian origin avg. 0.33 0.30 27.26 2.14 0.57 2.42 11.80 Sample average 0.45 0.40 21.59 1.02 0.59 3.79 11.72 PANEL B: TESTS OF MEANS (T-STATISTICS) Common vs. civil law -0.91 3.12' 3.16a 3.97a 1.33 1.23 -1.06 English vs. French origin -2.68 3.29' 3.16' 4.50' 2.29b 1.97c -0.28 English origin vs. Latin America -2.34b 1.97' 2.29b 3.21a 3.42' 1.93b 0.46 French vs. German origin 3.29a -2.38b -1.85 0.78 -3.39a -1.96' -2.48 French vs. Scand. origin 3.32' -0.91 -3.31' -5.45' 0.82 0.97 -0.33 Rest of French origin vs. Latin America -0.54 -0.35 1.36 2.41b 2.20" 0.88 1.22 a = Signficant at 1% level; b = Significant at 5% level; c = Significant at 10% level. 1 million people is 3 5, vs. 21.6 for the world mean; and the gin. However, German civil-law countries are a mystery. number of IPOs per million people is 2.2, vs. 1.02 for the We conjecture that a possible explanation of the German- world mean. Finally, equity markets in countries of Scan- origin anomaly is that firms in both Germany and Japan dinavian origin are smaller but broader than in countries of have large liquid assets and, therefore, our debt measure German origin. To summarize, external equity markets overstates their true liabilities (Rajan and Zingales 1995). line up pretty well with shareholder rights and legal insti- tutions: They are smallest in French civil-law countries and Regression Results largest in common-law countries. We present two sets of regressions on Tables 4.6 and 4.7 for The last column in Table 4.5 shows the aggregate debt each of our measures of ownership concentration, external measure. The ratio of total debt to GNP is 45 percent for equity finance, and debt markets. The first set of regres- French civil-law countries, 57 percent for Scandinavian sions (Table 4.6) includes legal-origin dummies, whereas countries, 68 percent for common-law countries, and 97 the second one (Table 4.7) includes our measures of share- percent for German countries. It is also interesting to note holder and creditor rights. In both specifications we regress that Latin America, with a ratio of total debt to GNP of our measures of capital markets on two control variables only 29 percent, is an outlier even within the French civil- and on law enforcement. The first control is the growth of law family. Low creditor rights line up with small markets GDP on the theory that growth affects valuation and that when we compare French, Scandinavian, and English orn- in turn may affect ownership patterns as entrepreneurs are 81 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE 4.6 Regressions of External Finance and Legal Origin Ordinary least squares regressions of the cross section of 49 countries around the world. The dependent variables are: (1) Ownership Concentration (the mean of each country); (2) External Cap / GNP; (3) Domestic Firms / Pop; (4) IPOs / Pop; and (5) Debt / GNP. The independent variables are (1) GDP growth; (2) Log GNP; (3) Rule of law; (4) French origin; (5) German origin; and (6) Scandinavian origin. Robust standard errors are shown in parentheses. INDEPENDENT VARIABLES DEPENDENT VARIABLES: OWNERSHIP EXTERNAL DOMESTIC CONCENTRATION CAP / GNP FIRMS / POP IPOS / POP DEBT / GNP GDP growth -0.0077 0.0584S 1.0111 0.1938 0.0251 (0.0084) (0.0259) (1.3676) (0.1112) (0.0148) Log GNP -0.0436a 0.0038 -2.9127 -0.0662 0.0370 (0.0132) (0.0420) (1.9117) (0.1193) (0.0281) Rule of law -0.0031 0.0417 4.8422 0.2122 b 0.0698 (0.0067) (0.0272) (1.4708) (0.0926) (0.0163) French origin 0.1217 a -0.3225 a -21.9070 a -1.5982 a -0.15 16 (0.0317) (0.1131) (7.9944) (0.3902) (0.0817) German origin -0.0013 -0.2962 a -25.1485 a -2.8119 a 0.1080 (0.0660) (0.1629) (9.1683) (0.6257) (0.1116) Scandinavian origin -0.0589 -0.3391 b -22.2680 -0.3123 -0.2764 (0.0434) (0.1494) (10.9897) (0.9516) (0.1145) Intercept 0.9889 a 0.0336 33.0486 -0.9201 -0.3496 (0.1314) (0.4001) (22.2848) (1.4532) (0.2786) Observations 45 45 49 41 39 Adjusted R' 0.5517 0.3840 0.3497 0.5671 0.6647 a = Significant at 1% level; b = Significant at 5% level; c = Significant at 10% more willing to issue at attractive prices. We also control three shareholders than the average country of E]nglish for the logarithm of total GNP on the theory that the cre- legal origin. ation of capital markets may be an activity subject to The first regression in Table 4.7 has the same controls increasing returns to scale. If this theory is true, we should and legal enforcement variables as Table 4.6, but instead of observe that larger economies have larger firms, which legal origin it tests for the significance of stronger share- might therefore have lower ownership concentration.' As a holder protection in the form of more anti-director rights measure of quality enforcement we chose the "rule of law" and the existence of one share-one vote rules. Looking at index, but the results we present are representative of other Table 4.7, the coefficient on the logarithm of GNP remains specifications with alternative enforcement measures. significant, showing that larger economies have less coni- The regression results in the first column of Table 4.6 centrated ownership. The results also show that legal show that larger economies have lower ownership concen- enforcement significantly reduces the concentration of tration and that, although not significant in this specifica- ownership in the regression. A 4.65 point increase in the tion, better enforcement leads to lower ownership concen- rule-of-law score (roughly the distance between New tration. In addition, this regression confirms the sharply Zealand and Argentina or Mexico) reduces average owner- higher concentration of ownership in the French civil-law ship concentration by 7 percentage points. Similarly, coun- countries: Controlling for other variables, the average tries with stronger shareholder protection, measured by country of French legal origin has 12 percentage points our aggregate score of anti-director rights, also have a sta- more concentrated ownership in the hands of the largest tistically significantly lower concentration of ownership. A 82 INSTITUTIONAL REFORM IN MARKETS: THE CASE OF THE FINANCIAL SECTOR TABLE 4.7 Regressions of External Finance and Shareholder and Creditor Rights Ordinary least squares regressions of the cross section of 49 countries around the world. The dependent variables are: (1) Ownership Concentration (the mean of each country); (2) External Cap / GNP: (3) Domestic Firms / Pop; (4) IPOs / Pop; and (5) Debt ! GNP. The independent variables are (1) GDP growth; (2) Log GNP; (3) Rule of law; (4) Anti-directors rights; (5) One share - One vote; and (6) Creditor rights. Robust standard errors are shown in parentheses. INDEPENDENT VARIABLES DEPENDENT VARIABLES: OWNERSHIP EXTERNAL DOMESTIC CONCENTRATION CAP / GNP FIRMS / POP IPOS / POP DEBT / GNP GDP growth -0.0124 0.0604a 1.3926 0.1433 0.0311' (0.0097) (0.0176) (1.3976) (0.1251) (0.0184) Log GNP -0.0312' -0.0205 -4.7687' -0.1814 0.0667 (0.0116) (0.0306) (1.7815) (0.1541) (0.0270) Rule of law -0.0151' 0.0456 4.8174' 0.2824 0.0615' (0.0060) (0.0214) (1.4273) (0.0887) (0.0142) Anti-director rights -0.0385' 0.1244, 6.0688' 0.5761' (0.0385) (0.0378) (1.6293) (1.438)) One share - One vote 0.0044 0.1433 0.4189 0.0226 (0.0461) (0.1278) (7.0938) (0.5267) Creditor rights 0.0518' (0.0287) Intercept 1.1041' -0.3219 20.9494 -1.1172 -0.8622' (0.1304) (0.2614) (17.3374) (1.4796) (0.2763) Observations 45 45 49 41 39 Adjusted R2 0.4862 0.4549 0.2671 0.4541 0.5993 a = Significant at 1% level; b = Significant at 5% level; c = Significant at 10% 1.6 point increase in the anti-director rights score (roughly deviation of the variable is 25) firms per million popula- the distance between common-law and French civil-law tion, and a further 0.67 (the standard deviation of the vari- averages) reduces ownership concentration by 6 percentage able is 1.5) IPOs per million population. However only the points. Finally, the existence of a one share-one vote rule in last two results are statistically significant. Second, legal the corporate law turns out not to be significant for owner- origin matters. Relative to common-law countries, French ship concentration. In sum, regression results confirm our civil-law countries have a ratio of external equity-to-GNP previous finding that the protection of shareholders 32.2 percentage points lower, 21.9 fewer publicly traded through legal institutions is an important determinant of firms per million population, and 1.6 fewer IPOs per mil- ownership concentration. lion population. German and Scandinavian legal origin are As with ownership concentration, we present two sets of also associated with smaller and more narrow stock mar- regressions for external equity markets on Tables 4.6 and kets than English origin but the effects are not as pro- 4.7. Two key results emerge from the analysis of the regres- nounced as with French legal origin. All estimates are sta- sions that use legal origin (Table 4.6). First, rule of law has tistically significant with the sole exception of the effect of a large impact on all three variables: A move from the Scandinavian origin on the number of IPOs per million world mean of 6.85 to a perfect score of 10 is associated population. with an increase of 13.1 percentage points (the standard The results on anti-director rights and one share-one deviation of the variable is 37) in the ratio of external mar- vote (Table 4.7) are easy to summarize. One share-one vote ket capitalization to GNP, an additional 15.3 (the standard has the expected sign, but it is never significant. In con- 83 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER trast, anti-director rights has a large impact on equity Conclusion and Policy Implications financing both in statistical and economical terms: A move In this chapter we have surveyed the evidence on laws gov- from the world mean of 2.5 to a perfect score of 6 is associ- erning investor protection, the quality of enforcement of ated with an increase of 43.5 percentage points in the ratio these laws, and their effect on the availability of external of external market capitalization to GNP, an additional financing on a sample of 49 countries. The analysis sug- 21.2 publicly traded firms per million population, and 2.0 gests three broad conclusions. additional IPOs per million population. Finally, as in previ- First, investors in different legal jurisdictions have very ous regressions, rule of law has a large impact on equity different bundles of rights. Therefore, investor rights are financing, and it is now statistically significant everywhere. not inherent to securities but rather are determined by Overall, the results on external equity finance in Tables laws. In particular, French civil-law countries protect 4.6 and 4.7 show that rule of law and shareholder rights investors the least and common-law countries the most. have a large impact on the availability of external equity Countries of German and Scandinavian legal origin take an financing. With the exception of the number of IPOs per intermediate stance toward investor rights. million population in Scandinavia, the regressions also Second, law enforcement differs a great deal aroand the confirm our earlier results that the legal institutions of world. French civil-law countries have the worst quality of civil-law countries reduce the size and breadth of the stock law enforcement, whereas German and ScandinaviaLn civil- markets. law countries have the best quality of law enforcement. As for the results on the size of debt markets, the last Law enforcement is strong in common-law countries as column on Table 4.6 shows that French civil-law countries well. These rankings also hold for one critical input into have a ratio of debt-to-GNP 15 percentage points lower law enforcement in the area of investor protections: than common-law countries. Similarly, Scandinavian ori- accounting standards. gin countries also have much lower (28 percentage points) Third, the evidence surveyed in this chapter suggests debt-to-GNP ratios than common-law countries. In con- that large capital markets require that countries protect trast, German origin countries have higher (11 percentage financiers against expropriation by entrepreneurs and pro- points) debt-to-GNP ratios than common-law countries. vide them with good enforcement mechanisms to exercise Finally, both regressions show that rule of law has, as in the such rights. In the absence of a good legal environment case of equity financing, a large and statistically significant financiers are reluctant to surrender funds in exchange for effect on the level of the debt-to-GNP ratio: The move securities, and hence the scope of capital markets is lim- from world mean (6.85) to a perfect 10 is associated with a ited. Specifically, we see evidence that poor legal institu- 20 percentage point increase in the debt-to-GNP ratio. tions result in high levels of ownership concentration, low The results using legal origin are confirmed in regres- availability of external equity financing, narrow equity sions that include creditor rights. In particular, Table 4.7 markets, and small debt markets. shows that the creditor-rights index is statistically signifi- Latin America offers investors a rather unattractive legal cant at 10 percent. The point estimate implies that a move environment. Both creditor rights as well as the quality of from the world mean (2.30) to a perfect score of 4 is asso- enforcement lag behind the rest of the French civil-law ciated with a 8.8 percentage point increase in the debt-to- origin countries while shareholder rights are marginally GNP ratio, which is economically significant when com- better than the rest of the French legal origin average. As a pared with a world mean debt-to-GNP ratio of 59 percent. result, credit markets are exceedingly small, and stock The size of the debt market does vary with rule of law, markets are both small and very narrow. creditor rights, and legal origin. As a result, French and It is clear that improving corporate governance should Scandinavian civil-law countries have narrower debt mar- be at the top of the policy agenda in Latin America. The kets than common-law countries. immediate reaction to the evidence surveyed in this chap- In sum, the results in this section show that the protec- ter is to call for wholesale legal reform. Clearly, rminority tion of investors through legal institutions is an important shareholders would benefit from the existence of a mecha- determinant of ownership concentration and the size and nism to redress expropriation, and there is plenty of room breadth of capital markets across nations. to strengthen voting rights and to enhance disclosure 84 INSTITUTIONAL REFORM IN MARKETS. THE CASE OF THE FINANCIAL SECTOR requirements. Similar arguments can be made for creditor Notes rights. 1. By every single measure, richer countries have a higher quality However, to the extent that improving the efficiency of of law enforcement. Nonetheless, even controlling for per capita the judicial system and asserting the rule of law are slow income, French civil-law countries still score lower on every single measure, and statistically significantly lower for almost all measures, processes, it is important to incorporate those constraints than the common-law countries do. The regression results continue in the policy design. For example, it may be particularly to show that legal families with investor-friendlier laws are also the valuable to adopt an oppressed minority mechanism, per- ones with stronger enforcement of laws. (See regression results in La haps similar to that of Chile, that minimizes the involve- Porta et al. 1998a.) ment of the courts even if its more mechanical nature 2. Ownership concentration per se may be efficient, because the results in less fair outcomes. Similarly, mandating existence of large shareholders monitoring management reduces the agency problem between management and shareholders (Jensen and enhanced disclosure requirements may not be sufficient in Meckling 1976; Shlelfer and Robert Vishny 1986). But large con- countries with weak legal institutions. In such instances it centration comes at a cost as it creates another agency problem: the may be desirable, for example, to require that institutional expropriation of minority shareholders by large ones. An additional investors only be allowed to invest in companies that meet cost of heavily concentrated ownership is that the core investors are minimum corporate-governance standards as determined not diversified. by independent best-practice commissions. 3. La Porta et al. (1998c) find that, for a cross-section of countries around the world, various measures of dividend payout ratios are Finally, in the area of creditor rights one may want to lower in countries with poor investor protection than in countries emphasize bankruptcy procedures that minimize the with high investor protection. This evidence suggests that companies involvement of courts. The United Kingdom's administra- in countries with poor laws and poor enforcement of those laws do not tion procedure is an example of bankruptcy procedure that build reputations by paying high dividends to their shareholders. puts most of the discretion on the hands of commercial 4. The results presented below hold for that uncorrected ratio as banks rather than on courts. Along the same lines, another well. departure fo cuetpaie5. In alternative specifications presented in La Porta et al. (1998a) we also controlled for the Gini coefficient of each country on the the- forces in the bankruptcy process by auctioning off bank- ory that more egalitarian nations have lower ownership concentra- rupt firms much the same way that state-owned enterprises tion; and for the logarithm of GDP per capita under the theory that are currently privatized (Hart et al. 1997). richer countries may have different ownership patterns. The coeffi- cients on these two variables are insignificant in most specifications, and their inclusion does not significantly affect the results presented here. In the case of GDP per capita, a further reason to eliminate this variable from the specifications is that its correlation with rule of law is quite high (0.87). 85 PART 3 Institutional Reform in Hierarchies CHAPTER 5 Reforming the School in Latin America and the Caribbean: An Institutional Analysis T EHE EDUCATIONAL SYSTEM IN LAC CONTINUES TO BE UNSATISFACTORY, DESPITE RECENT efforts to improve the quality and equity of education, including new national education policies, sustained financial support to basic education, and the implementation of educa- tional reforms. Achieving the goals of those initiatives depends largely on the institutional structure and capacity of the public sector, which is the main provider of basic education, and on the interaction of the people involved in providing education at the school, local, and national levels. Persistent inequality and low quality characterize basic education systems in the LAC region. Education inequalities-in access to school, readiness, attendance, quality of teaching, and learning outcomes-per- petuate income and social inequalities and contribute to making Latin America and the Caribbean one of the most inequitable regions of the world. Figure 5.1 illustrates the patterns of differences in preschool enrollment for poor and non-poor children in rural and urban areas in Costa Rica and Chile. Similar pat- terns exist in secondary education participation; in Brazil, for example, 40 percent of non-poor urban chil- dren are enrolled in secondary school, but only 10 percent of rural poor children are. Poor children attend poor schools, have fewer opportuni- of student learning confirms this observation. In science ties to complete their basic education, and perform below and mathematics, Colombian students consistently per- their counterparts in private schools. Figure 5.2, which com- form below their counterparts in other regions of the world pares rural public schools with elite private schools, shows (see Figure 5.3), and regional testing carried out by typical differences in educational achievement between poor UNESCO (1995) demonstrated that Colombia's perfor- and non-poor children in LAC countries. Moreover, the level mance is typical of the region. of attainment of the poor in several Latin American countries Many of the quality-related problems also apply to pri- compares unfavorably with that of other regions of the vate schools, although there is evidence that elite private- world, such as East Asia or East Africa. For example, 46 per- school students in several Latin American countries score cent of Brazil's poor complete Grade 5, compared with 89 higher on achievement tests than their peers in public percent of the poor in Zimbabwe (Filmer et al. 1998). schools (UNESCO 1995)-see Figure 5.2. Privately man- Most education experts of the region agree that the out- aged schools in Chile perform better on national achieve- comes of school systems are disappointing. School attain- ment tests than municipal schools in Chile (Larrafiaga ment and completion rates continue to be low, and levels 1997). Privately managed schools (for example, Fey Alegria of student learning do not meet the expectations of educa- schools in 12 Latin American countries) also show higher tional reformers (Rama 1992; UNESCO 1994). Empirical student retention and attendance rates than similar public evidence from national tests and international comparisons schools (Swope and Latorre 1998). 89 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER FIGURE 5.1 Preschool Gross Enrollment Rates in Costa Rica and Chile, Poor vs. Non-poor 50 - 45 - 4 40 - poor non-poor poor non-poor Costa Rica Chile 1992-93 1992 Source: World Bank estimates based on household survey data. The limitations and challenges of LAG education sys- Many of the proposed changes, such as creation of mecha- reins were discussed and debated at the Summit of the nisms for evaluation and participation by civil society, Americas (OAS 1998 and PNUD 1998). There was general require improvement of institutional frameworks thlat will agreement about the symptoms of the problems of educa- lead to alteration of traditional social behaviors. tion in the region, and the region's political leaders agreed The patterns of behavior of the actors involved in edu- on the directions the required changes needed to take in cation and the interaction among them-which are shaped the Summit Action Plan (Summit of the Americas 1998). hy the rules of the game (both formal and informal)-have FIGURE 5.2 Achievement Test Results, 1992 TEST SCORES 10 20 30 40 50 60 70 80 90 100 Argentina Math 6 3 Science p O Colombia Math p O Science p p Costa Rica Marh p Science p p Dom. Rep. Math p p Science p Venezuela Math i p Science p Rural public schools Elite private schools * p Sorce l UNESCO/OREALC in Schiefelbein (19951. 90 INSTITUTIONAL REFORM IN HIERARCHIES FIGURE 5.3 behaviors of the main actors involved, despite the institu- Average Mathematics Achievement Test Scores, Grade 8 tional changes on the national level. The question for the next generation of institutional reforms is how to facilitate 700 - change in values and behavior where it matters most-at the classroom and school level. Singapore 600 - Korea, Educational Institutions and the School Rep. of 1 E W W i l i i S i _The Latin American school has been a remarkably stable 500 - Ireland 527 system. Not only have the key actors remained unchanged for centuries, but the shared values, expected behaviors, 400 - Portugal - and rules of the game-the institutional framework-also Colombia have been remarkably stable. Only recently has a first gen- eration of education reforms attempted change. A major 300 South Africa challenge to implementation of these school reforms, in fact, has been the strength of the system and its resistance 200 to change. The traditional school and its principal actors are shown in Figure 5.4. The school includes a director, teachers, and 100 > students and is influenced by a number of external actors, including parents, supervisors, ministries of education, 0 teacher associations, and politicians. The interactions between these actors, both within the school itself and Soarce: Beaton et al. (1996). between the school and the extcrnal actors, are very com- plex. What is important, however, is that in many cases the combined interactions and behaviors of these diverse actors not changed significantly with the initiation of educational are dysfunctional, so adequate learning does not take place, reforms. Misallocation of resources, inefficiencies, and lack despite additional public spending and despite new educa- of accountability are still predominant attributes of the tion laws. In the pages that follow, we describe some of the organizational structure of education in LAC. Parents are major interactions, starting with those inside the school still not demanding better performance from public itself. Although the discussion focuses on public schools, schools, schools lack capacity and resources to respond to new institutional environments, and clientelistic practices FIGURE 5.4 continue to dominate the decision-making in several coun- Principal Actors in the Traditional School tries (PNUD 1998). A decade of educational reforms in LAC countries has achieved legal changes, placed education as an important theme in social policy, and demonstrated the possibility of radical change on a relatively small scale. Indeed, most LAC countries have developed new education policies, and some of them have adopted general education laws (e.g., / External Actors: Argentina, Brazil, Colombia, and Mexico). Nevertheless, the effect of the legal changes and their impact on the pro- Director Parents vision of education at the school level, and on the improve- Supervisors ment of education nationally, remains to be seen. Few Teachers I Teacher Representatives ,, | | ~~~~~~~~~~~~~~~~~~~~~~~~~Teacher Trainers reforms are affecting schools and teachers at the local level, Students T |ae and there are difficulties in changing value systems and l 91 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER which enroll 80 percent or more of all primary-school stu- that may not succeed, to seek assistance from school super- dents in LAC, the implications are also relevant for private visors, the director or peers, to spend time discussing stu- schools, which are on the increase in several LAC countries. dent performance with parents, to collaborate with peers in Private secondary-school enrollment in Bolivia, for exam- assessing students, to collaborate with peers and the direc- ple, increased from 17 percent in 1980 to 26 percent in tor in managing the school, and to allocate time to non- 1990, and from 24 percent to 42 percent in Chile during school activities, including those of the teacher associa- the same period (UNESCO 1994). tions. Each of these decisions in and of itself is complex, but each is also heavily influenced by the behavioral norms Actors in the School and expectations of other actors. Excessive teacher absen- Students are the fundamental principals of the education teeism, for example, may be condoned by peers as "beating system, but they have little power to influence the quantity the system" rather than "violating the rights of the child." and quality of services they receive. Nevertheless, they have Teacher associations may protect the employment rights of considerable power to affect their own learning through the teacher as a matter of principle, even when students personal decisions about how to use their time, including, suffer. Teachers may view impoverished parents as too for older students, whether to participate in the labor force. poorly educated to understand their children's learning And student decisions about their own classroom behav- problems. ior-which are, in turn, strongly influenced by the values Directors of the traditional public school are selected and behaviors of peers, parents, and teachers-affect learn- from the teaching faculty and sometimes return to the fac- ing and schooling outcomes. While students clearly have ulty after serving this administrative duty. Usually, direc- some choice, research shows that negative expectations of tors retain membership in the teacher association. The teachers and parents about student achievement, which responsibilities and authority of directors are severely con- occur frequently, are often self-fulfilling. Thus, the very strained, especially regarding the recruitment and evalua- high probability of failure and repetition in many Latin tion of teachers. Their principal duty is the administrative American schools may in and of itself adversely affect enforcement of basic employment rules. They usually are learning. not expected to initiate change, evaluate staff, or provide Teachers are the fundamental agents of the school, and support to teachers with performance problems, and they they are agents for multiple principals in addition to stu- have very few ways to reward good performance or address dents, including the school director, parents, and the special needs of individual teachers. Besides, directors typ- teacher associations. Research on learning shows the pre- ically receive no specialized training in school management dominant influence of the teacher on learning in the class- that would equip them to carry out these responsibilities, room. Teachers have the problem of divided loyalties that so they seldom know the arcane and complex requirements all agents face: They have self-interests that may not coin- for removing or disciplining teachers, even in cases of cide with the interests of their employers or bosses, yet extreme cause. Like teachers, directors have divided alle- they have strong professional interests that are consistent giances. They are agents of the education ministry, but with those of the principals. their day-to-day work cannot be carried out without the Most teachers first enter the classroom after graduating cooperation of the teaching faculty, whom they may one from teacher-training programs in universities or in spe- day rejoin. cialized teacher-training institutions, or normal schools. These training programs inculcate values and behaviors, in Local External Actors addition to providing knowledge on subject content and External to the school, but closely tied to it, are four pedagogy. Teachers make numerous decisions that affect extremely important actors-parents, locally electecd repre- learning in the classroom, including whether to be present sentatives of the teacher association, the appointed local or absent, to take training to improve teaching skills, to representatives of the education ministry, and the trainers use their own time to develop lesson plans or to grade of teachers. examinations, to carefully assess individual students and Parents are by far the most important determinant of individualize teaching accordingly, to try new methods learning by children. Their own education (that of moth- 92 INSTITUTIONAL REFORM IN HIERARCHIES ers, especially), the value they attach to doing well in tion, especially, but not exclusively, around salary negotia- school, and the interest they express in their children's per- tions at the regional or national level. From the perspective formance-all affect the child's motivation and effort. On of the individual teacher, the representative is available and the other hand, parents have very little "voice" or "choice." responsive. If they are seriously unhappy with their child's education, Local ministry representatives are typically former teachers. they have little recourse-especially if they are poor-save Often they have received little specialized training to pre- lengthy appeals through the education bureaucracy. Since pare them for this role. These frontline representatives or the costs of such action are high and the probability of a supervisors are responsible, in principle, for ensuring qual- positive response low, parents seldom undertake it. Highly ity by monitoring and assisting teachers and directors, and motivated parents may be active in their parents' associa- they communicate policy and regulatory changes to school tion at the school; however, they quickly discover that staff; however, their primary role in fact is that of enforce- most school authorities want them for their potential to ment of government regulations and policies. In most contribute financially and not for their interest in the edu- countries of Latin America, their supervisory role is not cation of their child. Indeed, these parents may receive a taken seriously. There are few supervisors, and they gener- hostile response from school staff if they are perceived to be ally make infrequent school visits; supervisors in rural questioning the work of trained professionals. In short, areas, where the need is greatest, often lack transportation parents have learned to be passive actors in the system. to school sites. In short, the local representatives of the In the traditional school, parents depend heavily on the ministry are not highly visible and are not viewed as state to monitor and regulate education. Given the limited sources of help. knowledge most parents have to evaluate pedagogic prac- Teacher trainers offer both pre-service and in-service tices or the quality of instruction, they depend on the state training and are most commonly situated on university to do this for them. Indeed, ensuring quality and informa- faculties and specialized teacher-training colleges. The tion about quality is an essential role of the education min- trainers of teachers have heavy responsibilities. They must istry. But since parents rarely fight their battles up the edu- take the theory of pedagogy found in the textbooks and cation bureaucracy, the staff of education ministries seldom convert it into instruction on good teaching practices. find out about their problems, making monitoring partic- This requires that the trainers themselves have extensive ularly difficult. experience in the classroom and, perhaps more impor- We should point out, however, that parents are not tantly, that they have continued intimate contact with powerless. Those with adequate resources may elect to send actual classroom practice. The trainers have the responsi- their children to private schools or to move to neighbor- bility for knowing classroom reality, identifying good hoods where schools may have better reputations. Parents teaching practices in light of that reality, and communi- may also purchase additional instruction after school hours, cating those practices to their student teachers. They are, and they may work with and encourage their children to in effect, responsible for carrying out "action research" on engage in learning activities in the home. The home envi- teaching to permit the education system to "learn" from ronment is a powerful factor in determining students' aca- its experiences. In several respects, the typical teacher demic success (Kellaghan et al. 1993). trainer in Latin America falls short of this model. Those Local teacher representatives, elected by majority vote, who have studied the issue cite several problems, includ- generally reflect the perspective of the average teacher, ing failure to maintain close contact with practice in the although, as with all democratic systems, large numbers of schools; failure to effectively translate theory into practice; teachers disagree with the positions taken by their associa- excessive emphasis on theory over practice in instruction; tion (Scobie 1998). Local leaders serve three main func- failure to carry out practical research and feed lessons tions: They protect teachers from violations of their con- learned back into teacher training; and failure to maintain tractual rights; they communicate to teachers the contact with newly graduated teachers (Villegas-Reimers association's perspective on educational policies and pro- 1998). In short, the deficiencies in teacher training pose a grams advocated by the education ministry; and they mar- major obstacle to raising the quality of learning in the shal vocal teacher support for the positions of the associa- classroom. 93 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER National External Actors budget decisions. However, the ministry has no equivalent Also external to the school-but physically remote from ally when it comes to making the argument for non-salary it-are other important actors who directly affect the expenditures. Thus, there is a persistent distortion in school. These include the national ministries of education resource allocation in favor of salaries. and finance, the national teacher association, and political The weak capacity of education ministries to carry out actors, including political parties, the church, non-govern- rigorous analyses, to prepare well-justified budgets, and to mental think tanks, and, increasingly, private business manage itself reflects in part the dominance of the political leaders concerned with the quality of skilled labor. over the educational goals of the ministry itself. The weak The national ministry of education is itself a highly com- capacity of the education ministry, combined vvith its plex organization with numerous principals and agents, political role, constrains its ability to carry out two key some of whom are physically separated by great distances. responsibilities of the state-ensuring quality ancd ensur- The minister is the agent of the president and brings to the ing equality of educational opportunity. ministry a number of high-level appointed officials who are The ministry of finance is headed by someone who is as responsible for policy development and overall manage- much a technical appointee as a political appointee. He is ment of the sector. For typical ministers, the education selected by the president of the country for his ccmbina- ministry is a short-term step in their political career. Their tion of political skills and credibility as the guardian of the major challenge is to balance the demands from the finance treasury. The importance of maintaining a credible macro- ministry to keep a lid on costs with demands from the economic policy also results in the finance minister's hav- teacher associations to increase wages and demands from ing longer tenure in office than the education and other politicians to provide political payoffs in the form of new social-sector ministers. The seriousness of purpose of the schools or employment for political supporters. The single finance ministry usually translates into a determined largest risk he faces is a prolonged teachers' strike, which attempt to control the fiscal deficit by constraining probably would force him to resign. In short, the principal increases in public spending, including that of the educa- wants an agent who generates political benefits, not costs. tion ministry. Since salaries represent a very high percent- The minister is the principal to a group of appointed age of total education spending, there is a natural conflict executive officials, who by and large have goals congruent of objectives between the education minister, who needs to with those of the minister. The difficulty arises in the rela- increase salaries to appease labor, and the finance m,inister, tionship between the appointed officials and the ministry's who needs to limit salaries to maintain an acceptable fiscal civil-service bureaucracy, which has seen many ministers, deficit. which has dealt with many new, short-lived policy initia- The leadership of the teacher association, is democratically tives, and which has the power to defeat policy and man- elected and usually has much longer tenure in office than agement initiatives that threaten its own power. the education ministry leadership. While the leadership of Venezuela's failure to decentralize education, for example, the association or union can guide its membership, the has been attributed to the resistance of the rank and file democratic nature of the organization means its positions staff of the central ministry (Hanson 1995). never can deviate too far from those of the average teacher. Public administrators often view the education ministry The association has as its first and foremost objective pro- as being among the weakest of all government ministries tecting and enhancing teachers' employment rights. This in its capacity to manage its sector and to implement new translates into ensuring stable employment, adequate programs and policy initiatives. It also tends to have poor salaries, and satisfactory working conditions. The pursuit capacity, relative to other ministries, to plan and imple- of employment stability means the association, and at least ment its budget. This particular weakness puts the educa- a majority of its member teachers, are steadfastly opposed tion ministry at a technical disadvantage relative to the to any political or management change that might threaten finance ministry when it comes to justifying its budget the employment or even location of employment of the request. However, the education ministry has a political teacher. Naturally, this translates into vehement opposition advantage in the ever-present threat of a teachers' strike. In to market-oriented proposals to increase parental choice in this sense, the teacher association is its political ally on the selection of schools. 94 INSTITUTIONAL REFORM IN HIERARCHIES The real power of the teacher association is its capacity lic school curriculum; the other is the desire for continued to convince teachers to take to the streets.2 Successful state support for church schools. Since church schools com- teacher associations recognize that constant and effective pete with public schools for students, there would appear communications with the rank and file are critical to main- to be a natural conflict with teacher associations, many of taining this capacity. Association leaders who fail to listen whose members are also supporters of the church. This to their membership will find that eventually their call to conflict is avoided in most countries by giving public sup- strike will fail, and possibly will result in a demand for a port to church schools in the form of school teachers who change in union leadership. Highly successful associations, are recruited and paid by the government, thereby mini- like FECODE in Colombia, publish sophisticated education mizing the employment risk to public-school teachers. news journals for their membership, sponsor educational conferences, and in other ways maintain constant and effec- Problems with Educational Institutions tive communications with teachers. The sophistication of The traditional school as characterized here is an inefficient this communication is in stark contrast with the communi- organization. The description given above of the various cation programs and policies of most education ministries. internal and external actors in the school illustrates a num- Teacher associations in Latin America have another char- ber of the concepts of institutional economics. For one acteristic that significantly influences their own behavior thing, the information asymnmetries between principals and and complicates labor relations: They are highly politicized, agents are large. Both ministry staff and parents have little with leaders who frequently aspire to political office. information about the performance of particular schools or Indeed, many associations are so closely aligned with par- teachers. Other more general information problems are also ticular political parties that their leadership becomes a step- severe. Teachers have little information about the perfor- ping-stone to an important party candidacy. These political mance of individual students or even their own classroom aspirations sometimes result in the leadership's expressing relative to others; hence, their capacity to diagnose instruc- positions on numerous issues unrelated to education. tional problems is constrained. Educational policy analysts Political parties have two main effects on education. and budget analysts often do not have the information First, some political parties depend on the political support required to do good analyses, and ministry policy-makers of teachers and teacher associations. These parties, of lack the information to make good policy decisions. course, are expected to deliver benefits to teachers if and The general problem of lack of information, however, is when they take office-although they may not always be less important than the shortage of incentives to seek and use able to do so if the political benefits of increasing teacher information. Teacher trainers are often criticized for not pay are outweighed by the economic risks of an increased being sufficiently "connected" to the real world of school fiscal deficit. On rhe other hand, at least those parties are practice and school problems, but it is within their own likely to consult with teacher associations when making capacity to obtain this information. Both education and major educational policy changes; parties not having the finance ministry officials are keenly aware that budgets are political support of teacher associations have no political distorted in favor of salaries, with serious, negative impli- obligations to pay and, therefore, no need to consult. cations for the performance of schools and teachers, yet the The second effect of political parties is their demand on distortion persists. When important educational policy the education ministers to deliver political benefits to loyal decisions are made primarily on political grounds, good party members in the form of employment and contracts. information and good analysis of education issues may be In its extreme form, this political clientelism results in the irrelevant. familiar phenomenon called "ghost teachers," i.e., teachers Principal-agent problems abound inside and outside the who receive salaries but do not teach. These phantom school. If students are the fundamental principals, poorly teachers, who have traditionally been a significant share of informed and poorly educated parents may fail to ade- the total teaching force in areas like Northeast Brazil, rep- quately represent their interests. Even when parents think resent a significant loss of resources for the schools. they know what constitutes a good education for their chil- The church is a unique actor in education with two major dren, they have almost no voice to express their prefer- concerns: One has to do with the values taught in the pub- ences. The transaction costs of expressing preferences to the 95 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER ministry are extremely high. In addition, there is the moral problem with the traditional educational organization is hazard problem that parents depend on the ministry to its slow evolution in the face of glaring and obvious prob- regulate the schools their children attend and, hence, have lems. There are excellent schools and excellent teachers little incentive to use their own time and energy to moni- and excellent teacher trainers, supervisors, directors, etc., tor school performance. within any school system. There are also very bacl exam- Other principal-agent problems are also evident. School ples of each. The richness of the range of educational expe- directors have little power to manage resources, resulting riences provides a fertile ground for learning and improv- in a high degree of autonomy among their agents. How- ing-in short, for evolving. Creating a learning ever, the problem is still more serious. Teachers and direc- organization will require changes in values and behaviors tors, who have the best information about the school, have that go beyond solving information problems. It will little authority to act on that knowledge. They cannot require instilling a sense of personal responsibility and change the curriculum, select different textbooks, or in accountability among all the actors for the performance of other ways change production relationships within the the school. school. While teachers do have control within their own classrooms, this control is constrained by what happens to The Theory and Practice of Education Reforms students outside their classroom and by the low level of Many of the problems analyzed above have been addressed technical assistance and support provided to them by the in several education reforms undertaken in Latin America representatives of the ministry. and the Caribbean over the last decade. These reforms have These problems translate into a lack of incentives for concentrated on: school directors to set organizational goals and monitor the * Increasing the emphasis on learning outcomes, over school educational outcomes. School principals do not simply increasing inputs posses the resources to motivate teachers and give them * Providing opportunities for broader social and com- incentives to improve their teaching skills and innovate munity participation their classroom practices. Teachers are not rewarded for * Giving more management autonomy to individual their teaching quality or their level of effort for helping schools students to learn. There are no standards for the teaching * Promoting the development of new curricula, and profession, nor evaluation mechanisms. Parents do not localizing some of the content have incentives to participate in school management or the * Upgrading teacher competencies, and information to monitor school performance. In the major- * Raising the accountability and social responsibility of ity of cases their choices are nonexistent. the system (Alvarez 1998). The final and perhaps most important underlying The mechanisms employed by most reforms to carry out FIGURE 5.5 LAC Education Refonns YEAR MANAGEMENT REFORM STUDENT INFORMATION PARENT DECENTRALIZED INITIATED EVALUATION SYSTEMS PARTICIPATION MANAGEMENT Argentina 1994 X X X Brazil Minas Gerais 1991 X X X X Parana 1995 X X X X Chile 1991 X X Colombia 1989 X X Dominican Republic 1990 X X X X El Salvador (EDUCO) 1992 X X X Mexico 1991 X X X X Nicaragua 1992 X X X X Paraguay 1995 X X X Source: Based on data in AED (1996). 96 INSTITUTIONAL REFORM IN HIERARCHIES these goals are summarized in Figure 5.5. Most reforms and the MERCOSUR countries in the World Education have seriously addressed the information problem by pilot- Indicators Project means there is now high-level demand ing national student-evaluation systems. The magnitude for such information. and importance of this change is immense. A decade ago, Perhaps the most common element of the last decade's only Chile had a national student-testing system, and even reform is decentralization. Several countries (Argentina, it failed to make local results available to the public. Today, Brazil, Colombia, Mexico) increased the education respon- most countries in the region have some kind of testing sys- sibilities of regional governments; in Argentina and Brazil, tem, as shown by Figure 5.6. Although few of these sys- central government ministries now have very little respon- tems have matured sufficiently to provide feedback to par- sibility for financing and managing primary and secondary ents and teachers about the performance of individual education. In other countries, most notably Chile, the fis- schools and individual students, most countries now have cal and administrative management of education has been the capacity to sample schools and students and to monitor largely turned over to the municipalities. Decentralization national progress in learning (Rojas forthcoming). In short, of education to subnational governments has moved the the capacity of education ministries to monitor and provide locus of much decision-making closer to parents, but it has information on quality has clearly improved in the past not, in general, affected the school. decade, and this improvement is widespread throughout More important for the school than government decen- the region. tralization has been the movement toward increased school Similarly, many countries in the region have made sig- autonomy and community participation in a few countries. nificant investments in modern management information In Chile this has taken the form of empowering teachers systems that allow decision-makers and budget analysts and school directors to develop school-specific develop- and planners to carry out their work with more timely and ment projects. In El Salvador, the EDUCO program has more comprehensive and accurate information on student given parent councils the authority to employ teachers and flows, educational costs, and expenditures. Advances in allocate resources. In Minas Gerais, Brazil, school councils computer hardware and software have permitted more comprising parents and teachers have the authority to rapid improvements in these systems within shorter peri- select the school director, who is required to present a ods of time than was true with past investments in infor- school development strategy as part of the selection process mation. While the sustainability of these systems is still in (see Box 5.1). doubt, the participation of countries like Mexico in OECD Finally, decentralization and movements toward school FIGURE 5.6 Assessment of Educational Quality: Latin American Testing Systems, 1986-97 25- 2 ; 0 -. . . . . . . ............................................../K..... Cr *1 0 5- . ... .. . . ... . . . . . . .. ....................... . . . . ..... .. C, Z O 1986 1988 1990 1991 1992 1993 1994 1995 1996 1997 Somrce: Rojas (forthcommg). 97 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER autonomy have been accompanied by changes in financ- Second Phase of Education Reforms ing arrangements. Several countries have introduced The first phase of education reforms over the past decade competitive financing mechanisms under which schools set the stage for new large-scale social interventions that can compete to receive financing for special projects. will more fundamentally attack the institutional and other Minas Gerais, Brazil, introduced capitation grants to problems of the school. The first generation of reforms schools, giving school councils a degree of budgetary emphasized decentralization, legal changes, development authority and ensuring at least a minimum level of of pilot innovations, and creation oftesting and evatuation spending on non-personnel inputs. In the early 1980s, systems. The second phase of reforms will need to broaden Chile introduced capitation grants to municipalities and and deepen these reforms by focusing on changing rules privately managed schools, providing strong incentives and institutions that explicitly address the information and for schools to compete for enrollment and retention of accountability problems that shape the behavior of actors students. In 1997, Brazil reformed national revenue-shar- at the school level. Table 5.1 illustrates this evolution. ing rules to ensure that all jurisdictions receive education First, the new phase of reforms needs to move from a revenues sufficient to guarantee a minimum level of focus on the national system to a focus on institutions at spending per student ($350). the school level. The reforms of the past decade have fre- 98 INSTITUTIONAL REFORM IN HIERARCHIES TABLE 5.1 in small slices in diverse areas. There need not be one Focus and Goals of Two Generatons of Educational Reform in LAC grand, top-down education reform (see Box 5.2). FIRST SECOND Second, with the legal framework often in place, the Focus on system Focus on school next set of reforms needs to focus on the much more diffi- Decentralized decision-making Teacher and parent empowerment cult problem of making sure the legal framework leads to Pilot innovation projects New roles in policy-making a change in the values and behaviors of the actors. This Large scale implementation of piloted projects transformation will require navigating uncharted waters, and, to be successful, will require a systemic approach that Changes in legal framework Changes in value systems and rifre hne twl edfiutt uti hnea Top-down reform efforts behavior reinforces change. It will be difficult to sustain change at New educational policies Popular reforms the school level unless ministry officials, teacher-associa- Priority given to education in Behavioral changes in all actors tion officials, and parents also change their behaviors. social policy While the first generation of reforms has centered on Development of management Development of accounrability the top, emphasizing legal changes, the second generation information systems mechanisms informatin systemsmechanism of reforms must start at the bottom, emphasizing behav- Creation of basic data bases Broadening the use of information Establishment oftesting services by all actors ioral changes. For example, for local decision-making to Using evaluation information at occur, there needs to be training to allow the school coun- the school level cils (or teachers) to exercise their new powers fully. The Changes in structures Changes in pedagogical practices irony is that if grassroots efforts are to succeed, they need New role of governments Professional development of Emphasis on states teachers to be promoted by the educational leadership, at both the Use of advanced technologies ministry level and in the teacher association. In short, top- Changes in classroom activiries down reforms need to be converted to popular reforms sup- Emphasis on learning processes ported by a broad, national consensus. This will require, Basic re-engineering Development of learning among other things, that ministries develop the capacity to Blueprint approach institutions Social learning approach effectively publicize and sell reform efforts. As the case of Application and utilization of Colombia shows (see Box 5.3), grand education reform research findings designs are not necessarily implemented if key actors, espe- cially the teachers, do not buy into the reform. The imple- mentation of Paraguay's 1990 second-phase educational quently decentralized decision-making, improved infor- reform followed an initial top-down reform effort. The top- mation, and legally empowered parents with relatively lit- down approach used in the first phase "delivered" the tle impact on the school. For example, the municipaliza- reforms to teachers via traditional in-service teacher train- tion of primary and secondary schooling in Chile had ing aimed at changing classroom behavior. The initial fail- relatively little impact on the municipal school (Espinola ure of this approach led to a radical change. Teachers were 1997). The extensive constitutional and legal reforms in invited to participate in discussions and "learning circles." Colombia have affected the school only very slowly (Mon- They took on ownership of the reform-implementation tenegro 1995). In contrast with these legal reforms, the process and committed themselves as individuals and second generation of institutional reforms needs to members of small groups to improve their teaching behav- empower parents and teachers or, where these actors have ior (Hobbs and Rentel 1997). been legally empowered already, focus on the actual imple- However, not only the behavior of teachers needs to be mentation of the law to ensure that parents and teachers changed. Parents need to be given voice so they can con- begin to play out their new roles-that is, change their tra- tribute to the second phase of reforms. One way of empow- ditional behavior. In some cases, this will require giving ering parents is by expanding their choices and promoting school councils real, rather than advisory, powers, as is competition among schools. Case studies on the provision beginning to occur in the states of Minas Gerais and of education and health services conducted in Chile, Costa Parana, Brazil, and in Nicaragua (Winkler 1997). The case Rica, and Venezuela suggest that competition improves of Chile demonstrates that real education reform can occur efficiency, regardless of the public or private nature of the 99 BEYOND THE WASiL Education Refonii m CMos lb. Whdt Exceeds the Sum of Its Puts SiId the bginnin of the 1990s Chile has 4jached purees in schools and create on line learning cnmmiini- edw5arioii !efbiin with a strategy of coiitii an4 lucre- ties connecting secondaty and primary schools to s nitaiciiaAgs. initiated by schools rarhet edhipre- national network arid then to the Internet Other projcs h&nsiv lineai action. p an Four basic fhir iP1ain the focus on a variety ofareas, including teacher training spec stccess of the reftrm proiies cialized radiitig methodology nd curriculuni for ru+a educaeioit ysrem for oloy ihete ar& iekral cominoiialries amoh the ?MEs; over a decade in spite of changes in national and based on new critena for education pohcy both at the sys. ministerial administrations tern level and at the level of the actors * Macroeconomic stabihty and growth making avail * A consensus must be built in the proces able the financial resources to sustain reform and national education policies * Widespread public acceptance of proposed polucies * Decentralization increases responsr'ienes which focused on the need to modernize the ethics demand tion seccot rather than on addresB33=g or B?} ih S>#Br13 ptibli its<83 ?inin82Bg8gisBg3B. B?Bg., BB.B BBaBBB;;B8B3B 3: andrcmedros caton' BeP~tc g idehoe on itehing school4 083:??3:go 'Te aeoec fte R seoe yvsn eac,admngmn I shwvrwti h the'S>.B B Be?spon???sibl foatas<'' in th? perfo3+a8BBBB3BBrmace'5B of?3BB scho? power of theB,Ba3* ?R toB meaur stda leanin tkrou i 5?, 33?,a;-,?h in ai M itefoEdctoReew whorprsdrel tnarie et atog t sntcretydigs tov3 Bx pariaen rathe than the Biuse B4B BBii~ts To I New Ze- ad the prncpa ain reaiosi hasBB BBB'BBB .' caryou hs rsPnsbliy~th mit6 f dii~nxn ben edfie ad ad ffctvei te duaio10eco INSTITUTIONAL REFORM IN HIERARCHIES lems that are solved today will reappear tomorrow. If the actors could bring support and direction to the reforms. benefits of educational reform are to be sustained, teachers The private sector, for example, now understands the need must learn to carry out their own action research to iden- for better education of the labor force for its own success in tify problems and seek solutions; school supervisors must the global context. The first stage of the educational reform play an important role in facilitating this kind of research in Dominican Republic, initiated by the private sector, by teachers; and teacher trainers must also actively support illustrates the possibilities of cooperation between the pri- teachers and communicate to future teachers the lessons vate and the public sectors. The ministers of finance, who being learned. There are no hard-and-fast rules about how traditionally had considered the social sector only as a to create learning organizations, but the extensive research source of expenditure, are now more committed to the on successful organizations can provide clues. National development of human resources. The declaration of the ministries and regional secretariats of education will cer- Summit of the Americas is an expression of this new cli- tainly play a key role in the creation and maintenance of mate of political interest in education by all government learning, and this will require fundamental changes in the agencies. On the other hand, teacher associations continue organization and functions of ministries, with a new focus to resist the educational reforms that introduce account- on support rather than control. ability, decentralization, and performance-based pay. Parental Voice and Choice. Parents may fear that local par- The Political Economy of Reform ticipatory school management will cause a decline in gov- Latin America has by now acquired considerable experi- ernment financial commitment to schools. However, pro- ence in educational reform. These experiences, which have viding voice and choice to parents and communities helps been reviewed in detail elsewhere (see Alvarez and Ruiz- balance the interests in favor of children, the ultimate prin- Casares 1997), suggest that the political economy of cipals in the educational organization, and introduces some reform needs to be divided into two distinct stages- degree of competition among schools. Various countries are design and implementation. trying out creative financial alternatives to respond to the In the first stage, a reform is designed, legislated, and new reform philosophy. Several countries, including Chile regulated. When reforms can be designed within the exist- and Colombia, have experimented with voucher systems ing legal framework or when governments enjoy over- with varying results. Trinidad and Tobago schools have whelming majorities in the legislature, the first stage of employed school donation schemes with in-kind local con- reforms is not complicated. Reforms can be designed by tributions, and Ecuador has tried competitive mechanisms technocrats inside or outside the education ministry and for public-school improvement through the Red de Apoyo subsequently adopted and announced by the government. Tecnico. When governments do not enjoy majority representation Incentives for Change. The negotiation process underlying in the legislature, or when proposed reforms would reform efforts has varied from country to country with threaten the interests even of party members, sophisticated diverse results, given different political circumstances and political strategies are needed. In cases such as the Domini- reform packages. In some cases, compensation schemes for can Republic (see Box 5.5), El Salvador (see Box 5.6), and teachers, such as increasing their salaries or maintaining Chile, a process of consensus building and participation the purchasing power of their salaries (in Chile, for exam- accompanied the development of the reform framework. ple) have been used to obtain their support. Other coun- This process, however, does not guarantee the successful tries have provided bonuses for participating in reform implementation of the reforms. activities, as in the Dominican Republic. Non-monetary In the second stage, reforms are implemented. Regard- incentives, such as participation in the reform planning, less of the political power of the government, implementa- and professional development have worked in Paraguay tion is a complicated business. Latin American experience, (Hobbs and Rentel 1997). Appealing directly to teachers however, suggests a few factors to consider in the imple- and generating community support were successful strate- mentation of an educational reform. gies to the Escuelas Integrales (Integrated Schools) reform The Interests of Principal Actors. This involves opportuni- program in Venezuela (Lowden 1996). However, linking ties as well as challenges for reform implementation. New incentives for educators to school performance, and align- 103 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER theIoreticaland practical difficul ties. Inetives forteauchr satgopfhilyrfeinledcosinhm- trainersaren ando c writiea n inthe im olmentaicon ofpulicg intr of edueatefomn Cholenwthm previou slowienc The 'ptedrm eie'rl t~aiga educational cagsShol reform imatement ation poec aag'emennht hand resenrh chngaed thd organozad tep uetini ofly . onEs nei aS se as a tool for abandon S1ed4 cmuictin theSSS <9 vs§ion of a natonS edcto pSa in *NENkU Oraiz i S v = fatSS SSS SnS NISe SrySS chaneSSiSS an efrmnewt broade social cgoal cntv s fo steacheresns topmades inr the stucce,su omplerenation, n cofnreforms. trinersare althe dritialou incltded imlentatciaorwnd of orgniag-n nPed(hmiistry of education,i hl,wt rviu xeinei femedusaional adacmnhandes Scolreogni mlmettion forojc managmpent and rexsti c,hngebha doa ptthei orgniz- depe addiutimaeyon tohevitalitin educatio tieahn prorcesso tion el cutourevandpoiedntoaladitentoa Orgniz~gatinal Capablitconse fsus helped to streniatthen.the ediblinancoathfareorm. Tereowas (Angllc of6) Asatsocal' 'Thueda im sle en seto h ococnmcsnation of Thom eensas nh ti eang lessntrise of tedomncatina epriencem requresevalua- tudeos tr and educaiors ona nereduras. anThe poiiangta inadkoldedevelopment m~te onreswsigt nertakesm prcinpadition . otio d sisandtens nivealctivit iieosevea fclronts suc gals cnd sofial lneantie ang mros ntialvationa otonsenu wh s lea.dt-e thceortical andtractgical difterficultes.Inentvs fore crtachlere prcsmalgoposihypoesinleuaosi.h trinrsar ls citca n heimlmetaio o on- styofedcaio n hie,wthprviusepeiecei INSTITUTIONAL REFORM IN HIERARCHIES . - BOX 5.6 . ing -a- -od-O.: After years of CI War in i a i Sa i es, teac ers' organizations ad Mon- vador, the Chapultepec A&ords ed in 199,2,and vetted weekly.Ai.cu.. stohdisuss bseratn c us ...ions and agreement was reached on the prima-y importance of recommendati f n. This was a valuable,process ationa educatio reform. With the achievement ofof cons s building that enabled:a:coherence between coeconomic slityin the ountry, ec ic growth the various compon s of e reform and ha continue resum.ed.- Steay comitment:on the pair of the govn- throughout implementation, ae t e meut and especially t inistry o e onM E) ispro . Ine provirde the appropriate political enfor Jmple- icsiCommittee, responible frdisseminain the nt&nration of the 1992 N ational Plan of Action for- Edu- results of the res and Adi- Group dissionsto cation to begin, allowitg the establishment of legal, po ical fleader in all pties, international fing administrative,andfinancia manageent mecani teachers and students, and MINEDsperso enable the effective transfibmation ofthe:education S.ctor. t heAdvir Grup qe Theprxncipa compnenit of th refrm, Educatio ith nam r as r. at Community Participation (EDUCO>, began as a pil erelevantactors in the e dcation refo jeer i: 1991 and fouses on presch 1d ar-lvel proce,theforulai f edu n in rural are ougCmmunity Eucation ae acion p an Associations (ACEs) with democraticaelectd dectorS, El Salvado's 1995 naiona edu io reforr effortwas EDUCO involves the partiipation iofparents ad comdrnu-i intated to tagt ble of limited access, systemic nities in the design and administrariowof formal educatio ineiciencs,an low ltyhre fiact £C in rual areas. The ACE assume responslit or hring significa t contry's o-going success in addr- t S: aid inmtoring their performance manageeit: ing those problems: of sol bud ad maintenance of school libarmies and * Soe mc stablit and steaat- gernmnent clsrooms. This transfer of resources shifts some responsi- commitmen to educational reform hiliry for the quality of educational services to locat corn- e Increasedm local paad .munities creatig a aSenseg oflocal ownership in ration an management, and doin proess. EUCO proved effective ad ibecame ion of a art bw inspiratiOn .an modelf&or the 1995 Educational Reform, a e£ment and the Po i t s pporte a £ion.agTo al-tfton th ho Sre and. . - 1-erplan for sector-al transformation.. sutained the reform --UndLerscorig the- government's commitment to and£ Te 9 ducational Refrm and it principle strat-t design ofomad min of l paticipatis gcooet etheval of incr -an ongoing d ialogue between MINED an the popula- orating participation of all relevant organizationsan - on which btgan in. 1993 with the esalsent ofthe actois in£ the reform of tional education systems, Advisory Group. The A yG o upA.-frmed from son, applic b to a ontri h res archers who undertook.the baseline study of the ediu- reinvent educ.atinal servi cation system and represtetatives fm -v go - - ental and noTgo entatl organizations, -incudiig S-n; Bejr (1997) a Mea (1997) - Prospects for Reform in the global marketplace of the next century. This aware- There is an increasing awareness on the part of society that ness has given education political visibility and has a highly skilled labor force and its capacity to use knowl- increased the political payoffs to leaders who undertake edge flexibly will be key to a country's ability to compete reforms. However, this payoff probably is considerably 105 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER higher for initiating and legislating a reform than it is for cially, teachers to raise awareness about the benefits of implementing a reform already in place. And it remains reform. As noted earlier, education ministries have tradi- true that reform incurs high political costs up front with tionally been ineffective in communicating the intentions the expectation that the educational benefits of the reform and expected results of reform, especially relative to teacher will take many years to surface. Still, the prospects for ini- associations. One example of how to do this comes from tiating the second phase of reforms are encouraging. Victoria, Australia, where an education minister who The focus on education is, also, not likely to recede in wanted to aid reform efforts created a television studio in the near future. Several countries in the region have elected the basement of the ministry and began broadcasting pro- to participate in international testing of educational gramming of interest to teachers directly to schools at the achievement and in the development of internationally end of every school day. comparable indicators of educational inputs, processes, and The dialogue between government and the teacher asso- outcomes. These international comparisons have the poten- ciations is, of course, a critical one, which is made more dif- tial to stimulate political debates on education and to ficult by the fact that the associations may be aligned with increase the political payoffs to reform. One result may be political parties that oppose fundamental reform efforts an increase in the political prestige of the education minis- and, hence, see little benefit in the government's succeed- ter, with increased likelihood that presidents will appoint ing in its reform efforts. Despite this political opposition, politically skilled individuals and give them the tenure it is critically important for governments to have continual required to carry out reform. and open dialogue with teacher associations and to listen The increased political visibility of education addresses, and be receptive to their views. Strong opposition by in part, the fact that the immediate beneficiaries of educa- teacher associations to proposed education reforms at the tion reform-students, parents, school directors-are dif- very least means increased difficulty in implementation. To fuse and not well organized, and the real educational bene- avoid resistance and conflict over reforms in El Salvador, fits are long-term in nature. Meanwhile, teachers and their the education minister created an advisory group compris- associations often view themselves as incurring the costs of ing diverse actors in civil society to advise on the reform the reform, and, as shown earlier, teachers are well-orga- effort. The result was the development of a national con- nized, vocal, and often represented by their own politically sensus on reform. skilled leaders. A major challenge for governments wishing Most reform efforts provide ways to compensate teach- to introduce-and more importantly to implement-edu- ers for what they consider the costs of reform. In particu- cation reform is simultaneously to help the diffuse benefi- lar, the emphasis on upgrading the skills of teachers pro- ciaries organize themselves for a more effective political vides the opportunity to increase teacher compensation, voice and to win over the teachers to the cause of the reform. linked to demonstrated improvement in teaching skills. Various actions are possible to help implement this Other reforms-for example, in Chile and Uruguay- strategy. The first phase of reforms created parent and entail the creation of single-shift schools, with increased school councils in many countries. Governments (or others, hours of teaching and class preparation. These reforms can such as private business groups) could help these local offer teachers both increased pay and improved profes- councils create effective national councils who elevate the sional status. views and concerns of parents, drawing national attention. The educational institutions that exist now are the The first generation of reforms increased marginally the result of an evolution in which social learning has played a role and powers of the school director. Education ministries decisive role. The "rules of the game" and the standards could, also, help these actors to organize themselves into and expected patterns of behavior cannot be changed more effective voices for further reform. Countries where except by a broad social and political learning process education has been decentralized could help organize local rather than by a mechanistic, often top-down, interven- authorities to lobby constituents more effectively for tion. For any learning to occur, there will be a need for change.3 motivation and incentives as well as continuous reassess- Another potentially effective action is the creation of ment and creativity. This represents the challenge of the effective systems of communication with parents and, espe- educational reforms in LAC at the turn of the century. 106 INSTITUTIONAL REFORM IN HIERARCHIES Notes 3. The Brazilian Council of State Education Secretariats is a good 1. While there are multiple associations, we assume, for purposes example. The state education secretaries meet periodically, on their of exposition, there is only one. own, to exchange experience and to develop common positions on 2. There are many examples in Latin America of this power-and national educational issues. The Brazilian model contrasts with its consequences for education ministers. Recently, one minister was councils organized by and managed by education ministries. forced to resign because he planned to transfer teachers across schools to ensure a minimum class size. 107 CHAPTER 6 The Challenge of Judicial Reform A S THE NATIONS OF THE LATIN AMERICAN AND CARIBBEAN REGION HAVE BEGUN TO modernize the machinery of government, reform of the courts, the public prosecutor, and the other organizations that together constitute the judicial system has emerged as a pri- ority. This has been as true for those countries in the region whose legal systems are based on English common law as it has been for those whose systems are rooted in European civil law. Whatever a nation's legal tradition, its judiciary performs a set of functions essential to the governmental process: the enforcement of the criminal law, the resolution of conflicts between pri- vate citizens and those between the citizen and the state, and the determination of the law itself (Shapiro 1981). No political system can work well without a judicial system that preserves social order, fosters vol- untary exchange, redresses complaints against the government, and contributes to the lawmaking process.1 The very nature of the judicial system makes reform difficult to realize. Judiciaries are deliberately insu- lated from short-term fluctuations in social preferences and the relative power of social groups. The aim is to ensure that judicial decisions are based on neutral legal principles, not the identity of the parties or the facts of the particular dispute. While the mechanisms that provide for judicial independence, and their effectiveness, vary greatly from nation to nation, in every instance they complicate the reform effort. Walling off the judiciary from political pressure means that popular as practitioners, civil-law judges typically are career demands for reforms will not by themselves guarantee that bureaucrats. But they share with their common-law coun- reform is initiated. terparts an independent, craftsman's approach to their Another obstacle to reform is the specialized nature of work, which conflicts with such basic management tech- law and the judicial process. In every system, law is the niques as standardization of procedures, organizational domain of the professional expert: the jurist, the academic, guidelines for prioritizing attention to tasks, or quantified the lawyer. It is this community that controls society's dis- output targets. Furthermore, judges traditionally are kings course about law and that enjoys a monopoly, or near (or queens) in their courtrooms and frequently reject the monopoly, on the knowledge required to design and imple- notion of officially delegating even logistical decisions to ment a reform project. Therefore, at least some members of professional managers or of sharing support staff.2 Profes- this community must be enlisted in the reform project for sional norms and judicial independence (however much it to succeed. violated in fact) also conflict with measures to monitor per- Professional roles and self-image also affect the judi- formance, to increase accountability, or to deal openly with ciary's willingness to undertake reforms and the pace at disciplinary and ethical issues. No profession likes to air its which they are implemented. Unlike common-law judges, dirty laundry in public, but judges, to their collective who typically are appointed to the bench after long careers detriment, seem particularly reluctant to do so. 109 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER This resistance to change manifests itself in many ways. Thus, there has been a move throughout the civil-law Judiciaries have never been the leaders in adopting modern world to substitute oral for written procedures. management techniques or new technologies, and it is not Judicial training is another example where the content, uncommon for them to be decades behind the rest of the form, and integration into the judicial career differs public sector in this regard. Arcane personnel practices and depending on whether judges are politically appointed or procedural requirements are the norm, not the exception. whether they are career bureaucrats. Most reforms attempt So is antiquated equipment; the computer, the fax, or more to increase the merit element in appointments, but devel- efficient methods for recording data are usually adopted oping the criteria for assessing merit is still debated, even late and only after considerable hand-wringing over their within individual systems. How countries define problems "legality." and what they deem acceptable remedies also affect Finally, many systemic vices or distortions are com- reforms. American advisers working in Latin America in monly encountered across legal traditions. Judicial corrup- the early 1990s found that their local counterparts were tion, a retreat to legal formalism (deciding to the letter but often less concerned with reducing delays than with mea- not the spirit of the law) in the face of external threats, and sures to combat corruption and increase judicial indepen- decisions skewed by partisan or other biases are frequent dence. While Latin Americans were interested in acdopting complaints, especially in societies undergoing rapid and more adversarial criminal proceedings (which they per- fundamental change. While often initiated by those out- ceived as more effective and less susceptible to abuse), they side the judiciary to cut individual transaction costs, they were less receptive to the idea of plea bargaining, a proce- can become part of the informal organizational culture and dure many U.S. jurists regard as critical to making Latin thus embedded in a network of vested interests that will justice systems work better. While mediation, arbit-ration, oppose their elimination. The opportunities for, forms of, and other forms of alternative dispute-resolution mecha- and incidence of corruption vary widely, but there is no nisms have now achieved regionwide acceptance, even five judicial system where corruption is unknown. years ago many judges and lawyers in the region saw them Not surprisingly, then, whether the legal system is as abominations, arguing that they threatened judicial based on common law or civil law, similar measures will integrity and basic due-process rights. improve the operation of the judicial system and the per- It is the means for introducing change, however, that formance of the judges, clerks, and others who work within seem to vary most consistently between civil- and com- it. Common prescriptions include higher budgets and mon-law traditions. In those countries where the legal sys- salaries; more staff, equipment, and infrastructure; training tem is based on common law, the judiciary is typically for judicial and administrative personnel; the adoption of more independent of the political branches of government, modern administrative practices for individual courtrooms and the judiciary itself more often takes the lead in intro- and entire judiciaries; the introduction of new categories of ducing and implementing reform. In the United States, judicial and administrative staff, revised judicial appoint- the legislature will often mandate reform but leave it up to ment systems and qualifications for candidates; and the the courts to design and apply the measures required to introduction of performance and ethical standards and implement it. In civil-law systems, particularly those monitoring as well as disciplinary systems. The rewriting influenced by French law, the judiciary is much less inde- of substantive and procedural laws is also a frequent ele- pendent (Merryman 1996). In these systems matters are ment of a reform package. more often left to the executive branch. This tends to be Despite the commonalities of the issues, differences true whether the judiciary is managed by the executive between civil- and common-law countries do produce a branch or is responsible for its own governance. This has different mix of elements and a different emphasis. Reform frequently produced conflicts in Latin American r eforms in civil-law countries, for example, has often laid greater because of their judiciaries' historical aspiration for more stress on procedural reforms. Civil-law systems tradition- autonomy in the U.S. style and a long tradition of exten- ally rely on written rather than oral proceedings, but as sive if usually illegal executive intervention in court affairs. caseloads have grown, the time and expense required to These conflicts raise a final dilemma, which in one form reduce everything to writing has exceeded system capacity. or another confronts all judicial-reform efforts. Even 110 INSTITUTIONAL REFORM IN HIERARCHIES reforms managed by a non-judicial agency, and certainly modernize their justice systems and have encouraged exter- those headed by the judiciary, usually combine two goals- nal development agencies to provide funds and technical greater efficiency and greater independence. Both inter- assistance to support them. nally (at the level of individual judges) and externally (in Whether because of the reform efforts themselves or the relation to other political and economic forces), the judi- combined weight of societal pressures, the region's judicial ciary is usually not conceived as a command system. Indi- systems have undergone a substantial transformation in vidual judges and judiciaries are expected to carry out their recent years. In the past decade alone almost a dozen coun- functions in compliance with the law, rather than because tries in the region have created judicial councils (Consejos de of the instructions of a superior or extra-judicial actor. As la Magistratura) to oversee judicial appointments and man- the Latin American case suggests, the two goals are related age their judicial systems, and several more are considering in that the judiciary's lesser efficiency is commonly blamed them. Six nations have either amended their constitutions on external intervention. Nonetheless, whatever the diffi- or adopted legislation requiring that a fixed percentage of culties of making an organization and its members either the national budget, ranging from 6 percent in El Salvador more efficient or more independent, accomplishing both at to 2 percent in Panama, be earmarked for the judicial once is extraordinarily challenging. branch. More than a dozen have established judicial-train- ing programs or separate schools to train judges, and over Recent Changes in Latin American Judiciaries the past decade virtually every nation in the hemisphere Most judicial systems in Latin America and the Caribbean has raised judicial salaries, added more courts, and intro- have suffered from a common set of failings. Resources to duced computers and other modern technologies (Correa support courts and judges have been few. Procedures for Sutil 1998). Finally, almost all countries in the region have resolving disputes have been slow and cumbersome. undertaken legal reform programs, rewriting the codes Appointment and promotion systems have been skewed by governing judicial procedures and the substance of the law. political interventions. And a variety of external pressures It will be some time before the full impact of these have often made it difficult for judges to produce neutral, changes can be assessed, but there are at least some pre- unbiased decisions. One consequence of these flaws has been liminary signs that they are having an effect. Judiciaries that citizens either have had to wait years for their cases to have become more effective in curbing human rights be heard or have been denied access to justice altogether. abuses, including those once perpetrated by some of their These problems have also fostered such public disdain for members. Individual countries have reduced case backlogs the system that many qualified individuals have been and times for resolving at least some types of cases; more unwilling to become judges, while many private actors have clients are being attended; judges appear more knowledge- preferred to use alternative, and often extralegal, mecha- able of the law, less arbitrary in their decisions, or at least nisms to resolve their conflicts. Worst of all, long delays in less flagrant in their abuses; some courts have begun to processing cases, biased decisions, and low public esteem remove corrupt and incompetent judges and administra- have created a culture of corruption that has infested many tors; countries where elites broke the law with impunity Latin American and Caribbean justice systems. have begun to try prominent citizens and officials; and Thanks to the spread of market forces and democratic judicial governing bodies (supreme courts or judicial coun- political systems, judicial reform has become a priority in cils) appear to be taking their jobs more seriously. the region. Markets require a judiciary that can quickly Still, the amount of improvement has lagged behind the and inexpensively resolve contract and property-rights dis- growing demand, and many of the traditional complaints putes. The expansion of political participation has created persist, fed in part by heightened public expectations that a demand for greater access to justice. Rising rates of con- reform efforts have encouraged. Despite the progress in ventional and unconventional crime, with negative impli- accelerating the handling of some types of cases, most liti- cations for both political and economic development, have gants continue to wait years for their cases to be heard, and also brought concerns about the systems' ineffectual the majority are frequently without any access to courts at response. These forces have spurred nations throughout the all. Reports of systemic corruption persist, as do com- hemisphere to embark on long-term, systematic efforts to plaints that the augmented budgets are not being used effi- 111 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER ciently, that at most production, but not productivity, has ing it still more difficult to modify them as their negative increased, and that access to donor funding has encouraged as well as positive consequences become evident. a proliferation of new offices and services that the recipient Finally, it is likely that too much has been asked of countries will be hard-pressed to sustain. reform. The best-performing judiciaries or entire justice This poor performance is also reflected in public evalu- sectors will not eliminate crime or gross social inequalities; ations of the system. A World Bank survey of more than nor will they compensate for inadequate physical infra- 3,600 entrepreneurs in 69 countries found that more than structure or resources in attracting external investment. In 70 percent of those in the Latin American and Caribbean short, the task was far more difficult, lengthy, and uncer- region identified the judiciary as an impediment to doing tain than initially envisioned, and it thus is not surprising business (World Bank 1997a). In Peru more than 90 per- that the results are less than what was originally hoped. cent of the respondents in a recent survey reported they had little confidence in the nation's judges, and in Argentina Improving the Current Approach the figure was 87 percent (Dakolias 1996). A recent com- There is surely room for improvement. One way is to facil- pilation of public opinion polls from 12 Latin American itate greater exchange of information among reformers nations shows that only in Uruguay does a majority trust working in different countries. Few cross-national studies of the courts (Martinez 1996). reform have been published (for a notable exception see There are several explanations for these disappointing Correa Sutil 1994). As a result, too many national programs results. First, reform takes time, and there is a still greater continue to repeat the same mistakes or reinvent the same lag between its impact on judicial operations and on pub- procedures. Donors are critical here, since they have the best lic perceptions. Where judiciaries are doing better, the access to comparative experience, an advantage they unfor- public may be the last to recognize it. Second, the very tunately don't always use even in their own projects, societal forces encouraging reform have also complicated While the recent emphasis on a systematic approach to the task. Demand is not a fixed target but continues to reform has increased the scope of recent reform projects, increase, and meeting last year's goals may not satisfy this there is still considerable debate over the relative impor- year's objectives. Promises of better, often subsidized ser- tance and timing of the different elements. Until this vices have exacerbated this problem, as have factors like the debate is resolved, or at least some progress made in that growing crime rate and economic privatization. direction, doing what comes naturally or logically will lead Third, reform has itself been a learning process, and it is to early paralysis (when the easy part is done, the reform only in the last few years that participants seem to have ends), counterproductive or redundant results (a computer come close to assembling all the pieces. The initial ten- system is installed before procedures are changed), or a loss dency to rely on a few isolated interventions (computers, a of the benefit of complementarities or synergies (while new code, training programs, or a higher budget) has been logic dictates that training programs should come later, replaced by a systemic strategy that acknowledges that there are additional benefits to doing some training early, multidimensional problems require multidimensional to overcome resistance to change and to develop more solutions, and that these in turn require the appropriate information on the state of the system). mix and sequencing. Just what that mix and sequencing Another area for improvement is better and more realis- are is still a matter for debate, but at least no one is any tic definition of objectives, both the problems to be longer relying on magic bullets. resolved and what can be achieved. Judicial reform by itself Fourth, reform has also had to confront some of these will not resolve fundamental societal ills. Furthermore, magic bullets or traditional solutions, which have hung on many reform programs contain internal inconsistencies- despite their having been discredited in practice. Most of goals which at least over the short run cannot be achieved them (earmarked percentages of national budgets, external simultaneously. For example, as a study of the Spanish judicial councils, new mechanisms for the selection and judicial system demonstrates (Pastor Prieto 1993), provid- tenure of judges) offer partial remedies for real problems, ing subsidized services for all citizens may increase the but have been oversimplified or reduced to mere formulas. level of demand so much that it will increase rather than Many have already been set in national constitutions, mak- reduce delays. It may also drive out the search for more 112 INSTITUTIONAL REFORM IN HIERARCHIES effective alternatives that could reduce the pressure on Whatever their wider social costs, the persistence of courts by filtering or diverting some kinds of complaints or dysfunctional organizations rests on vested interests, incen- encouraging other means for reducing transaction costs. tive structures, and ingrained patterns of behavior and The setting of realistic objectives must also address expectations, none of which will be reversed automatically costs and the sustainability of new or higher levels of ser- and many of which can be easily transferred to "reformed" vices. The apparent inattention to this issue has been exac- organizational settings. Sometimes this is because reform- erbated by the easy availability of donor funding. Once ers have placed too much faith in the power of technology these funds begin to disappear, national systems may find (as in the frequent observation that computers are simply they cannot sustain the special services, ambitious training used to automate inefficient systems, leading to automated programs, and equipment financed by external resources. inefficiency) or legal norms. More often it is because social In an era of cutbacks in public-sector budgets and staffing, engineers have not looked deeply enough into the factors justice institutions have enjoyed a special exemption, but reinforcing undesirable behaviors; they change the form or it is one that is already being questioned. titles of organizations without altering the pressures and A related consideration is the definition of judicial roles incentives shaping the actions of their members. An exter- and powers. As programs produce more independent and nal judicial council composed of judge-like members is no often more activist judiciaries, some negative reactions have more likely to oversee administrative matters any differ- already been registered. These are especially evident where ently or any better than the internal governing body (court activism and independence have run ahead of more basic or council) it replaced. More highly paid judges may have reforms, but they are also encountered where more profession- one less reason to succumb to bribes; the other reasons, alized judicial systems have come into conflict with executive unfortunately, are still in place. and legislative bodies (e.g., Costa Rica). It was perhaps obvi- Some of the resistance to change is self-interest, always ous that a more independent judiciary would affect the pow- difficult to overcome but especially challenging where the ers of other branches of government, and that this would lead likely losers are also the organizational and political lead- to some unaccustomed clashes. However, a balance-of-power ers charged with directing reform. When leaders commit equation that is still evolving in the United States and Europe to a reform, they frequently do not realize the full implica- is having to be worked out much more rapidly in Latin Amer- tions of that commitment. Once they do, their enthusiasm ica and the Caribbean and may be facilitated by a little more may decrease sharply, either because of what they will lose forethought as to where the respective societies wish it to go. directly (privileges, power, rent-seeking opportunities) or because of the costs they incur in fighting the resistance of A Different Paradigm others. Not all resistance is calculated self-interest. In the All of the above suggestions for improvement work within design and implementation of reform programs, the very the dominant systemic model, which even as regards its institutional constraints targeted for elimination are likely political choices still takes a fairly mechanistic, techno- to impede progress.3 Aside from the obvious problem of cratic approach to introducing change. Once the basic endemic corruption and thus the threat that reforms will choices are made, it places enormous faith in the powers of be misused or directed to private gain, these include factors formal rules, organizational structures, and technological like weak planning and management skills, a lack of innovations to reorient behavior. The current reform understanding or appreciation of non-traditional disci- agenda could be enriched by the addition of another per- plines and technologies, excessive reliance on relational spective, one that draws on the insights of institutional networks as opposed to merit or expertise in selecting staff, economics in its depiction of institutions, "the rules of the and an approach to setting goals that is formalistic or prin- game," as "the humanly devised constraints that shape ciple-driven rather than instrumentalist or results-based. human interaction" (North 1990, p. 3). This would allow Increasingly attention has thus focused on who should a focus on some of the more perplexing issues of reform: direct a judicial reform and how to ensure a continued how an organization comes to transform itself, and what "political will" to change. The obvious choices-the judi- that means for its leadership, its members, and those who ciary itself or the executive, or some kind of judicial coun- benefit from the existing pattern of activities. cil-all have drawbacks. Most likely, a better solution will 113 BEYOND THE WASHINGTON (cONSENSUS: INSTITUTIONS MATTER have to draw in other stakeholders, including the general question of greater independence is still caught up in a public and organized civil society, to create a sort of ad hoc series of traditional and seemingly oversimplified formulas. principal. But including more interests does not resolve However, before much progress is made in addressing the the problems of inadequate subjective models and other third, more technically based set of questions-on how to cultural constraints, inexperience, or more broadly inclu- reorient judicial actions-considerably more attention sive but still self-serving agendas. This is an area where the should be directed to the issues of what the reformed judi- participation of external donors may be useful in support- ciary will look like as an organization, how its objectives ing the formation of these ad hoc principals, providing and values will be defined, and how it will reconcile the them with technical input, and helping to set and enforce goals of increased independence and accountability, both the procedural rules shaping their operation. Still, if reform internally and externally. Unless these issues can be is to succeed, in addition to a good strategy, an agency or answered more creatively, the most likely alternative devel- principal who can oversee its enactment is obviously essen- opments are either total bureaucratization or independent tial, and this is equally true whether the reform enjoys irrelevance. external assistance or is entirely domestically supported. The third area, and one where institutional analysis sup- A second set of questions, deriving in part from the plies some answers as well as questions, is that of shaping first, has to do with the nature and internal governance of the actions of individual judges. One reason why so many the new judiciary itself and its relationship to its broader of the region's judicial reforms are failing to meet expecta- institutional environment. Whatever its prior level of tions is that they have largely ignored a central insight of dependence or independence and whatever the external the new learning on public-sector reform: that an organi- participation in its reform program, the judiciary has its zation is ultimately a collection of individuals with their own formal and informal rules of operation, institutional- own goals, and that unless these goals are aligned with ized roles, and culture. These internal characteristics are those of the organization itself, performance will lag. The shaped by the larger institutional setting, but they also dis- assumption implicit in most reform initiatives has been tinguish members' understandings, motivations, and that judges, lawyers, and others within the judicial system behavior from those of any other social actor. Although the are motivated solely by an interest in satisfying the public ultimate goal of reform is to transform the organization's demand for a reliable, reasonably priced justice system. impact on its environment-its output-this obviously From this it has followed that if the system is not fulfilling requires changing internal relationships and understand- this demand, it is for want of adequate resources, lack of ings and the environment's impact on them. properly trained personnel, or because of problems with It may be true that "organizations will be designed to the way the judiciary is organized. The continuing prob- further the objectives of their creators" (North 1990, p. lems many systems are experiencing in spite of almost a 73), but in the case of the judiciary the situation is far more decade of reforms argue for reconsidering this assumption. complicated-first, because the creator (reform principal) Successful reforms of other parts of the public sector show is at best a proxy for (or agent of) the public or society, but that reform is far more likely to succeed if it begins with a also because this organization and its members are realistic assessment of the incentives faced by the people intended to have some degree of external and internal who make up the system, and if it is followed by measures autonomy. A simple hierarchical chain of command that bring these incentives into line with the public's goals between the agent (judiciary or individual judge) and prin- (Boston et al. 1996; T. Moe 1984). cipal (the "public" or a court or council) will not work, and thus a more complex internal structure and set of external An Incentive-Based Approach to Redirecting relations are required. The goal at all levels is enhanced Judicial Behavior independence and accountability in pursuit of a common This section suggests how this approach can be applied to set of objectives. To date the objectives remain as unclear a central problem in judicial reform: increasing judicial as the means for setting them and the identity of those who productivity. It bears repeating, however, that applying the will do it; the mechanisms for internal and external techniques without resolving some of the foregoing ques- accountability are just beginning to be debated, and the tions-as to overall goals and internal organizational 114 INSTITUTIONAL REFORM IN HIERARCHIES dynamics-may well produce more obedience in the pur- characteristics. The first step is to identify what those char- suit of the wrong objectives. Unfortunately, their resolu- acteristics are. Judicial systems vary in the demands they tion is far more difficult, and far more political, than what put on judges and therefore on the kind of individual best- follows. suited for a judicial career. At one extreme is the Ameri- can-style judiciary, where politically appointed judges Attracting the Right Candidates wield significant power over the other branches of govern- Judges are the heart of the judicial system. Unless they ment. At the other lies some civil-law systems, where apply themselves diligently and honestly to the resolution career judges typically defer to legislative or executive of the cases brought before them, procedural reforms, com- action. The profile of a person who will perform well under puterization, and other changes to the system will have lit- the one system is unlikely to resemble the profile of one tle impact. But while many reform measures try to boost who would perform well in the other. judicial output by providing judges with better tools and One trait that would seem to be desirable regardless of more training, comparatively little attention has been the system is a willingness to expend a significant amount given to the reasons why people become judges, what of time and effort early in one's career mastering a large motivates them to work hard once they are a judge, and body of learning. Judges in any system need to know a great how this hard work translates into better performance of deal about the law to do their job well. Those willing to put the judicial system as a whole. in long hours early in their professional life studying the Job security, pay, prestige, ego, and a sense of public ser- law are the kinds of individuals whom judicial systems vice are among the reasons why law school graduates might should try to attract. A second trait would be a willingness choose a judicial career. Whether these are in fact what to forgo short-term gain in favor of longer-term rewards. attracts individuals to seek a judgeship, whether there are Like many professions, the rewards from being a good other reasons, and what the relative importance of the dif- judge-peer respect, a sense of craftsmanship, and profes- ferent reasons are, has yet to be explored in any systematic sional accomplishment-accumulate over time. Those fashion. Doing so is critical, because the rewards a job looking for a quick payoff are unlikely to perform well. offers determine the kind of individual who will apply for They are also more likely to succumb to the temptation to a position. If the current set of rewards are attracting the accept a bribe than those with a longer time horizon. wrong kinds of candidates for the judiciary, they need to be Arrufiada (1996) argues that a profession that requires altered. applicants to pass a stiff entrance exam for admission, that Brennan (1996) uses the example of tenure for univer- gives them an opportunity to advance during their career sity professors to illustrate how this can be accomplished. by scoring well on additional exams, and that pays less Universities wish to grant tenure only to those interested than what they could earn elsewhere early in their career in genuine scholarship, but some portion of those applying but more than what they could earn later, will attract indi- want tenure solely because it offers them the quiet life. viduals with these traits. The exam requirement will screen One way universities can change the distribution of true out those unwilling to master a significant body of profes- scholars and shirkers in the applicant pool is to offer a sional learning. Offering those already in the profession a lower wage but supplement it with graduate student assis- chance for more rapid promotion if they do well on subse- tants, better research facilities, and other non-cash quent tests serves as a further screen. Deferring compensa- perquisites appealing to the true scholar. Shirkers would tion until later in life will tend to select out those inter- presumably be interested only in the cash wage, and thus ested in short-term gains. such a compensation package should change the distribu- tion of scholars and shirkers in the applicant pool. Incentives for the Best Judges Just as universities could vary the package of financial Just as more research is required to determine what kinds and non-financial rewards offered to those seeking tenure of incentives now make judging attractive and whether to screen for desirable applicants, the region's judicial sys- these need to be altered to affect the kind of individual tems could likewise vary the compensation package offered seeking to become a judge, so too is research required to prospective judges to attract candidates with the desired ascertain what incentives to work hard those who are 115 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER already judges have, and whether and how they might be Disclosing how individual judges voted when decisions altered to encourage better performance. In Latin America, are made by a panel of judges could have a similar effect. FIEL (1996) has identified the factors that create disincentives Secret balloting can foster irresponsible voting, whereas for Argentine judges to perform well, but research on an disclosure, by facilitating public scrutiny and criticism, appropriate incentive structure remains to be done. Work may again supply a motive for a judge concerned with rep- conducted in the United States shows that trial-court utation and prestige to decide cases carefully. In some judges are motivated by the chance of promotion to a countries in the region decisions are issued in the name of higher court, the desire to exercise discretion in their rul- the full court and dissenting votes suppressed. If further ings, and a wish to avoid heavy workloads. To these, an research confirms the link between transparency and better American appellate judge, Richard Posner (1995), adds performance, this practice should be reconsidered. considerations of prestige and reputation, the ability to Traditionally, the chief means for holding a judge's exercise power, the deference accorded judges, and the intel- work up to public examination has been the written opin- lectual and emotional pleasures associated with resolving ion. It discloses why the judge ruled for one party and not conflicts. Given the differences between the U.S. judicial the other and the reasons for the ruling. But in many coun- system and those in the LAC region, without empirical tries in the region little attention is given to judicial opin- research it is impossible to say whether these same factors ions, particularly those of the lower courts. They usually are important determinants of judicial behavior in LAC. are not published, and it can often be difficult to obtain a Once these factors are identified, a second priority is to copy of an opinion even from the judge who wrote it. find ways to capitalize on them when crafting reform mea- Again, if public scrutiny creates an incentive for judges to sures. Some steps in this direction are already being taken. work harder, reform measures ought to lay greater stress on Whereas in the past promotion to a higher court was fre- the publication, dissemination, and public evaluation of quently based on personal or political ties, El Salvador, the judicial opinions of lower-court judges. Honduras, Colombia, and Chile are among those countries that have recently introduced merit-based promotion sys- The Remuneration Issue tems. Establishing the appropriate criteria is proving to be Ironically, although it is commonly thought that financial a challenge, but if it turns out that the chance for promo- incentives play a secondary role in motivating judges, most tion is as powerful an incentive for first-instance judges in of the attention that has been paid to how the existing LAC to work hard as it is for American trial-court judges, reward system affects judicial behavior has focused almost the effort will be well worth it. exclusively on the financial side of the equation, and in par- Besides the possibility of promotion, Posner suggests ticular on the relationship between salary and corrupt that greater transparency in the conduct of judicial busi- behavior. The theory has been that if judges are paid bet- ness is a way a judge's interest in a good reputation and ter, they will be less likely to take a bribe. Yet even on the desire for prestige can be harnessed to improve judicial narrow issue of salary level and corruption, reform mea- productivity. When judges must conduct trials and other sures have failed to apply what is known about the impact proceedings in the open, their conduct is exposed to pay has on employee loyalty to maximize the deterrent scrutiny by lawyers, litigants, the press, and the public, effect higher pay can have on an individual's willingness to and to the extent they value reputation and prestige, they accept a bribe. have an incentive to acquit themselves well in public pro- A judge offered a bribe must decide on economic ceedings. Evidence to support this link between public grounds whether the amount offered is worth the risk. The scrutiny and better performance comes from a recent report gain is the size of the proffered payment. Assuming that a of a committee weighing the merits of allowing trials in judge discovered taking a bribe would be immediat ely dis- New York state to be televised. It found that when cameras charged, his or her loss equals the amount of income that were present in the courtroom, judges did a better job of would be forgone if fired. A judge who would earn conducting trials and other proceedings, and were also $100,000 for the rest of his or her judicial career, but only likely to behave in ways that fostered respect for the judi- $50,000 for the same period if forced off the bench, thus cial system (New York State Committee 1997). risks losing $50,000 by accepting a bribe. But while this 116 INSTITUTIONAL REFORM IN HIERARCHIES fox 61~~~~~~~'fa - Measudng th eromnc fa Jdicia Syte '-'On,e -w,ay to m e'asu' re te pefo 'mance o a' judiciA'sysrm rcentage of c that fall outsid a: bn around th- t throughoPini nys. itigan o o w aver im. direct' cotact with the syster can be asked about their Other quantitative measures include the percentage of periences. n j a st y h cases reversed on ap pa anthe number of je sub -eminent of NewSouth Wales asked:255 indiuls wo to disciplinary action in. a given' pero Tea hdbroughtlawsuits intivof the error ratei o the ler whil thought trhe proceedg h been and -hether they were he latter is an indicator fdical per- -~~~~~~ soih!nert fi iit tenogeA S £et satisfied wihthe' resolution; of: their cases. Alterniatively sonnet. In: its recent study of the retn jdica -the. publc at large can be su vedout itSpdaci derivestigacones c '-the courts ard th other components of. their countrys noam' ricanas (PI'L) argued that both weretimportant justice systetm a technique the 6rld Bank sometimes meas of the peffor mance a jdic Ia system_'. ,uses when assessitg a nationts juiiary. Efforts to evaluate judl svery recent. The A - A-second way -to,-- ee'jnmeasure World: Bank ject underway to de uniform, qu i i orsf output. ata on caseloads and cr meas of jr Pr o couirt expnses- c beus todevelop measures of e h ost ou t t t speed with which courts- riesolve c d t formance Is t a rveo the Australian fiederal ivlved, icrly In in A ms state, aid territorial -cous coductd byst teering important. in icatrs of iudicial emanceis t aver- 'Co itt othe R t age delay litigants encounter before their case is resolved. stion. In is most recebt performna.nce reviewn ( data on88 o A simple-measure is the. ratio o the n of cases - co fs di a c pendingto the number aduicatedorwithdrawnina recourtsystemswscombinedwit avdiabe given od. A mote is m as Buscaglia ad ve data To me a jti Dakoias (I99) note, s i ate sigte ol W dardsfr difs fenttyp:eso cases,and then calculating the accessibe the were to ths iv n area& would suggest that a judge would have to be offered at tion and censure remains very low. Pay increases must be least a $50,000 bribe before it would be ev'en worth his or complemented by measures that beef up enforcement her while to consider it, a second factor enters into the cal- efforts, and where pay levels are already reasonable, addi- culation as well: the probability of detection. If the chances tional increases might be put off in favor of greater spend- of detection are 50-50, then the amount of the bribe ing on policing corruption. would only have to exceed $25,000 (½/2($100,000 - A second implication is that the economic incentive to $50,000)) for it to be in the judge's economic interest to take a bribe can be dampened just as much by measures that take it. Likewise, if the chances of detection were very reduce what those removed for corruption can earn after small, say 1 in 20, it would be worth accepting any bribe leaving the bench as it can by a pay raise. In the example over $2,500. above, cutting the amount a judge forced off the bench can Although this example might seem to be merely a way earn from $50,000 to $40,000 has the same effect as raising of illustrating numerically the common intuition about pay $10,000. The surest way of reducing post-judicial corruption and wage levels, putting it this way brings out income is by incarcerating those discovered taking bribes, certain dimensions that might not be so obvious. One is yet judges suspected of corrupt behavior are frequently per- that pay raises aimed at dampening the incentive to take mitted to resign quietly. The stigma of resignation may be bribes will have little impact so long as the risk of detec- sufficient by itself to keep post-judicial income low, but 117 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER where it is not, incarceration is essential to maintain the Although the primary function of a judge is to decide deterrent effect the threat of lost income provides. cases, judges may have other responsibilities as well. In What this analysis also shows is that a judge's economic some countries in the region they not only resolve disputes incentive to decline a bribe depends on tenure, age, and but confirm name changes, issue land titles, and adminis- pension income as well. Those appointed to the bench for ter estates as well. a fixed term have far less to lose in terms of future judicial As a first approximation when designing a pay-for-per- income than those with lifetime tenure. Among those with formance scheme for judges, judicial output might be tenure, older judges, because they are closer to retirement defined solely in terms of the core judicial function of dis- age, would forgo less in future income if forced from the pute resolution. This output can be measured in a couple of bench than younger judges. Similarly, the size of the pen- ways. There are several methods, for example, by which the sion affects the calculation. When the pension is generous, speed with which a judge resolves disputes can be measured. even older judges may be deterred on economic grounds How many cases a judge decides in a given period is one. from considering a bribe. The number awaiting decision, the judge's case backlog, is a Where judicial corruption continues to be a problem, second. A third is the number of cases that have been pend- reformers should incorporate these factors into their pro- ing for more than a given period, say one year or more. gram. There may be advantages to appointing judges for a If judges were paid solely by the number of cases fixed term, or replacing some or all of the bench periodi- decided or the size of their backlog, this would create an cally, but these advantages must be weighed against the incentive to resolve cases without sufficient deliberation impact such policies have on the economic incentive to just to meet numerical targets. Clearly some measure of the accept bribes. Salary differentials between younger and quality of the judge's decisions is required as well. For older judges should be also reviewed. If the wage scale has lower courts, the traditional measure of the quality of a been compressed, consideration should be given to raising judicial decision is whether it is reversed on appeal. The the pay of older judges. Where a general pay increase is more time and effort the judge puts into determining the contemplated, some or all of it might better be deferred facts and the applicable law, the less the chance a higher until retirement. court will find reversible error. Thus, to offset the bias that would be introduced if only speed were considered, the Boosting Productivity percentage of cases reversed on appeal could be made a part While level of pay and the way total compensation is dis- of an index of performance measurement, too. tributed over time can help deter corrupt behavior, experi- Good judging, however, consists of more than the ence from the private and the public sectors shows that pay speedy and accurate disposition of cases. The way a judge levels can also be manipulated to boost productivity. A treats lawyers, litigants, and witnesses can either promote common way of motivating workers in the private sector is respect for law or foster cynicism and disdain. Such intan- to make their compensation depend on how hard they gible factors as temperament and demeanor a.re thus work. Rather than paying a fixed wage, the employer important, too. Although these can't be measured directly, rewards them on the basis of output. Public-sector reforms surveys of the lawyers, litigants, and witnesses that have have adapted this technique to government bureaucracies. appeared in the judge's court is one way to take them into In England, for example, teacher pay is linked to test account. scores. The better the students do, the more the teacher earns. Drawbacks Judicial pay could also be tied to output. Those judges All these measures have their drawbacks. Collecting and who produced more than their colleagues would be paid maintaining the information needed to measure the speed more, or be given bonuses, while those producing less with which judges dispose of cases is costly, and until the would be penalized. Linking judicial pay to judicial per- recent wave of reforms few judicial systems in the region formance requires measuring judicial output, however, and took the time and trouble to gather it. Also, no matter how this exercise is not nearly as straightforward as tying pay to hard and how efficiently a judge works, his or her caseload factory output. First of all, what is it that judges produce? may grow because the legislature failed to add new courts, 118 INSTITUTIONAL REFORM IN HIERARCHIES because of a rise in the number of cases filed, and so on. Iso- implemented is still under consideration. The approach is lating those variables that are under a judge's control from new and the hazards outlined above suggest that perfor- those that are not may not be possible. mance-pay schemes must be approached with caution, but The rate of reversal on appeal can be misleading as well. no serious reform program concerned with judicial produc- A judge may do a poor job of deciding a case but reach a tivity should overlook the possibilities they offer. result the appeals court agrees with by chance. The cases appealed may not be a representative sample of the judge's Conclusions work. And even when the judge is reversed, it may be for The new learning on public-sector reform is still in its reasons outside his or her control: changes in the law, mis- infancy, and many of the techniques it has borrowed from takes by the appellate court, and so forth. the private sector are still being adapted to the unique cir- There are problems with user surveys, too. It may be cumstances found in the public sector. This is even truer difficult to get a representative sample of litigants or wit- when it comes to applying them to the judiciary. But this nesses to participate in the survey. If lawyers were to rate chapter has suggested that the task is not impossible. judges, judges might begin handling cases in ways that There are surely risks-including the danger that a judi- pleased the attorneys, say by drawing out the proceedings ciary will come to resemble a tax-collection agency rather to increase fees. than an independent branch of government. However, Combining these different measures into a single index assuming those shaping the reform have nobler objectives of judicial performance also raises questions of weighting. in mind, a little demystification of the judicial role might Is it more important that judges process cases quickly, or be just what is needed. It may also help address some of the that their rate of reversal on appeal is low? What is the more transcendent questions as to the direction of reform, tradeoff between the two? How should low scores on both who will oversee it, and what the reformed judicial organi- be balanced against high ratings on user surveys? zation will look like. The job will not be easy, but if the Despite all these difficulties, a performance measure result is a better functioning judicial system, then surely that combined survey responses with rates of reversal on the effort will be worthwhile. appeal and one or more measures of speed could well pro- vide important information about judicial productivity. If Notes nothing else, the scores of judges sitting on the same court 1. Even in France, whose legal tradition formally denies judges could be compared. Those with high rates of reversal, large any role in the lawmaking process, it is now widely acknowledged backlogs, and poor.scores oDuser sur-veys are almost cer- that the judicial function inevitably entails lawmaking (Salas 1998). backlogs, aind poor leseffores ohn uhoser o s s arel amosthcer- 2. Of course, unofficial and even illegal delegation is a common problem, but aside from denying its existence, judges usually prefer indicators. to retain formal control. Only very recently have policy-makers in the region 3. North (1990, p. 111) observes that "ideas and ideologies" considered making judges' pay depend upon their perfor- shape the way individuals "interpret the world around them and mance. FIEL's 1996 study suggested using a combination make choices." A more concrete application became the official jus- tification for Peru's current reform program, much criticized for its of performance measures to rate Argentine Judges. A recent erosion of judicial independence. As President Fujimori noted in Chilean statute provides that a portion of judicial pay is to confronting his critics, "When we begin a reform, we don't consult be based upon performance, although how the law is to be with its targets." 119 CHAPTER 7 Public Administration in Latin America and the Caribbean: In Search of a Reform Paradigm TE HERE IS NOW A GROWING CONSENSUS, AS WAS REFLECTED IN THE LONG MARCH, THAT public administrations urgently need to change if the process of economic and social reform in the LAC region is to succeed. The public administrations of many countries of the region are typically inefficient, unable to deliver services to the most needy, and bastions of oppor- tunistic behavior. Without effective public administration, how can services be delivered more efficiently? How can governments regulate effectively? How can they reach the poor? And how can they function in a way that does not threaten fiscal discipline? The need for reform may be indisputable, but the way to do it is not. Governments and donors (partic- ularly since the 1 980s) have put a good deal of effort into reform, in LAC as in other areas of the develop- ing world. But no clear reform paradigm has emerged, certainly not to the extent it has in such areas as macroeconomic policy, market regulation, and privatization. The success stories are limited, and the jury is still out on how effective the current variety of reform efforts in LAC will prove. The World Development Report 1997 provides a good tour of the state of the art of reform. Its perspective is institutional, and it takes a realistic view of the problems. Box 7.1 provides a pr6cis of the report's main messages about reforming the state and its administrative apparatus. This chapter uses an institutional approach (see Chapter administration characterizes the "models" of public admin- 1) to examine the performance of public administrations in istration that have emerged in the more advanced countries LAC. Public administrations are those permanent bodies to tackle these problems, contrasting a classical, "hierar- within government that execute policies made by a politi- chical" model with an emerging set of ideas and practices cal executive or by a legislature. The chapter's main con- coming under the umbrella of the "New Public Manage- cern is how public administrations do their work, rather ment." The new ideas are the fruit of a lively debate that than what tasks governments should perform (and how has been cross-fertilized by ideas from other fields, espe- they should devolve tasks). Its concern is primarily the cially theories from the new institutional economics. Pub- national public administration, although most of the issues lic administration in Latin America and the Caribbean and the are also relevant for subnational administrations. regional reform experience examine the LAC region's experi- The problem of pubhlic administration: an institutional per- ence of public administration and its reform, and how it is spective, the first section of the chapter, outlines a frame- difficult to fit into any of these models. In fact, in many work for diagnosis of the problems of effectiveness, effi- LAC countries the rules are not what they seem. The ciency, and responsiveness that face public administrations. underlying informal rules that guide the public adminis- This framework relies largely on an institutional perspec- tration are the product of institutional conditions that typ- tive in which politics play a central role. "Models" of public ically differ from those of the more advanced countries. 121 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER The reform models of the more advanced countries thus PMbtIcSe _tor _sfrnmr A Wai~kproviprovide an uncertain guide, and as shown in reform: con- straints and possibilities, there is still much to learn about il'he1997W'o IdD dqnntnz 1Zep~rt was the first wh6lly effective methods of reform of LAC's public admLinistra- tions. What is clear is that politics are central. frcutgThe Problem of Public Administration: An *' ethnk he oleof he tar, wat t podues' Institutional Perspective and hw- it prodces. .he state soul where. v ........Good democratic government provides the products the poss}ble be a marker facilitator electorate wants effectively, efficiently, and honestly.i The Mthhrlefhesetoica ty problems in doing this can be grouped under two head- usethe resources of the market and of civil society: ings, monopoly and the "principal-agent" problem. Weak s~&~es sha~ildcondciimre oki. _ht' Jlifidamen- The monopoly problem is straightforward and well t e ~righ±s, a c s.........................ty...... known. The government mostly provides services for which it is a monopoly supplier. This is by definition the case with ti6n of the destitute, ~public goods: The government's monopoly of legitimate force makes it the dominant purveyor of defense, internal Promoing acental caacityfor ormulting security, judicial services, and so on. Governments alIso have and coordinating polcy .dominated the production of many other services, such as Promoting eficient and effetive service-d y health, education, and communications. Monopoly provides systems rhinugh a^ varieyofmeansfperfornan62 an incentive to produce inefficiently. oriented measures, better nierirocracies better The principal-agent problem, which occurs in market information, strengthening voice ) taored 'tot contracts as well as in contracts within hierarchies (see the circumstAnces (if the activity and the ~ountry Chapter 1),2 is more complex. It arises when the agent Fighting corruption tthrough A miiitipionged (who is being compensated to provide goods or services to approach including reduced regulation, grearer the principal) has information about product quality (or transparency and> ^ use of markets~ and, bettA changing states of nature) that the principal does not. enforcemenjy<,t v.-. Because the agents wish to extract from the principals as Promoting otivagted and§}0 capable sgg3rg;ff 0 l much net value as they can (in the form of rent, ability to using appropriate compensation Policies ansI shirk, political power, and so on), the asymmetric informa- biuldingesprit dt,corps n gYgtion problem enables the agents not to act in the princi- Thi.k how toseqiience reforms pals' interest. In the public sector, the principal-agent relationship is f c at ..........................e....... rea ''e.t ,encompassed by the long chain of accountability that runs Be cautinijuiri introducing the nA~ tec~niqiies. from voters, through political representatives and public oyfy public management or instane fullblwn.o bodies, down to public bureaucrats. There are thus at least contrctin witin te pulic ecto migr be two stages to the principal-agent relationship in democra- tic government. 3 In a first stage, voters are principals, and flexibility will he y S 4 possible only when effecti political actors (individuals, parties, and elected officials) inutco tol are in.S","s'2:2 ' placeare their agents. In a second stage, elected politicians (and While an effective professionalized civil setsice. i parties) or government actors (such as presidents and min- isbeitig put iAplae.S i~ipr6ve p~Iicpmakipg isters, legislatures, or judiciaries) are principals, and the and ci~qtdin~gion; etc po~sible use n~rket public bureaucrats or their agencies (such as ministries, - 2 SYS>S> - mechanEIi si ni, Zsu as. c6sisfi~ outsand statutory boards, or state-owned enterprises) are their inecliasiismso . t i . agents. At this level of generalization, the two-stage prin- cipal-agent problem is the same as the well-known prob- 122 INSTITUTIONAL REFORM IN HIERARCHIES lem of the shareholder-owned firm in which the sharehold- Third-and related to the information asymmetry- ers exercise their ownership rights, through a representa- rules and practices of political parties affect their respon- tive board, over the firm's managers who in turn have to siveness to voters. Candidates selected by "closed" parties manage their employees. dominated by oligopolies have greater latitude to act in In Stage 1, voters' control over politicians is usually their own particular interest, or those of the interest groups imperfect for several reasons. First, there is a problem of that provide financial backing, rather than in the interests collective action (see Chapter 1): lt is difficult for individ- of voters. Candidates chosen by the broader membership of ual agents to cooperate to further their common interests. "mass" parties will be able to change party leadership, and Second, the political institutions purporting to represent thereby influence the party's policies, presumably in the voters-the political parties-also face problems of collec- direction of the general interest. tive action. Third, voters' information on the appropriate- Finally, successful collective action is "path-dependent" ness, quantity, and quality of public goods (and technolog- (North 1990): Different countries start from different his- ical options for producing them) is imperfect. torical situations; informal norms change more slowly than In Stage 2 it is similarly difficult for politicians to con- formal rules; and success is achieved after repeated trol public employees. The particular difficulty lies in the attempts at collective action. So, even if formal political problem of defining and measuring with any precision rules can be easily written, the practice that ensures com- many of the outputs of public administration (a problem pliance with the spirit of these rules takes time. Thus, greater than in the private firm)-hence the difficulty of democratic practice will take time to mature. monitoring the performance of employees. It is not easy to submit typical government goods and services to the test Special Interests and Other Problems of the market.4 To the extent that voters lack control, special interests- economic, political, or civil-become important. They Constraints on Collective Action of Voters can, typically through campaign financing, influence Collective action by voters seeking to have politicians rep- politicians (Stage 1) and they can, by "colonizing" govern- resent their interests (Stage 1) is difficult for several rea- ment agencies (through money or influence) limit politi- sons. First, voting rules that encourage fragmentation and cians' control of the public administration (Stage 2). multiplication of parties serve particular interests (provi- Where democratic practice is weak (i.e., where voters have sion of private goods), rather than the general interest (pro- a collective-action problem) and where societies are frag- vision of public goods). The more the fragmentation, the mented-notably by economic, social, ideological, or eth- more parties will represent particular rather than general nic differences-public goods, including good public interests, while in order to create governing coalitions, par- administration, are under-produced. This is akin to Olson's ties will trade particular interests (Cox and McCubbins (1993) argument about the conditions giving rise to "rov- 1996; Geddes 1994). ing bandits," rulers whose hold on power is sufficiently Second, information asymmetry may harm the "true" weak that they consume public revenues rather than invest interests of voters and lead to a clientelist system of trad- them in producing public goods. ("Stationary bandits," ing personal favors for votes. Low voter education and poor rulers with a better hold on power, have an interest in pro- communications and information, including deficient viding some growth-inducing public goods so that they media, will enable politicians and parties to win votes by can continue to tax in the future.) This underproduction of providing private goods, such as access to state benefits, at public goods occurs for two reasons: First, fragmentation of the partial (though not complete) expense of public goods. power, as we already argued in the case of party fragmenta- Politicians and political parties need an army of people to tion, leads to the trading of particular interests in political provide these favors, and they need to reward this army in decision-making. Second, when fragmentation expresses return, typically with government jobs (Geddes 1994; see itself in terms of political instability and discontinuity over also Box 6.2 of World Bank 1997a, which describes the time, politicians, mindful of their low probability of stay- "political machine" that characterized many U.S. cities in ing in power for long, will look to providing private goods the nineteenth century). that create a short-term advantage for them, rather than to 123 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER public goods that would be more likely to assure their the professional public administration. Other institutions, longer-term support by the citizenry. Constitutional notably the judiciary and a public comptroller, provide arrangements often have "veto points" in the name of independent, external oversight. checks and balances (Cox and McCubbins 1996), such as Hierarchical and centralized forms of organization of the presidential-congressional checks, judicial review, and public administration ("command-and-control") are meant term limits on presidents. These checks and balances exist to provide a means of creating professionalized public ser- in the name of prudent governance, but they also often vants who are responsive to a broad political mandate, but reflect competing interests that do not trust each other. who are insulated from individualized political influence. In In addition to the problems posed by interest groups, its idealized form, more or less as Weber (1968) defined it, Stage 2 problems are exactly the same as those faced in any the hierarchical model is characterized as follows: organization. And the larger the organization and the more * Functions are organized into specialized units where difficult its outputs and behaviors are to monitor, the accurate financial and technical information flows up, deeper will be the problem. Organizations seek to over- down, and across. Commands are obeyed, agencies come the problem through arrangements that determine cooperate, and decision-making is delegated to the how authority is exercised (the nature of the implicit con- appropriate level (including decentralization to polit- tract with employees); offer rewards; promote loyalty to the ically independent units of government). organization; and promote standards of behavior (Simon * Qualified personnel are employed on a strict merit 1991).5 The difficulties of imposing these kinds of arrange- basis and under rules that relate pay scales to defined ments are no doubt affected by different "cultural" traits. jobs and provide open and fair procedures for hiring, Notably, in societies where social capital-"norms of reci- firing, and promotion. Typically, they offer a system of procity and networks of civic engagement" (Putnam 1993, lifetime tenure and other non-pay incentives. p. 169)-is greater, it is likely that the management of * Budgets (planned income and expenditure) are accu- organizations will be easier. rately forecast, and financial systems check that expen- ditures are made honestly and in accordance with these "Models" of Public Administration budgets. - Records are accurate and procedures are codified. The Hierarchical Model In general, the hierarchical model has worked well in the The first modern model of public administration came into more advanced countries to circumscribe the freedom of being among the modernizing countries of the North politicians and public servants to act outside the public Atlantic area in the nineteenth century. It developed as a interest and to create a professionalized public service. response to prevalent political interference, corruption, and Although political interference, corruption, incompetence, lack of professionalism-honesty, discipline, and compe- and so on, do exist in the public administrations of these tence-in the then-public service. This now-classical countries, generally they have honest and competent public model, which we will call the hierarchical model, is still administrations delivering a wide range of public services. broadly intact in its essentials. It uses two principal-agent The hierarchical model is nonetheless under extreme "techniques" to foster more efficient and honest govern- pressure. It was born when government was small. In ment in the public interest-political checks and balances, 1890, less than 10 percent of expenditures in today's and a hierarchical form of organization with centralized OECD countries came from the public sector. But the role control. of the state expanded enormously-until the 1980s, at Checks and balances are usually embedded in a consti- least-and by 1995 this share had grown to almost 50 per- tutional separation of powers, which provides for special- cent (Figure 1 of World Bank 1997a). This growth has led ization of functions, "second opinions" in decision-making, to large public organizations difficult to control and easier and oversight of behavior. Elected politicians in the execu- for special interests to "capture." It has led to inefficiency tive and the legislature have a dominant role in making and inflexibility. And citizens, more distrustful of govern- policies.6 The politicians oversee the execution of policies, ment than in bygone times, have expressed increasing dis- but they are bound by rules that limit their interference in satisfaction. 124 INSTITUTIONAL REFORM IN HIERARCHIES The New Public Management It is premature to talk of any consensus, but there is a For these reasons all the advanced countries have experi- coalescence of ideas around a promising new model emerg- mented to modify the hierarchical model. The same checks ing from the debate. We shall refer to the new approach as and balances remain, and the experiments have been on the the New Public Management.7 It has evoked widespread hierarchy itself. But it is difficult to identify any unique interest in LAC, as well as other developing and transi- pattern (see Laking 1996). Governments have constantly tional economies. The emerging model, as it applies to experimented in new budget techniques, but not always public administration, can be characterized along four successfully-indeed, new techniques have often led to a lines, whose common theme is to borrow from the man- loss of budgetary control. They have exploited new infor- agerial methods of the private sector:8 mation and communications technology. And more * Devolution of decision-making. Reducing the burden of recently, they have exercised greater control over the hierarchical rules and fostering greater discretion at growth of public employment. lower points in the hierarchy-operating agencies, Now, however, two particular reform themes-one con- regional agencies, subnational governments-allows cerning what governments do, the other how they do it- decisions to be made effectively because decision- herald the emergence of radical new ideas that move in the makers are closer to the problem and have clearer direction of a new model. objectives. • Virtually all advanced-country governments have * Performance orientation. Changing the accountability experimented with increasing "contestability"-i.e., relationship from an emphasis on inputs and legal by putting activities back into markets or simulating compliance to one on outputs provides incentives that market conditions. "Contestability" includes privati- lead to greater effectiveness. zation and corporatization, voucher schemes (where * Clientfocus. Reporting to and "listening" to the clients the state finances a service, typically education, but of the public sector allows governments to understand the market provides it), contracting out, having gov- more systematically what citizens want and to ernment departments charge for their services (even to respond with more appropriate outputs. other departments), transferring funds to users, and * Market orientation. Making greater use of markets or devolution of activities to lower levels of government. quasi-markets-typically through management and * A smaller number of governments is pursuing a range personnel contracts, competition between public agen- of options and experiments to change the way they do cies, inter-agency fee-charging, and outsourcing- business (see Reid and Scott 1994; Holmes and Shand improves the incentives for performance orientation. 1995; Nunberg 1995; Laking 1996). A ferment of As these kinds of reforms are progressively applied, they new ideas focuses on improving performance-by are typically intended to change the organizational charac- switching the emphasis from controlling inputs to teristics of public administrations in the direction of: controlling outputs, thereby allowing greater manage- * A more strategic orientation to policy-making rial discretion-and on promoting greater trans- * A change in the way agencies are arranged, for parency and accountability. The new ideas have been instance splitting policy-making from implementa- serviced by a lively academic debate reflecting com- tion and financing from service delivery peting approaches. A public choice approach empha- * A financial-management system that emphasizes out- sizes the need to constrain bureaucrats' freedom puts, provides a full costing of inputs and outputs, through top-down controls. A principal-agent and decentralizes ex ante expenditure controls; typi- approach concentrates, somewhat to the contrary, on cally, controls become less specific (broader budget the use of incentives and information to allow bureau- categories, for instance) and external controls are not crats more freedom (see Aucoin 1990 on the tensions eliminated, but shift from ex ante to ex post, and between these approaches). Some authors have exam- * A decentralized personnel system placing larger ined the limitations of the new ideas in addressing the emphasis on rewarding performance. problems of advanced countries' public administra- Since the 1980s, most industrialized countries have tions (see, for instance, R. Moe 1994 and Savoie 1995). advanced on various fronts related to the new model. A 125 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER small number of them, most prominently Australia, New rather than public-sector reform per se.10 Paul (1998) Zealand, and the United Kingdom, have carried out large, describes the successful use of a "report card" to provide integrated reforms in the spirit of the new model. Iceland, feedback on public services in Bangalore, India, and Sweden, Chile, and Singapore have also undertaken reform thereby to apply pressure to get services improved. Using in a similar vein. Australia, for instance, has redefined cab- "voice" shares some common ground with the customer- inet, ministerial, and departmental management responsi- orientation shown in many New Public Management ini- bilities to create stronger central control (th-ough strategic tiatives, such as the Citizen's Charter in the United King- decision-making, budgets, and evaluation) and to give dom. But its proponents tend to emphasize its use where greater managerial discretion to departments. The United populations are politically marginalized-in our terminol- Kingdom has so far devolved two-thirds of public employ- ogy, where the principal-voter cannot control the agent- ment to executive agencies in order to separate policy-mak- government. ing and implementation and to provide those agencies with greater managerial discretion related to performance. Public Administration in Latin America and the The New Zealand case is of special interest. It is the coun- Caribbean try that has applied the model in the most complete, con- ceptually rigorous, and integrated way (Box 7.2).' The Performance of Public Administrations Reforms in New Zealand, Australia, and the United The countries of LAC have more or less copied thc hierar- Kingdom are reaping many benefits of greater efficiency chical model in creating their own public administrations. and better services, though not always without controversy. But the hierarchical model has not worked well. In fact, New Zealand and the United Kingdom are the only OECD the dominant diagnosis is that the region's public admin- countries that have been able to cut core public-sector istrations are overcontrolled, and most of them remain employment in the 1980s and early 1990s. Surveys in New poor in the delivery of public services. It is thus not sur- Zealand and Australia generally support the view that the prising that the techniques of the New Public Manage- benefits of improved efficiency outweigh the costs. But it is ment have attracted attention. too early for a comprehensive verdict on the desirability or A starting point is to ask how well the governments and feasibility of the new approach. public administrations of the LAC region have responded Some advanced countries remain wary of the New Pub- to the crises that have beset them since the 1980s. The lic Management (see Nunberg 1995 and Laking 1996). answer, which has to be impressionistic, is mixecl. Many Japan, Germany, and France have been careful to maintain governments of the region have a reputation for the poor basic bureaucratic traditions, though with some piecemeal quality of their services and excessive regulation. But there reforms in the direction of performance orientation (such as is another side: The region has made substantial achieve- more intensive use of formal evaluation of existing policies, ments in economic stabilization. It has also done much to greater use of performance measures, at least in reporting, change the role of the state through structural reforms that decentralization, more extensive contracting out of con- emphasize privatization and deregulation. Many countries testable services, and more attention to quality of service). have developed effective regulatory institutions, in,cluding They have maintained a hierarchical tradition on the autonomous central banks, supervisory agencies in finan- grounds that it is necessary to maintain an ethic of public cial markets, and agencies to safeguard competitive mar- service, equitable administration of public law and policy, kets and regulate natural monopolies. And these countries efficient management of the public service, and control of have considerably decentralized their governments. The the size of the public sector and of public finances. region is the only one in the world to have reversed the growth of government in recent years: Public expenditure Voice in the countries of the LAC region averaged 14 percent of The New Public Management emphasizes a client focus. GDP in the early 1960s; it rose to 26 percent in the early One way it does this is through what we call a "voice" 1980s, but fell to 22 percent in the first half of the 1990s model, which emphasizes improved public-sector perfor- (World Bank 1997a, Figure 1.2). But this may be as much mance through "listening" or "participatory" techniques a sign of fiscal crisis as it is of fiscal responsibility. The over- 126 INSTITUTIONAL REFORM IN HIERARCHIES New Zeutandt A Leafing Exanupl of NewkPubic o Mngmenit he: goverent that cat w in New Zealand in. as they wish.: E i h p s n 1984 fatcin~g a deep economic cris, undertook a program work uniderpiatao lawf (and un der the same 4->Lih*--i am - : priat i labo .-. tun a-,osia afto'-, ,' '' tlye ta fdameta acroeconomic and structuralrefor bt c i a a iva fir it still found itself saddled with a t, cent, . Cif - -ireponsiepbc setr The firs Ste was.'toisitt exctie -arersosbet.iitr,' ne anua widespread program to corpor tize1 thn privatize a se- partofthe pu bli srector tte, mae epoible -for outcomes: ( he Near e.-with the aidof ne-w toiofe apublic mnge effet o ftheseon th comu ni).Ministersare free -ie goernment mbt fiom 1988 a rocess of core . u i o oh p i o public-sector reforem more radicalthan h been seen in any efecuties t an capi e ' Oti '- SXw ' dU££t - t s '- -'re.66 . 're d ' e,e,ld'S orhet~ country duigthscntr.The reom ee drreiIonsT avoid conrflict of interest,plc inspii-redbythe idea that th incentiv inthe pblic sector, adc a v dever have m- ostly been ' sepa * -were.wrong and that they could beimproed b......y repicating:.rated i'roQdifferenr agencies.. -what wafound 'in the private sector. TMe reform ias bn Re 'mwrirg, and cordinatin. 'Mni dominated by the idea of establishing contract-like l p th q qo rions between *t goverment as purchsers nd gohvea - e e a m ef -ex uives -anro ff'ent agencms, as su iers (iC.e, the S 2 princpal-a budgeting.and regutlar fnancial reporing o rela,tionshipdescribed in this chapter). accrual exactly. as in thep d- the New Zealand public-sector reforms are ca racter-' -in the treatent ofasset a' de iato M . ed by the following main elemen tees coorinate strategic pcb ei Ac iy and n ye re n Dep t (i.e, pub isI outcomes, £ hrough spcial- -mityisrYheads lose th lifetime re'te ad no i interministerial committees, a-id by, sp ttig wo-, rk a, s- ch-ie exeecutives with specifi perormnance the i of oi a contracts negotiated with the relevant mhinster. The- : department heads a ree to run their departmnts(199) d Schick (1998) all impression is of a set of countries that, with some excep- What lies behind this poor performance? Again, a tions, has had the capacity to dismantle much of the old familiar, anecdotal picture is painted in different studies of state, but has not had the resources or capacity to recon- an overcentralized, overregulated, overrigid, under-moti- struct a public administration appropriate to the new state. vated public administration: The problems of government are reflected in the mea- Many studies of bureaucracy and public administra- surement of investors' perceptions of the quality of gover- ton in Latin America conclude that certain charac- nance in the region. In terms of three indicators of foreign teristics appear to be persistent despite constant investors' perceptions reported in Chapter 1-the level of reform efforts. Among the frequently cited charac- corruption (Figure 1.5), law and order (Figure 1.6), and the teristics are excessive centralization of authority, quality of the bureaucracy (Figure 1.7)-the average score for the countries of the LAC region in 1998 is low. Even superisory insalylegalism,nadeq t com - nication, and incomplete staff management (Hop- though the indicators have seen notable improvement kins 1991, p. 701). since 1984, they still show levels of governance compara- ble to those of Africa and the Middle East, and well below Most Latin American public sectors are bedeviled those of East Asia and the industrial countries.l with counterproductive civil service policies and 127 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER practices that impede their efforts to recruit and most have neither the capability of pursuing collec- retain highly qualified employees and motivate them tive goals in a predictable, coherent way nor an inter- to perform to the best of their abilities. ... Exces- est in doing so. sively rigid procedures and practices impede the ability of public managers in LAC to significantly We shall characterize this difference between the appar- affect the performance of their employees. ... Finally, ent rules of Weberian hierarchy and the real rules as "infor- the institutional structure within which the public mality." Informality occurs when actual informal bureau- sector managers must operate in LAC is often so full cratic behavior does not correspond with the formal rules. of red tape and overlapping responsibilities that it is Bureaucracies often may appear to comply with rules, extraordinarily difficult to take any significant while in reality they are broken or bent. In effect, the rule actions, let alone effectively manage complex pro- of law is undermined.12 grams and large numbers of employees (Reid and Informality can be characterized in terms of the main Scott 1994, p. 39). functions of government-decision-making, procedural rules, and personnel and financial-management systems It is not surprising, in light of such views, that the dom- (Box 7.3 provides the example of informality in Peruvian inant diagnosis of many LAC governments is that their government): public administrations suffer from overcontrol-a diagno- * Decision-making processes are concentrated among a sis shared, to an extent, by the international financial insti- small number of people and agencies, and the benefits tutions. This diagnosis says, at least implicitly, that the of delegation (and of contact with end users) are for- traditional approach to public administration has failed. gone. This reflects an absence of cooperation and trust. This is why the new market-based approaches to manager- And yet decision-making is also fragmented; techni- ial flexibility are attractive, and, thus, there is a close par- cians at lower levels often hoard technical information. allel with the diagnosis that has been made for the public * The formal rules of procedure are excessive, because administrations of more advanced countries. managers do not trust public employees. Yet the rules have limited effect, either because an excess of rules Informality in the Public Administration means they are contradictory (sometimes reflecting a For the LAC region, however, this diagnosis appears pre- legal framework that has been designed in pieces that mature. In particular, it does not take account of the insti- do not fit together; very often, when one law is seen to tutions in the region that drive performance. The adminis- be ineffective, another is enacted, typically with insuf- trations of LAC possess formal rules and structures very ficient effort to annul the first), or because there is no much consistent with the hierarchical (and centralized) effective mechanism of enforcement, and written rules model prevalent in more advanced countries. But actual can simply be ignored. One effect is to minijmize the bureaucratic behavior is different. As Evans (1992, pp. number of administrative interactions (i.e., transac- 176-177) remarks: tions). Another effect is to encourage informal solu- tions. If line-ministries do not work, new agencies Bureaucracy is in under-, not over-, supply. This is outside the formal administrative structure are cre- not only a problem in the post-colonial societies of ated. If career public servants are ineffective, ad hoc the sub-Sahara. Even in countries like Brazil that appointments are made to do the job. Another effect, enjoy relatively abundant supplies of trained man- of course, is to encourage opportunistic behavior- power and a long tradition of state involvement in corruption and shirking. the economy, predictable, coherent, Weberian * Merit-based personnel rules are circumvented in favor of bureaucracies are hard to find. The standard percep- procedures that allow employment for reasons of tion to the contrary flows from the common tendency patronage or personal trust. Typically, the merit system for patrimonial organizations to masquerade as is used as a cover to practice patronage or to choose Weberian bureaucracies. There is an abundance of employees who will be personally loyal. Or else the sys- rule-making or administrative organizations, but tem is circumvented through ad hoc appointrnents. 128 INSTITUTIONAL REFORM IN HIERARCHIES BOX 73' Infrni In thea Peruvian e; Two or three years after t rok -offic in 19901 a new dry appoin1tments are imporant pa government iniPeru implemented a sweeping program cicularly in- t key posts, in t absence of efetive -of economic, -social, and politial change. The govern- r t p t The men- achieved some major steps in re-establishing cred- wak c c wit t-promoting tles in the ible and effctivegoo en - The szeandsco f ci th s ru re public sector have been reduce to fir norem semblace of ty proporttins, ad nature of goerent rm has shifted m Third: th budgetand -ttm substlly from ownership ad interventon to market is the cunte f a r regulation and provision of services targeted to the poor, s trte hi entrafized system weaknsh ci ut -most pubic s.ev ices are stiil and he procests;t it c Reforms of the public sector hav been highly selectiive overight and a cash-m.anagemient bypassing existing ineffective agencies to conente sys l sh rent to concentrate on its a few - privileged agencies (some ministries a sdome cs targets. autonomous agencie ) that:have:spearheaded the reforms In an effet tensions etween writtenlaw of the government. and inform pts have had th The Most evident feature of Peru' system of Publ to correct mani fet d eects b administratio- one that h eas nt chaned with the. mare laws, very often wihout poerly eIod ne: gewoverirnment-is-that the de- jure system is in reaity lws. This ha led- ' i' o -mostly overtaken by a very different de facto, or. iformal, ywhih unermidnles ther One nifestation oi f system. A ifirst lev of faw (the constitutio a leadin isistr government. ls) precribes an i`deal fbrm" of public ainistration For instance difent minuistries, autonomous ,a inthe- imagefthe adnced iindstrial coties, atca- and I hae fli n coornatedjui t:retized by checksandbalances, public accountabil ty, lictions in' th same secialy in the area o hierarchical form of organizton, and a professionalized s and ifirastmct investmti). Ther is a public; service But the first' lvel olaw is systetically- ivious cIrcle of ki ane controls are adedt circuimvnted. by second-level law; sometimes these are eistin lislatio suc asperoel oY, rbueta s- actual law, butrusualy the are lowet-level imptement-. lation, t an effort to mae it: work at ouh. BUt the hag rules. This second lejvelproduces, in effect, elements outcome is, it ad, a regulat ry jungle. In addition, laws of paratlel governm t with sever main attributes: are often poorlyenforced. Pont enomnta - irst, ' a. parlel organizationd stcture (eseiy son feed off eahother, and their c ation povser- - key autonomous agencies) bsanti passesthe ily ud ese ca4binet atn ministerialistructure. e I I . - erallyapliee publ aminisration The management rules for setting and executing bud- description of the vicious circle of fiscal uncertainty gets often have little meaning. Many public initia- and informality characterizing budget systems in tives, such as tax breaks, escape the budget.13 Budgets developing countries. are unrealistic; unexpected increases in spending and Informality is an institutional arrangement with sudden decreases in revenues are the norm. Informa- national costs and benefits. Opportunism (corruption, tion from agencies on expenditures is inaccurate, fraud, patronage, and rent-seeking) and inefficiency flour- often deliberately so. Thus the budget as executed ish and are "institutionalized," while the costs of bureau- does not resemble what was originally planned. Box cratic transactions are raised. On the other hand, informal- 7.4 provides Caiden and Wildavsky's (1974) synthetic ity allows some essential transactions to be completed, in 129 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER spite of burdensome rules. There is often a vicious circle the advanced countries, and some Latin American and whereby the failure of the state breeds more corrective Caribbean countries are doing better than others. rules, which both reformer and opportunist applaud, the Chile, for instance, has reduced informality thtough a reformer under mistaken formalistic notions about how to long process of construction and reform, and tocay it is reform and the opportunist in the knowledge that reform close to the practice of the more advanced countries (see will be frustrated and his opportunism can continue. Box 7e5). Costa Rica, Colombia, and to an extent Brazil Indeed, the existence of many laws-"egal pollution"- and Venezuela are also countries that have benefited from a can be the antithesis of the rule of law. Informality seems history of relative stability that has allowed a better public to reflect the gap between expectations of what the state service to emerge (though political crisis in Colorrbia has should be doing and the reality of what it can do. recently meant some deterioration). Some of the small To characterize informality is not, of course, a complete English-speaking countries of the Caribbean also provide diagnosis of problems of public administration in LAC. an interesting foil. These countries tend to have clearer There are other important elements, such as low wages and legal frameworks (with less overlap between agencies), a skills and a corrupted wage structure, that, though no more professionalized public service, more centralized and doubt linked with informality, have not been mentioned better enforced personnel and onancial controls, and less above. Moreover, our characterization of a dominant mode opportunistic behavior than in other LAC countries. Of informal behavior has been stylized and generalized. In lndeed, the control system works to the point that effective reality, informality is a worldwide phenomenon of public centralization and the inflexibility associated with this is- administration in developing countries. It is not absent in somewhat as in the more advanced countries-ome of the 130 INSTITUTIONAL REFORM IN HIERARCHIES principal problems that reformers have to tackle in the tics to flourish. Geddes (1994, pp. 86-89) describes how the English-speaking Caribbean and in Chile. political machine works in Brazil: Politicians win votes, in part, by using cabos eleitorais ("electoral corporals") as inter- The Origins of Informality mediaries who provide private favors, such as access to state Informality, in this sense of a weak rule of law, is a funda- benefits, from public resources. This system requires exten- mental characteristic of many countries of the region, per- sive patronage, in the form of public jobs, to work. vading the public and private sectors. It is not simply a Democracy in the LAC region is typically younger, and technical problem that can be corrected by better rules. has been more punctuated by non-democratic interludes, Some point to its origins in colonial history (see de Soto than in the OECD countries. For instance, universal male 1989 and Hopkins 1991), although this may apply more suffrage was achieved by 1921 in most of Western Europe to the Spanish-speaking countries than to Brazil or the and North America but not until the 1970s in a sample of English-speaking countries. Others point to the low level Latin American countries (Engerman, Haber, and Sokoloff of generalized trust, or social capital, that sometimes char- 1998). acterizes these societies and, correspondingly, the impor- As is expected under these conditions of a weak voter- tance of private networks in transacting in the modern sec- principal, interest groups dominate the political scene. To tor. But when governments do not want to, or cannot, obey the extent that interests were fragmented and none of these the rules they themselves have set, surely informality must had the expectation of commanding power over a longer also be seen as political. A recent study by ECLAC (United period of time, it might furthermore be expected that pub- Nations Economic Commission for Latin America and the lic goods would be underproduced (and public administra- Caribbean 1998) argues that non-transparent-in our lan- tion weakened), because these interests would have the guage, informal-fiscal arrangements typify many coun- incentives to act as "roving bandits" rather than "station- tries of the region. These arrangements, the study argues, ary bandits." are ways of getting around the lack of political consensus Despite the indications that this is the case in LAC, in society on the role of the state, including its size and there is a strong tide of democratization in the region, and how it is financed. as voters are becoming more effective, the pressures to for- The political problem can be addressed in institutional malize the public sector are growing (see below). terms. Earlier, in the problem ofpublic administration: an insti- tutional perspective, we suggested that the voter-principal The Regional Reform Experience would have difficulty controlling the politician-agent if There have been various approaches to public-sector reform voting rules led to fragmentation within or between polit- in the region. Although we still need to learn far more ical parties; if information asymmetry helped preserve about how successful these initiatives have been, we never- clientelist patterns of political activity; if party rules cre- theless can make a few generalizations, or at least put for- ated closed, oligopolistically controlled parties (rather than ward some hypotheses. Reforms affecting the public sector open, mass-based parties); and to the extent that democracies can be classified in two dimensions. First, these reforms are were "young." While it is difficult to collect conclusive either discrete (or "one-off')--a downsizing, for instance- evidence on the matter, it must be a strong hypothesis that or they represent permanent changes in process-for many countries of the LAC region meet these criteria. instance, the rules and procedures that control the size of Comparing six countries under democratic rule in post- the civil-service establishment. Second, reforms have been war Latin America, Geddes (1994, Chapter 5) shows that implemented either through "enclaves"-typically, spe- civil service reform was more likely to be legislated and cial-purpose bodies insulated from the rest of the public extended when power was more evenly and stably divided administration-or across the board, i.e., in the public among a small number of parties. administration as a whole. Typically lower educational levels, poorer communica- tions, and less developed media in the LAC region than in The Successes: "One-Off' and Enclave Reforms more advanced countries may lead to the kind of informa- Many countries of the region have, since the 1980s, suc- tion asymmetry that allows clientelist-or machine-poli- ceeded in many of the areas of "first-generation" reforms, 131 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Cinc lodng macroconomicstabiiainonn,fc# downsiic ng h hncntucigadipeetigte -aebe 'iespublic administration, deeuaisone,o prvthezbetion, an achiesedb smallroupsin f measple,suamey tear noeratd decnbrliczeatints a heproessavealiozedcmeentsm ancesfu h and usunAeally out prmaentcvlsrvns)eviedlvr refot-snd tsoperatiiones afrepredictbe de hivery, oasey SocialsuThese oftagen have nbeenrb tchieve pblicradringistlav- Inhattenetiv punds. sitiina~ o raedoro thasit outesidred the ennisterial nistructre cendotur answe Tpureacty ofpirt-gnealtgood reformsit -"yone-aour'uio reorm coiressl 'ton toe presid rent roltatte, annreasn pro ee fof experienoled an dismant.linge operlicgnpervniods rMather putsidae irnotorss moreoer,tn they habvi pronitectedfniong I 97 ad t h'w ~log,sez~enil evlomet-A te ane ~ie,ai'ewy'eterip idlec132an INSTITUTIONAL REFORM IN HIERARCHIES BOX I73 IcOedi growth -f the public sector. But governments*were ble to acterist f m d loig countries, makes the blic ke some modest enting gins throug theedepedentn the honesty and dedi- impr ovement of the s;kills-be of pulic servants, ratio- cation of its civil servan T reforme ste was nalization of budget procedures, and improvement in:the lrgely efetive, but in macroecono control than public- adintstrati,on-'s'capity or s nosis More- inalloation of resources 'ry f seces, be over, by the 19s policy-makers r ein the -t reliance o t cont-'ls. autonomous gecies ia am-ore favoble light Afte lit 90, ihe ne The mitliry government th t seiz power in 1973 demoraticgov soght to m hack the frontiersof the state 7but as: sericesi to the private secto, irov th erformance of as anything else iitchanged the way: the state did busi- reglatory agencies and resctate some social services. It ness-from owner. andserviceand also came to- reize t h fo a m from centralized to decentralized service pider To do etred St p ad he riidi- this, it benefited i frpm A professionaliedciv serv t rs o hie ia de o- the mpha wich came under-much pressure fr the hah fiscal sis in tab adjustments-of them d-1970s and mostoft 198s)But objct hacteristic of the anal-yica ut pragmatic to sustain the!reformns,' te government also hd to da approach toieform typfied by previour the wt the severepoe iinherited of a deterioratdgoverment nab refo i a grdualis, stp- Pubi dm istr to.t card out a number reforms Wise maner. ticipaive form strategc panng :to'entr itefia ci ontt1rol within ,te executive and was intd-i 1 II reduce the influence'of interest,groups; uniifyciviil-servat ies were pay ses and strengthen metitocra tic'm le an ,-sipif I introduced; these wrew substantially ful il di Are n the orgnzational sttuture and decentrialize ate . -Inthe same year.experiments in perr deve them to tprivate secret (In 'I 98 the gvern mance- p tr ced in some agencies and mente formaizedito thedis ction between pI wil cver agI 199 making ministries and autono-mous agncies pvn in ed, c ,services, -thu creatinga system akintoSweden's and to and 291 indicators by, 199 6-hAi 'the eo , some of the i'novative elemuen o pu.blic adiministration started producing annual performance reports and a system intrducd in the Unite kingdom and Nw Zealand Di f ev ating c pws ntrodued These tht 1980s These reforms ledto -a system charcterized b increases in the comibintion of a centaied, rigid coitror (i.e. inputs) with: decetra lied.implenetation-b udtge execution personnel ranagnent,a and . ( Marc 997 (Chile's Idcentralized impl tatioin, which is-nor char- (19%); V s (19893:> and are subject to different control rules, which usually approach to promote administrative reform in the 1930s allow greater latitude in salary policies and greater man- and President Kubitschek's success in using agencies to agement flexibility in using resources and rewarding promote his industrial program in the 1950s.15 performance. Autonomous agencies also played an important role in Enclaves have played important developmental roles in Chile's development since the late 1930s (Marcel 1997). a number of countries. Geddes (1994, Chapter 3) cites Peru has had substantial success in reforming its tax and Brazil, and President Vargas's success in using such an customs administrations and in regulating competition, 133 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER consumer protection, and utilities through autonomous Venezuela). Chile provides the region's best example of a agencies (Keefer 1995). But the enclave approach is prob- country that has successfully reformed its core public lematic, for two main reasons (see Manning 1998). administration across the board (see Box 7.5). First, enclaves undermine efforts to reform the core pub- There have also been some attempts to apply newer lic administration because they undermine the rules of the models. These, too, have had a mixed success, or else the core. This is to say that enclaves are a manifestation of outcome is not yet clear (see Chapter X of United Nations informality. The donor community bears some of the Economic Commission for Latin America and the responsibility; the project units promoted by foreign aid- Caribbean 1998). On the plus side, since the early 1990s, donors and the habit of bringing in consultants paid by Chile's public administration has adopted a progressively donors (as in Bolivia) to work alongside higher-level civil broader array of New Public Management instruments, servants constitute an extreme version of this problem. including strategic planning, modernization agreements Second, enclaves do not provide a permanent solution. with individual agencies, performance pay, performance In general, they have proved difficult to sustain (or at least indicators, and evaluation of public programs. And the to sustain as productive agencies) once their champion- Brazilian State of Ceara, over the period of a few years after usually a president-has left office. Jamaica's autonomous 1987, achieved dramatic improvements in the quality and agencies have in the past, for instance, created a substantial coverage of its services, much of this through a transfor- fiscal challenge (Manning 1998). The successful efforts of mation of its civil service (Tendler 1997 and World Bank one mayor of the city of La Paz, Bolivia, to combat corrup- 1997a, Box 5.7). Through civil-service rewards and careful tion in the early 1990s stalled after he left office (Klitgaard recruitment, flexible organization, and community moni- and Baser 1997). toring-a mixture of the New Public Management and Autonomous agencies are not bad per se. In fact, the "voice" reforms-the Ceara government created a sense of autonomous agency and the executive-agency arrange- mission and participation among its workers. But there ments favored by proponents of the New Public Manage- was a failure in Ecuador, where an attempt to introduce, ment have much in common. The problem arises when the sequentially, a contractual approach to administrative autonomous agency operates under a different institutional reform in all the ministries quickly foundered. Two umbrella. More than other countries of the LAC region, months after finalizing the rules, the government declared Chile has been able to bring its enclaves within the general all central administration entities "restructured," in order structure of government. to grant a salary increase, although none had met the eli- gibility conditions (Reid 1998). Reforms of the Core Public Administration: A A World Bank project is helping Jamaica transform 11 Mixed Picture existing entities into executive agencies on the U.K. We do not know enough about more ambitious attempts model, but it is too early to gauge the results of this exper- to reform public administration across the board. There iment. Brazil also has a system of performance contracts have been many of these in recent decades. Some have between the central government and agencies. In addition, shown disappointing results, but the jury is still out on Colombia is building a National Evaluation System for current efforts. Reform attempts have taken a variety of Public Sector Performance (see World Bank 1997b), and forms, including strategic planning, better civil service Costa Rica has also introduced a system of performance systems and training, better budget and financial manage- measurement and related incentives. ment systems, performance-oriented management tech- There also have been a number of successes in applying niques, greater use of "voice," and so on. Large amounts of newer methods to the reform of individual agencies. The resources have been spent on designing and installing inte- Brazilian Treasury has introduced a system of salary rewards grated financial-management systems in most of the larger based partly on team performance. Results-oriented bud- countries of the region. But with a few exceptions, such as geting has been introduced with some success in the Brazil, these systems have not yet shown that they can be province of Mendoza, Argentina. Several autonomous agen- effective in rationalizing public financial management (see cies have benefited from increased managerial discretion Reid 1998 on overambitious efforts in Bolivia and and contract-like arrangements (see Keefer 1995 for Peru). 134 INSTITUTIONAL REFORM IN HIERARCHIES In a nutshell, many LAC countries have done well in parency will also have only a marginal contribution under "one-off" and enclave-based reforms and less well on these circumstances. Fred Riggs made this point more than process reforms and across-the-board reforms. The suc- 30 years ago with the hypothetical example of a would-be cesses, achieved by circumventing the public administra- reformer "called upon to correct the evils of a chaotic filing tion, have come about because they changed the principal- system in a formalistic bureau. His technology takes for agent problem. Typically, the president puts the operation granted the existence of an effective demand for good writ- in his direct backyard, bypassing the cabinet and the min- ten communications. The most modern and scientific pro- istries-or at most only using a few trusted people in the cedures and equipment will not remedy the situation if ministries. But enclaves are a problematic instrument and such a demand does not exist" (Riggs 1964, pp. 17-18). typically lose their effectiveness, or die, when political regimes change. By contrast, many across-the-board Reforms that Address Informality reforms have not been good at addressing the principal- A precondition to any modern system of public adminis- agent problem. Chile's careful (though not always linear) tration based on rules and checks and balances is a change advance-its effectiveness in achieving core public-admin- in the political conditions that encourage informality. The istration reforms, its ability to sustain autonomous agen- LAC region is clearly moving in the direction of change, as cies, and its current experiments in mitigating the rigidi- a result of two powerful pressures-democratization and ties of hierarchy-probably comes nearest to a model of globalization (see Chapter 2). reform for the region. The democratization of the LAC region since the late 1970s-the replacement of authoritarian governments, Reform: Constraints and Possibilities constitutional change, democratization of subnational gov- ernment-has been remarkable and consistent. The Informality and Reform changes have been as deep as in any other region of the The core difficulty in getting good public administration world (see Figure 7.1 of World Bank 1997a) and paralleled in democratic countries can be usefully expressed in prin- only by the more abrupt democratization of Eastern Europe cipal-agent terms: how to make the chain of accountability and Central Asia. Globalization, meanwhile, has helped work-from voters, through their political representatives, bring about the opening of the LAC region's economies to down to public bureaucrats. To simplify, we could say that the forces of international competition. These forces are the advanced countries have answered this with two mod- putting pressure on the very interests that have benefited els: the older hierarchical model, which combines checks from informality in the past to improve public services- and balances (deriving from the separation of powers) and education, communications, dispute resolution, market command-and-control structures, and the recent New regulation and so on-so that their economies can com- Public Management model which combines checks and pete. Thus, we can expect that the space for public admin- balances and market-friendly performance-oriented istration reform will continue to grow. arrangements. The second is an offshoot of the first (and by The political demand for formalization can be supple- no means eschews hierarchy). Both are based on the appli- mented by public actions such as these: cation of universal rules, including the same checks-and- * Electoral reforms might help improve the chain of balances arrangements. And there might be more conver- accountability: Some political-science research sug- gence, in modern practice, than the champions of either gests that reforms which lead to fewer, stronger par- school care to admit. ties could strengthen the constituency for reform of Informality lies deep within the fabric of society, per- the public administration. vading both the public and private sectors, and reflects * "Voice" mechanisms-client surveys and scorecards, political structures. It follows that reforming public participatory practices-circumvent the poorly func- administration in the countries of LAC is not merely, not tioning chain of accountability. Reforms that increase even principally, a technical issue. Changing the formal the information available to the public, provide pub- rules will not help if incentives do not change. Providing lic undertakings on service standards, create mecha- hardware or systems to improve information or trans- nisms for the government to "listen" to clients, or 1 35 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER technically strengthen other branches of government, How should the countries of the region move in this such as the legislature or the office of the auditor gen- direction? How comprehensive should the reforms be? eral, can all contribute. What is the best order of reform? The answers depend, of * Policies that move the "informal sector" nearer to the course, on specific conditions in each country-on politics, modern sector-formalization of property rights, leadership, windows of opportunity, and the current state deregulation of business practices, tax reforms-may of the public administration. But there is also a lively also have the result of better incorporating the generic debate on how far and how fast the reform process "excluded" into the political system. should go in developing countries. Caution is a common * Economic reforms that "level the playing field," open theme. The World Development Report 1997 (World Bank up modern-sector activities to new entrants, and 1997a and Box 7.1) advocates many aspects of the new weaken oligopolies, may create a demand from the thinking, but advises caution in some areas. Reid (1998) private sector for better services from the public sec- discusses the steps necessary to make New Public Manage- tor. Schick (1998, p. 127) argues that "{nlorms, prac- ment reforms work in Latin America, and also advises cau- tices, and ideas migrate from one sector to the other," tion. Schick's (1998) advice is that "most developing coun- and that "[t]he emergence of open, robust markets is tries should not try New Zealand's reforms." Bale and Dale as much a precondition for modernizing the public (1998), on the other hand, maintain that developing coun- sector as it is for developing the private economy." tries can, if careful, learn from New Zealand's success(l If this is an eclectic list of reform possibilities, it points For countries where informality still rules in the public to the need for continuing to improve our understanding sector, the issue of sequencing is difficult to avoid. Formal of the interaction of political arrangements (who holds the rules must prevail before the system can be improved. Sev- power in countries and under what conditions), economic eral sequencing variants have been suggested. Schick (1998, arrangements (the economic rules and their impact on pp. 129-131) sketches out a possible sequential path. He who the economic players are), and bureaucratic arrange- sees private-sector economic reform as a necessary initial ments (principal-agent relationships in the governance condition. This should be followed by the establishment of process). reliable external control (centralized, top-down financial controls), realistic budgeting, and some investment in a Reform Options professional civil service. These basics will allow politicians The New Public Management provides a powerful set of to take control of public management on the basis of effec- ideas about the uses of hierarchies, markets, and voice that tive control of inputs: "They must be able to control inputs have reset the agenda for public sector reform. Many before they are called upon to control outputs." Once this OECD countries, and a number of developing countries, stage has been mastered, the system could move to internal have begun to move in the new direction. Devolution of controls, where agencies police themselves and are subject activities-to the private sector, to corporatized bodies, or to ex post audit from the center. This would give the agen- to lower levels of government-is an important common cies substantial managerial flexibility. After independence element. But as for the reform of the core of the public in 1963, Singapore rapidly followed this route from inter- administration, there is as yet no consensus on the new nal to external controls, and it was then able, by the mid- model-in particular how far to go in the direction of 1990s, to adopt a "budgeting-for-results" system that "marketizing" functions and systems. Nonetheless, there implements several elements of the New Zealand model. are some key ideas that provide something of an idealized Some have criticized sequentialism as a recipe for inac- menu for reform in the countries of the LAC region. The tion. A variant is to propose that different agencies "grad- ideas relate to the setting of strategic priorities; reorganiz- uate" at different times to a more performance-oriented set ing the way that public services are delivered; moving to a of rules, and that the timing depend on the ability of these performance-based (or results-based) system of account- agencies to shift from external to internal control. Reid ability; and backing up this system through financial-man- (1998) describes such an approach and how it has been agement and personnel reforms. These ideas are set out in used in the LAC region. It is being applied in Brazil, more detail in Box 7.6. Ecuador, and Jamaica. 136 INSTITUTIONAL REFORM IN HIERARCHIES AU Set Re,m ubeues Emerging from Nelw NN Management. A recent. World Ba nk re: o 'in results-orientd publi- m m ectve sector reform -il aco s ered s ckage of e e ent inclu leadershi .-,or ,.,efo-, , tid. M ;eon r, 0t- IngpRi.. - reforming publc man agmett systems, structures and resources, tinentives, freedlom man:ae, and iacentives. tiS e-merges a set of issues thata vantito thereirmofp sblic adinistrtions oftheLA Effive financial n Finncal ae- -region (Ann 3SofI W rld 'Bak 9' wnt-' se sudi * -lear strategic priorities. Governments must decAid on ig, executo, andaUdi f t . A prfor- strategic priori ies an use b t. llocations ie- at ti quire h st tives, and montoring to achieve these, a reources on an ouput basis, faly account - Optim.l ways of dvii ering public services. Govemments for, al resources u (kCin-cludig t- must decidie whetherpublic goods services are idectralizd exectito th blalances best. delive4rd through regulation, regulated pivate manageriall freed and accountabi for- prouction, or public po.tion. res u r ce ue.: -~~~~~~~~ q-0tionl. -Xlr es urc * Opimalwaysf oranizng pth/i agecies Incntivs * erfomance-infornauion: systems: Perforance-based: -maybeclari-fiedand informationr may flow better by systems rqt uire infiot that ias nt t rearranging the boxes." (For = instance,an agncy that the financial mnanagement. sjrstem.n .videbiased advice,) P.erfornce m-ay be enhanced bequire a changel in midser from a Ca creag c igovernm ture to e cultur *- S c b r o i Performance, can be * E.t persnne sn Dation of perso-D enhanced if obf etilves5are set an hei is n i m are to beheld to act ieve ihem.i accountable fo resul ts Those countries in the region where some success, Notes however limited, has come from the creation of 1. This characterization is meant to encompass what a political autonomous agencies may face an additional option-to scientist might expect from good government in terms of decisive- improve the sustainabilit ofautonomousagencness, capability, and responsiveness (see Cox and McCubbins 1996) imcorporove ngthemsustainabioiry offeautonomoi agenes by and what an economist might expect in terms of macroeconomic sta- bility, strategic prioritization, and efficient and responsive service the state, i.e., by "formalizing" them. This might be delivery (see Campos and Pradhan 1996). achieved, for example, by transforming autonomous 2. The principal-agent literature addresses how a principal agencies into executive agencies by creating a uniform set designs a contract that will provide the best incentives for (i.e. best of accountability rules and establishing client-service likelihood of) an agent's carrying out the principal's wishes. For standar. Pt oapplications of principal-agent literature to the functioning of the public sector, see Blau (1974); Morgan (1986); Scott and Gorringe tion reforms has been through incorporating its (1989); and Lane and Nyen (1995). autonomous agencies. 3. See also Annex 3 of World Bank (1997b) for a similar presen- The new reform ideas provide a challenge to many tation. countries of the LAC region because of the informality of 4. To present only two stages is, of course, to simplify. Not only their current systems of public administration. But under are there many stages (voter, politician, party, executive/president, the thrust of democratizing and globalizing forces, these minister, agency head, civil servant ... ), but the process is not linear, nor is it unidirectional: Executives, legislatures, and judiciaries all ideas also provide many opportunities for the region to have degrees of control over each other and can have competing con- move forward. trol over agencies. 137 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER 5. Wade (1997) seeks to explain the superior performance in the (1998) applies the idea of informality to the public sectors of devel- management of Korean irrigation systems compared with Indian sys- oping countries, arguing that there is a parallel incidence of infor- rems by showing how the organizations in each country used differ- mality in the public and private sectors. Klitgaard (1998, p. 336) ing arrangements to tackle the principal-agent problem. characterizes "sick institutions" by "weak information, ineffective 6. Sometimes the executive is subordinate to the legislature; at incentives, and chronic shirking and malfeasance." other times, notably in the United States, it is independent. 13. The United Nations Economic Commission for Latin Amer- 7. Others call it "the new managerialism" or "performance-based ica and the Caribbean (1998) characterizes these kinds of budget management." problems as indicating the weakness of the "fiscal covenant" between 8. The four topics follow a schema suggested by David Shand. government and citizens. 9. For a treatment of these reforms, see Scott and Gorringe 14. However, decentralization, which has been especially strong (1989), McCulloch and Ball (1992), Schick (1996 and 1998), and in health and education services, has a mixed record in terms of its Bale and Dale (1998). efficiency effects and has often created fiscal disequilibria. 10. Paul (1991), Salmen (1992), and World Bank (1992) have 15. See also Evans (1992) on other "pockets of efficiency" in adapted the concepts of "exit" and "voice" from Hirschman (1970). Brazil. In broad terms, they use "exit" as the ability of the customer/citizen 16. There are several reasons for particular caution aboult apply- to choose alternatives (i.e., within a market) and "voice" as the abil- ing a New-Zealand-style model emphasizing contracting. Contract- ity of the customer/citizen (through the government's "listening") to ing is technically demanding and risky: Public-sector outputs are exercise influence over the type and quality of service. typically difficult to measure, as is the performance of inclividuals; 11. These averages, of course, mask a substantial variation among and since intra-government contracts are not at arm's length, they the countries of the region. could be difficult to enforce. For these reasons, some fear that greater 12. De Soto (1989) played a leading role in characterizing infor- managerial discretion could open the door to greater opportunism. A mality as it affected the private sector and the populace. Informality study of state-owned enterprises in developing countries (World in the public sector has nor been newly discovered, though the spe- Bank 1995b) showed that the use of performance contracts did not cific terminology may not have been used much. Riggs (1964) do much to improve enterprise performance because they did not expounds a concept of "formalism," a term frequently used to char- reduce managers' information advantage, they rarely inclucled effec- acrerize Latin America's legal systems, more or less the same as infor- tive rewards or penalties, and governments showed little commit- mality: laws that are not put into practice. Following de Soto, Schick ment to contract terms. 138 TECHNICAL APPENDIX: Concepts for Analyzing and Designing Institutions Economists have developed some useful concepts and tions whenever relevant information about a particular insights to help analyze and design appropriate institu- transaction is not fixed. If the relevant information tends to tions. In this appendix we review concepts associated with change over time as conditions change, then there will six sets of issues: First, we cover issues related to asymmet- tend to be imperfect information. ric information and resulting principal-agent problems. Principal-agent problems arise whenever an interested Second, we describe exit and voice strategies as feedback party (the principal) delegates the responsibility to act on devices. Third, we discuss the relationship between trans- his or her behalf to an agent, but, due to information asym- actions, property rights, contracts, and enforcement mech- metries, the agent can relinquish such responsibility with- anisms. Fourth, we evaluate the implications of the exis- out the agent's knowledge. This type of problem affects tence of transaction costs for the design of institutions for numerous relationships: Politicians sometimes do not ade- markets and hierarchies. Fifth, we highlight the impor- quately represent the interests of voters; bureaucrats or reg- tance of institutions for the provision of public or collective ulators do not always act in accordance with the mandates goods. Sixth, we briefly discuss the importance of the pre- of politicians or voters; teachers may not act in the interest dictability and credibility of institutions for economic and of students and parents; and judges may not always defend social development. the interests of parties engaged in a legal dispute or the "public interest" (T. Moe 1984, Pratt and Zeckhauser Asymmetric Information and Principal-Agent 1985). The negative welfare consequences of principal- Problems agent problems can be prevented or reduced through insti- Two closely related concepts are those of asymmetric or tutions that establish effective monitoring or feedback imperfect information and principal-agent problems. The mechanisms and that make performance and outcomes basic information problem described the financial and edu- more transparent and measurable. Ideally, monitoring cation sectors is that agents (school teachers or directors mechanisms should also be accompanied by enforcement and firm managers) possess much more information about mechanisms that effectively punish opportunistic behavior what they do (teach or use other people's money) than those and reward good performance. who take their children to school or own shares from a firm, These problems can lead to adverse selection, moral lend to it or deposit money in a bank. hazards, and incomplete markets. These economic phe- Asymmetric and imperfect information are facts of nomena were described informally in the previous section. life. Asymmetric information exists whenever one or more In our example of the problems faced by bankers and bor- parties in a transaction have information about the quality rowers when they do not have sufficient information of their inputs, outputs, or other aspects of the economic regarding the borrowers' creditworthiness, we emphasized undertaking that is not readily available to the others (or that in the context of imperfect information only risky bor- that is too costly for the others to assess). Likewise, imper- rowers may be willing to take on loans at the high rates of fect information afflicts all economic or business transac- interest charged by the banks. Hence this type of situation 139 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER has been called adverse selection, reflecting the tendency moting the use of rating agencies enable shareholders to for the market to be dominated by exactly the riskiest cus- effectively "voice" their informed opinions about manage- tomers. This type of problem has been extensively analyzed ment decisions, or "exit" the firm, which in turn can pro- in the case of insurance markets, where only the worst risks vide incentives for managers to behave in a way that pro- tend to purchase private insurance (Rothschild and Stiglitz motes minority shareholders interests and to provicde them 1976). with better and fuller information. Organizations like Likewise, due to asymmetric information, firms may securities exchange commissions are part of the institu- tend to undertake too risky operations, expecting to bene- tional set to regulate and supervise the implementation of fit fully from good results while shifting, at least partially, such requirements. potential losses to their lenders. Banks can engage in the same behavior, expecting to shift eventual losses, at least Exit and Voice as Feedback Devices partially, to depositors. These are the so-called moral haz- In general terms, shareholders or households can provide ards. In the context of insufficient information about the incentives for improving the management of firms and health of banks, governments often provide deposit insur- schools respectively by either exit or voice strategies.2 In ance to protect the banking system from potential panics the former, shareholders or households use the market to by depositors who may rush to withdraw their deposits in defend their interests if they are dissatisfied with their response to events that may or may not have affected a spe- agents' decisions; they simply divest from the firm in ques- cific bank. However, the government-provided insurance tion, take their deposits to a different bank, or, t]heoreti- may reduce the depositor's incentive to avoid risky banks cally, withdraw their children from the school. In contrast, or to monitor the entity's financial health, increasing the the voice strategy implies the use of control devices, moral hazard problem. Thus, disclosure requirements, pru- whereby the shareholders or parents in our examples can dential regulation, and supervision are essential institu- induce changes in management of the firm or school. tions to cope with such problems. (These issues are dis- With this distinction in mind, it is worthwhile to cussed in detail in Chapter 3.) Moral hazards can also explore some of the challenges associated with education, emerge in the context of hierarchies. For example, tenured where exit is often not an available strategy. For example, employment contracts within organizations (including in rural areas of developed and developing countries, it is public bureaucracies) tend to create an incentive to work difficult to find more than one school within a reasonable less since employment is guaranteed.1 distance from household clusters. In addition, due to famil- Also, as illustrated earlier, some markets in the context ial and cultural ties to certain regions, households often are of information asymmetries may never develop. In general tied to their "roots," which again limits the exit strategy. terms, whenever private markets fail to provide goods or Thus, in addition to the aforementioned problems of insuf- services at costs below the price that individuals are will- ficient information available for parents to make decisions ing to pay, economists refer to this phenomenon as incom- about where to send their children to school, these consid- plete markets (Stiglitz 1986). erations about the lack of household mobility seem to limit One of the critical roles of institutions, then, is to the scope for a market-driven education system, such as reduce the information asymmetry and overcome princi- public education vouchers.3 In such situations, "voice" pal-agent problems. For example, teachers can be made mechanisms, through parent or community involvement more responsive to the concerns and interests of parents by in the decision-making in schools-as in the EDUCO pro- establishing performance contracts that pay teachers gram in El Salvador-are useful complements to "exit" according to their performance (and even attendance) in mechanisms. the classroom; the government may conduct standardized academic tests across public and private schools in order to Transactions, Property Rights, Contracts, and provide parents with information regarding the academic Enforcement progress of the children in each school relative to others, Transactions in an economy (through markets or hierar- etc. In the case of capital markets, providing minority chies) consist of transfers of property rights on assets, shareholders rights, establishing disclosure rules, and pro- goods, or services. Such transactions permit the exploita- 140 TECHNICAL APPENDIX tion of the potential gains from trade (internal and exter- vatization, a mining concession, or an acquisition or a nal) and specialization (both within markets and hierar- merger among private firms. In these cases, the informa- chies), which are essential to economic development. It is tion and enforcement requirements can be enormous. precisely one of the basic roles of institutions to specify In general, the magnitude of transaction costs depends who owns what (property rights) and how property can be partially on technology. For instance, the information rev- combined or exchanged through transactions. olution has reduced the costs of transferring and evaluating Transactions are regulated by formal or informal con- information. However, while the information superhigh- tracts that determine the conditions under which property way makes it easier to access information, it has not neces- rights are transferred, including prices, conditions for pay- sarily provided incentives for entities to provide informa- ment, the venue for finalizing the transaction, and so on. tion. Thus, the magnitude of transaction costs also Contracts are institutions-that is, rules that establish the critically depends on the quality of institutions. conditions that the parties must satisfy to reap the benefits Indeed, their magnitude depends on the transparency of the exchange in question. Contracts, in turn, are regulated and efficacy of the rules that establish and protect property by higher institutions-both formal rules (the legal system, rights and the characteristics of contracts, including the regulations, bureaucratic procedures) and informal rules efficiency and reliability of the judiciary and alternative (trust and acceptable business practices)-and are enforce- dispute-resolution systems, and on the extent to which able through courts or other arbitration mechanisms. they reduce the costs of information gathering, by estab- Although informal rules for resolving property and con- lishing clear incentives for actors to provide and seek infor- tract disputes are used throughout the world, especially in mation relevant to their particular undertakings-or to developing countries with unreliable enforcement of for- those of others, in the case of credit bureaus, rating agen- mal rules, it can be argued that in today's world of imper- cies, entities that rank the quality of universities, or any sonal transactions and non-barter economies, it is increas- organization or individual that makes a living out of pro- ingly important to develop the formal rules establishing viding information. In addition, the rules that establish property rights and contractual obligations. Some analysts markets and hierarchies affect the incentives to participate have argued that developing countries should focus on the in market activities (including risk-taking) and for mem- development of substantive and procedural rules of prop- bers of an organization (both private and public) to act col- erty and contract as much as on creating a first-class judi- lectively in pursuit of common objectives. ciary (Posner 1998). In the case of financial and capital markets, the adoption of sound internationally accepted accounting standards, Transaction Costs external audit requirements, disclosure rules for banks and It must be clear by now that transactions involve signifi- corporations, and consolidated balance sheets for financial cant costs. Transaction costs include costs of obtaining and groups can reduce the scope of asymmetric information that verifying information about the quantity and quality of afflicts these markets. Also, promoting the operation of credit goods and services, the partners in a transaction (and veri- bureaus and rating agencies for the same markets can help to fying their reputation, records, etc.), and the quality of improve the availability of relevant information that can then their property rights to be transferred, including the legal be used by depositors or shareholders to make informed deci- and contractual framework; as well as costs of designing, mon- sions about where to put their money. In the commercial area, itoring, and enforcing the contract of transfer, including any there are also important quality standards for trade, especially costs incurred in litigation and dispute resolution. Some of concerning international trade, which ensure that traders are the risks involved in a transaction are insurable; that is, getting what they paid for. In the field of education, we have insutrance costs are also part of transaction costs.4 already mentioned the potentially constructive role of stan- For some transactions these costs may be small, as dardized exams for teachers and students, and the relevance of reflected in our previous discussion of shopping for gro- certification requirements to ensure a minimum level of qual- ceries in supermarkets or spot-market transactions. In oth- ity for schools and universities. ers, transaction costs can be huge for all parties involved. Unfortunately, there are also numerous examples of Consider, for example, the bidding or negotiations of a pri- "excessive" regulations and information pollution that 141 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER increase transaction costs. For instance, licensing require- costs by eliminating the capacity of organizations and indi- ments for establishing businesses, with heavy and costly viduals to act in an opportunistic fashion. This role is par- information requirements, are often imposed as revenue- ticularly important for the production of public goods. raising or protectionist measures (they can unduly increase Broadly defined, public goods have two important char- entry costs) rather than for solving real information or acteristics. First, they are non-excludable: Once they are enforcement problems. Also, very often reporting require- provided to one member of a group, they cannot be denied ments within the public sector take enormous efforts to to the others. Second, public goods are characterized by a produce massive ill-designed data that nobody can use or non-rival consumption feature; that is, the consumption cares to use in an efficient way. In addition, there are insti- of a public good or service by one individual or organiza- tutions that induce excessive "risk-taking" by the private tion does not reduce the supply of such good or service for sector, such as the case of blanket universal deposit insur- other potential beneficiaries.6 For example, the benefits of ance schemes in the banking sector. police-greater personal security-cannot be denied to A useful way to assess the quantitative importance of an any member of the community once it is provided to oth- economy's transaction costs (and thus institutions) is by dis- ers. Similarly, economic and institutional reforms can be aggregating total costs of production into transformation costs analyzed as public goods, because the benefits from reforms (determined by technology, factor, and input prices) and will be absorbed by society as a whole. transaction costs (determined by institutions and technology). A problem arises in ensuring that all beneficiaries of a Wallis and North (1986), for example, estimated that trans- public good will pay for its production. Since consumers action costs exceeded 40 percent of total production costs cannot be excluded from its benefits, they have an incen- (GNP) in the United States in 1970. Although it should be tive not to pay for providing it. This is the so-called free- acknowledged that in practice there are difficulties in mak- rider problem, where beneficiaries have an incentive not to ing the distinction between transformation and transaction pay because benefits are not linked to contributions. Olson costs, such an estimate underscores the critical importance (1965, p. 14) equates public goods with collective goods, of institutions for efficiency in an economy.5 implying that organizations or groups of individuals with According to the transaction-cost literature, hierarchies common interests may face free-rider problems. Often are instituted to reduce the transaction costs that would be organizations impose membership requirements in the involved if the relevant transactions had been conducted form of mandatory contributions, such as dues imposed by through the market (O. Williamson 1989). Nonetheless, labor unions around the world (frequently supported in transaction costs within hierarchies can still be significant, law; i.e. coerced by the state). Another potential solution especially in large organizations. For example, many gov- for free riders is to establish monitoring mechanisms to ernments spend a significant amount of resources auditing ensure that members are not shirking on their collective and controlling the expenditures of public agencies. As commitments. mentioned, public organizations are often used as political instruments, which may lead to opportunistic behavior by Institutions, Predictability, and Credibility politicians. Of course, this type of opportunistic behavior In previous sections we argued that institutions maLtter for can also be ameliorated by designing institutions that long-term economic development because they establish effectively monitor performance, and perhaps that link the incentive structure (constraints) for economic and remuneration or the length of employment to such perfor- political actors. Consequently they determine the incen- mance. Indeed, as already mentioned, many countries tives that actors face to undertake business transactions, to around the globe are currently experimenting with various invest in productive activities and human development, to forms of performance contracts. innovate, and to take economic/financial risks. Institutions determine the perceived profitability of undertaking eco- Collective Action and Public Goods nomic and political activities by affecting the risks, costs, A key role of institutions relates to incentives for people to and benefits of economic transactions, by defining who shirk on their contractual or collective-action commit- owns what (i.e., property rights) and by establishing the ments. In other words, institutions may reduce transaction conditions for undertaking the transactions (i.e., con- 142 TECHNICAL APPENDIX tracts). In addition, institutions also determine the penal- ment that has arisen from this fact is a drive toward central ties and other consequences for parties that do not satisfy bank autonomy in constitutional and legal reforms in LAC the terms of an agreed transaction (i.e., enforcement mech- and around the world. The drive toward improved budgetary anisms). The main role of institutions (or institutional institutions is also influenced by such considerations. reforms), therefore, is to minimize transaction costs, thus In turn, recent analyses of the East Asian crisis of 1997 facilitating exchange and promoting investment in the have highlighted the role that inadequate institutions had in production of private as well as public goods. In summary, inducing banks and corporations to undertake excessive well-designed instituitions reduce transaction costs by ameliorating risks."1 The existence of "perverse" incentives, that induced information and enforcement problems. Thus, they make possible excessive risk-taking, resulted in catastrophic consequences the existence, efficiency, and depth of mnarkets and organizations. when their effects were magnified by recent exposure to huge In addition, the quality (and stability) of institutions may short-term capital inflows through capital account liberaliza- be important for maintaining stable, predictable, and credi- tion. It is clear by now that reforming and strengthening of ble "rules of the game" in economic and political life, which financial-sector and corporate-governance institutions is may affect the overall performance of an economy, both in absolutely critical for developing countries if they wish to the long run and in the short run. Surveys of investors and reduce the risks of costly financial and currency crisis associ- recent studies (see Chapter 1) have proved that investment ated with financial integration. Countries that have not liber- decisions and growth are affected by the predictability and alized their capital account should do it in a gradual and care- credibility of governmental policies.' Thus, institutions that ful way, while keeping pace with such institutional reforms. improve the predictability and credibility of policies are essential to reap the benefits of "sound" policies. Notes The efforts by governments of developing countries to 1. Pearce (1992) defines "moral hazard" as "the effect of certain "lock in" their structural reforms and market access types of insurance systems in causing a divergence between the pri- through external commitments that increase exit costs- vate marginal cost of some action and the marginal social cost of that thrugh amexternal p cmitmntse thatd inrease exgaizticosts- such as membership in the W'orld Trade Organization 2. See Hirschman (1970) for an insightful analysis of these strate- (WTO), Regional Trade Agreements (RTAs), and the gies in a variety of economic and political scenarios. like-is partially explained by the desire to convince 3. Education voucher schemes had been proposed a long time ago investors that the current authorities or future govern- by Friedman (1955), and have been tried across the United States ments will not display opportunistic behavior and roll back and in several developing countries. On the controversies associated the reforms. Witoutsuccrdiblit,ivewith these experiments, see Carnoy (1997) and West (1997). thke rheforms.ionsirhot such t credpibility invstos mayfrnot 4. 0. Williamson (1989, p. 142) defines transaction costs as "the comparative costs of planning, adapting, and monitoring completion trade and investment environments 8 under alternative governance structures." North (1990, p. 27) defines A similar problem arises in private participation in them as "the costs of measuring the valuable attributes of what is infrastructure and utilities. Privatization must be accom- being exchanged and the costs of protecting rights and policing and panied by substantial institutional reforms (regulatory enforcing agreements." laws and by-laws, autonomous regulatory agencies) that set 5. It should be noted, for example, that total transaction costs in an laws and by-laws, autonomous regulatory agencies) that set economy can be readily confused with the legal, accounting, and man- agement services that are part of the services sector of the economy. to the new policies. Without them, private investment 6. A succinct definition along these lines can be found in Pearce would not be forthcoming in the desired amounts or it will (1992). require excessive governmental guarantees that may impair 7. World Bank (1997a) presents evidence on the relationship the very objectives of private participation in infrastructure between credibility and predictability of policies and laws, based on Brunetti et al. (1997a and 1997b). (shifting risks to the private sector to encourage efficiency 8. Se Burki and P 1998)n and allow the state to concentrate scarce public funds in 9. See Brwi an. (1997). r ~~~~~~9. See Irwin et al. (1997). the provision of basic social services and public goods) and 10. Kydland and Prescott (1977), Calvo (1989), and Rodrik (1989) create huge fiscal contingencies.9 11. See, for example, Krugman (1998); Corsetti, Pesenti, and Credibility has also proved to be critical for the success of Roubini (1998); IMF (1998); World Bank (1998a); and Perry and stabilization policies.-" One important institutional develop- Lederman (1998). 143 DATA APPENDIX: Institutional Indicators Used in Figures 1-17 The data used were obtained from the monthly publication Corrzption: This measures the extent to which bribery is International Country Risk Guide (ICRG), published by present "within the political system." The higher the cor- Political Risk Services (PRS), a private international ruption level in a country the less the efficiency of govern- investment risk service company that employs analysts to ments and the lower the score. Forms of corruption given provide political, financial, and economic risk ratings of in the ICRG are related to bribes in areas of exchange con- countries. trols, tax assessments, police protection, loans, and licens- The "composite institutional index" used in Chapter 1 ing of exports and imports. The points of this component is made up of five components given in the ICRG: (1) range between zero and six, where low scores indicate high repudiation of contracts by government, (2) expropriation levels of corruption. risk, (3) bureaucracy quality, (4) corruption, and (5) law- and-order tradition. PRS defines the five components as Law-and-order tradition: This component is a measure of a follows: country's legal system and rule of law. The range of points given to this component is between zero and six. A high Repudiation of contracts by government: The minimum and score indicates the existence of a highly strong and impar- maximum amount of points given to this component are tial legal system and the citizens' acceptance of legal mech- zero and 10 respectively. A low score in this component anisms to settle disputes. means that the commitment by a country's government to fulfill previous or present contracts is low. The above components measure the quality of public institutions and the quality of public services. The "com- Expropriation risk: It measures the risk of "forced national- posite institutional index" was created with them for each ization-" As in the case of repudiation of contracts, a mini- country where there was data available from 1984 to 1998. mum of zero points and a maximum of 10 points are given. The index is the sum of the five components in each year. A high score in this component indicates a low risk In order to be able to add up the five components, the last of expropriation or confiscation of enterprise equity by three components were transformed to 10-point scales; government. therefore, the total number of points possible for a country is 50, where a country with a perfect quality of public Bureaucracy quality: This component measures the expertise institutions and quality of public services would have a and autonomy of the bureaucracy. A country with a com- score of 50. Annual averages of the five components were petent bureaucracy that is independent of "political pres- available for the countries in the sample from 1984 to sure" receives a high score in the ICRG. The minimum and 1996, 1997 data are from the January 1997 issue of the maximum points assigned to this component were zero and ICRG, and 1998 data are from the February 1998 issue. six from 1984 to 1997. Starting in 1998, they are zero and Table Al has the list of the countries by regions and table four respectively. A2 has the list of countries in the sub-regions of LAC. 145 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE Al Countries Included in the Sample EUROPE AND LATIN AMERICA MIDDLE EAST AND SUB-SAHARAN ASIA CENTRAL ASIA AND THE CARIBBEAN NORTH AFRICA OECD AFRICA Bangladesh Albania Argentina Algeria Australia Angola Brunei Bulgaria Bolivia Bahrain Austria Botswana China Czech Republic Brazil Egypt, Arab Rep. of Belgium Camerooss Hong Kong, China Greece Chile Iran, Islamic Rep. of Canada C&te d'lvoire India Hungary Colombia Iraq Denmark Gabon Indonesia Poland Costa Rica Israel Finland Ghana Malaysia Romania Dominican Republic Jordan France Guinea Pakistan Russian Fed. Ecuador Kuwait Germany Kenya Papua New Guinea Turkey El Salvador Lebanon Iceland Liberia Philippines Yugoslavia Guatemala Libya Ireland Malawi Singapore Guyana Morocco Italy Mali Korea, Rep. of Haiti Oman Japan Mozambique Sri Lanka Honduras Qatar Netherlands Nigeria Taiwan, China Jamaica Saudi Arabia New Zealand Senegal Thailand Mexico Syrian Arab Rep. Norway South Afiica Vietnam Nicaragua Tunisia Portugal Sudan Panama United Arab Emirates Spain Tanzania Paraguay Yemen, Rep. of Sweden Togo Peru Switzerland Uganda Trinidad and Tobago United Kingdom Zambia Uruguay United States Zimbabwe Venezuela TABLE A2 LAC Sub-regions CENTRAL AMERICA MEXICO SOUTHERN NORTHERN AND AND THE CONE CONE PANAMA CARIBBEAN Argentina Bolivia Costa Rica Dominican Republic Brazil Colombia El Salvador Guyana Chile Ecuador Guatemala Haiti Paraguay Peru Honduras Jamaica Uruguay Venezuela Nicaragua Mexico Panama Trinidad and Tobago 146 DATA APPENDIX TABLE A3 Description and Sources of Variables Used in Regression for Figure 1.1b VARIABLE DESCRIPTION PERIOD SOURCE 1. GDP growth Average annual growth of GDP per capita in constant 1984-95 WDI, World Bank local prices. This was measured as the log difference of the per capita GDP in 1984 and 1995 divided by the elapsed time period 2. Initial per capita GDP Log of initial per capita GDP in constant 1987 US$ 1984 WDI, World Bank 3. Education Average schooling years in the total population of ages 1985 Barro and Lee (1996) 15 years and over 4. (Export + Import) / GDP Ratio of the sum of exports and imports of goods and 1984 WDI, World Bank and services to GDP International Financial Statistics, IMF 5. M2/GDP Ratio of money and quasi money to GDP 1984 WDI, World Bank and International Financial Statistics, IMF 6. Gross investment Gross domestic investment as a percentage of GDP 1984 WDI, World Bank and International Financial Statistics, IMF 7. Inflation The annual percentage inflation measured in consumer 1984-95 WDI, World Bank and prices averaged over 1984 to 1995. A Laspeyres index International Financial formula is used Statistics, IMF 8. Terms of trade - average The annual percentage change in the net barrer terms of 1984-95 WDI, World Bank trade averaged over the entire period - 1984-95. Net barter terms of trade are the ratio of the export price index to the corresponding import price index (1987 - 100) 9. Terms of trade - standard deviation The standard deviation of the net barter terms of trade 1984-95 WDI, World Bank over the period 1984-95 10. ICRG 84 Composite institutional index, as described above 1984 See abote 147 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER TABLE A4 Countries Included in the Sample of Figure 1.1b EUROPE AND LATIN AMERICA MIDDLE EAST AND SUB-SAHARAN ASIA CENTRAL ASIA AND CARIBBEAN NORTH AFRICA OECD AFRICA Bangladesh Greece Argentina Algeria Australia Cameroon China Hungary Bolivia Egypt, Arab Rep. Austria Gabon India Poland Brazil Iran, Islamic Rep. of Canada Ghana Indonesia Romania Chile Israel Denmark Kenya Malaysia Turkey Colombia Kuwair Finland Malawi Pakistan Costa Rica Syrian Arab Rep. France Mali Philippines Dominican Republic Tunisia Iceland Senegal Singapore Ecuador Italy South Africa Korea, Rep. of El Salvador Japan Togo Sri Lanka Guatemala Netherlands Zambia Thailand Guyana New Zealand Honduras Norway Jamaica Spain Mexico Sweden Nicaragua Switzerland Paraguay United States Peru Trinidad and Tobago Uruguay 148 References Academy for Educational Development (AED) (1996). Angell, Alan (1996). "Improving the Quality and Equity Educational Reform Projects Supported by the W"orld Bank, of Education in Chile: The Program 900 Escuelas and the Inter-American Development Bank, and USAID in Latin the MECE-Basica." In Implementing Policy Innovations in America and the Caribbean. Washington, D.C.: Academy Latin America: Politics, Economics and Techniques, edited for Educational Development. by Antonia Silva. Washington, D.C.: Inter-American Akerlof, George (1970). "The Market for Lemons, Qualita- Development Bank Social Agenda Policy Group. tive Uncertainty and the Market Mechanism." Quarterly Arrow, Kenneth J. (1985). "The Economics of Agency." In Journal of Economics, 59:473-494. J. Pratt and R. Zeckhauser, eds., Principals and Agents: Akerlof, G., and P. Romer (1993). "Looting: The Economic The Structure of Business. Cambridge, Mass.: Harvard Underworld of Bankruptcy for Profit." Brookings Papers Business School Press. on Economic Activity, 2:1-73. Arruniada, Benito (1996). "The Economics of Notaries." Alesina, Alberto, and Allan Drazen (1991). "Why Are Sta- European Journal of Law and Economics, 3(1):5-37. bilizations Delayed?" The American Economic Review', Ascher, William (1984). Scheming for the Poor: The Politics of 81(5):1170-1188. Redistribution in Latin America. Cambridge, Mass., and Alexander, William E.; Jeffrey M. Davis; Liam P. Ebrill; London: Harvard University Press. and Carl-Johan Lindgren (1997). Systemic Bank Restruc- Aucoin, Peter (1990). "Administrative Reform in Public turing and Macroeconomic Policy. Washington, D.C.: Management: Paradigms, Principles, Paradoxes and International Monetary Fund. Pendulums." Governance: An InternationalJournal of Pol- Alvarez, Benjamin (1998). "Life Cycle and Legacy of the icy and Administration, 3(2):115-137. Educational Reforms in Latin America and the Aylwin, Mariana; Carlos Bascunian; Sofia Correa; Cristian Caribbean." InternationaljJournal of Educational Reform, Gazmuri; Sole Serrano; and Matias Tagle (1990). Chile 7(1):34-45 (January). en el Siglo XX. Santiago: Planeta. Alvarez, Benjamin, and Monica Ruiz-Casares (eds.) (1997). Bale, Malcolm, and Tony Dale (1998). "Public Sector Reform Paths of Change: Education Reforms Under Way in Latin in New Zealand and Its Relevance to Developing Coun- America and the Caribbean. Washington, D.C.: U.S. tries." The World Bank Research Observer, 13(1):103-122. Agency for International Development/Agency for Edu- Bardhan, Pranab (1997). "The Nature of Institutional cational Development. Impediments to Economic Development." Mimeo- American Bar Association (1989 and 1993). Multinational graph. Berkeley, Calif.: Department of Economics, Uni- Commercial Insolvency. Wisconsin: MG Publishing. versity of California, Berkeley. Ames, Barry (1998). "Institutions and Democracy in Barro, Robert, and Jong-Wha Lee (1996). "New Measures Brazil." Mimeograph. Pittsburgh, Penn.: Department of Educational Attainment." Mimeograph. Cambridge, of Political Science, University of Pittsburgh. Mass.: Department of Economics, Harvard University. 149 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Beaton, Albert E.; Ina V. S. Mullis; Michael 0. Martin; Caribbean." In Shahid Javed Burki, Guillermo Perry, Eugenio J. Gonzalez; Dana L. Kelly; and Teresa A. Smith and Sara Calvo (eds.), Proceedings of the Annual Bank Con- (1996). Mathematics Achievement in the Middle School Years: ference on Development in Latin America, 1997. Washing- lEA's Third International Mathematics and Science Study ton, D.C.: World Bank. (TIMSS). Chestnut Hill, Mass.: International Associa- Buscaglia, Edgardo, and Maria Dakolias (1996). Judicial tion for the Evaluation of Educational Achievement. Reform in Latin American Courts: The Experience in Becker, Gary (1983). "A Theory of Competition Among Argentina and Ecuador. World Bank Technical Paper Pressure Groups for Political Influence." QuarterlyJour- No. 350. Washington, D.C.: World Bank. nal of Economics, 98(3):371-400. Caiden, Naomi, and Aaron Wildavsky (1974). Planning Benston, George J.; Robert A. Eisenbeis; Paul M. Horvitz; and Budgeting in Poor Countries. New York: Wiley. Edward J. Kane; and George G. Kaufman (1986). Per- Calomiris, Charles W. (1996). "Building an Incentive- spectives on Safe and Sound Banking: Past, Present, and Compatible Safety Net: Special Problems for Develop- Future. Cambridge, Mass.: MIT Press. ing Countries." New York: Columbia University. Blau, Peter M. (1974). On the Nature of Organizations. New Calomiris, Charles W. (1997). The Postmodern Bank Safety York: Wiley. Net: Lessons from Developed and Developing Economies. Boston, Jonathan, et al. (1996). Public Management: The Washington, D.C.: The AEI Press. New Zealand Model. Auckland: Oxford University Press. Calomiris, Charles W., and Gary Gorton (1991). "The Ori- Brennan, Geoffrey (1996). "Selection and the Currency of gins of Banking Panics: Models, Facts, and Bank Regu- Reward." Chapter 10 in Robert E. Goodin, ed., The The- lation." In R. Glenn Hubbard, ed., Financial Markets ory of Institutional Design. Cambridge: Cambridge Uni- and Financial Crises. Chicago: University of C.hicago versity Press. Press. Brock, Philip L., ed. (1992a). If Texas Were Chile: A Primer Calvo, Guillermo (1989). "Incredible Reforms." In Debt, on Banking Reform. San Francisco: ICS Press. Stabilization, and Development: Essays in Memory of Carlos Brock, Philip L. (1992b). "The Government-Sponsored Diaz-Alejandro, edited by G. Calvo, et al. Oxford, Eng- Securitization of Mortgages in Nineteenth Century land, and Cambridge, Mass: Basil Blackwell. Chile: The Macroeconomic Consequences of the Caja de Campos, Ed, and Sanjay Pradhan (1996). "Budgetary Insti- Cr6dito Hipotecario, 1855-1880." Working paper. tutions and Expenditure Outcomes: Binding Govern- Brock, Philip L. (1998). "Financial Safety Nets and Incen- ments to Fiscal Performance," Policy Research Working tive Structures in Latin America." Paper commissioned Paper 1646. Washington, D.C.: World Bank, Policy for this chapter. Research Department, Public Economics Division, Sep- Brunetti, Aymo; Gregory Kisunko; and Beatrice Weder tember. (1997a). "Credibility of Rules and Economic Growth." Caprio, Gerard Jr., and Daniela Klingebiel (1996). "Bank Policy Research Working Paper No. 1760. Washing- Insolvency: Bad Luck, Bad Policy, or Bad Banking?" ton, D.C.: World Bank. Washington, D.C.: Development Research Group, Brunetti, Aymo; Gregory Kisunko; and Beatrice Weder World Bank. (1997b). "Institutional Obstacles to Doing Business: Caprio, Gerard Jr.; Michael Dooley; Danny Leipziger; and Region-by-Region Results from a Worldwide Survey of Carl Walsh (1996). "The Lender of Last Resort Function the Private Sector." Policy Research Working Paper No. under a Currency Board: The Case of Argentina." Wash- 1759. Washington, D.C.: World Bank. ington, D.C.: World Bank. Burki, Shahid javed, and Guillermo Perry (1997). The Long Caprio, GerardJr., and Philip Keefer (1998). "Regulations, March: A Reform Agenda for Latin America and the Institutions and Country Responses to Bank Insol- Caribbean in the Next Decade. Washington, D.C.: World vency." Reserarch Proposal (work in progress). Wash- Bank, Latin American and Caribbean Studies View- ington, D.C.: Development Research Group, World points Series. Bank. Burki, Shahid Javed, and Guillermo Perry (1998). Cariola, Patricio, S.J. (1996). "New Criteria for Educa- "Towards Open Regionalism in Latin America and the tional Policies in Chile." In Conference Report on Revital- 150 REFERENCES ization of Education in the Americas, December 4-6, 1995, Cox, Gary W., and Mathew D. McCubbins (1996). "Struc- edited by Benjamin Alvarez. U.S. Agency for Interna- ture and Policy: The Institutional Determinants of Policy tional Development Education and Human Resources Outcomes." Mimeograph. San Diego, Calif.: Department Technical Services Working Paper No. 11. Washington, of Political Science, University of California, San Diego. D.C.: U.S. Agency for International Development. Cull, Robert (1998). "How Deposit Insurance Affects Carlton, Dennis W., and Jeffrey M. Perloff (1990). Modern Financial Depth." Policy Research Working Paper No. Industrial Organization. Harper Collins. 1875. Washington, D.C.: World Bank. Carnoy, Martin (1997). "Is Privatization through Education Dakolias, Maria (1996). The Judicial Sector in Latin America Vouchers Really the Answer? A Comment on the West." and the Caribbean: Elements of Reform. Washington, D.C.: The World Bank Research Observer, 12(l):105-116. World Bank, World Bank Technical Paper No. 319. Carnoy, Martin, and Patrick McEwan (1997). Public Invest- de Juan, Arist6bulo (1987). "From Good Bankers to Bad ments or Private Schools? A Reconstruction of Educational Bankers." Washington D.C.: World Bank. Improvements in Chile. (November 1997 draft): Stanford, de Juan, Arist6bulo (1995). "A Sum Up of 'False Friends' Calif.: Stanford University. in Banking Reform." Seminar on Building Sound Bank- Chandavarkar, Anand G. (1996). Central Banking in Devel- ing in Transition Economies. EDRD Annual Meeting. oping Countries. Great Britain: Macmillan Press Ltd. de Soto, Hernando (1989). The Other Path: The Invisible Rev- Chong, Alberto, and Cesar Calder6n (1997a). "Empirical olution in the Third World. New York: Harper & Row. Tests on the Causality and Feedback Between Institu- Demirgiuc-Kunt, Asli, and Enrica Detragiache (1998). tional Measures and Economic Growth." Mimeograph. "Financial Liberalization and Financial Fragility." Washington, D.C.: World Bank. Washington, D.C.: Development Research Group, Chong, Alberto, and Cesar Calder6n (1997b). "Institu- World Bank. tional Change and Poverty, or Why is it Worth it to Dewatripont, Mathias, and Jean Tirole (1993). "Efficient Reform the State?" Mimeograph. Washington, D.C.: Governance Structure: Implications for Banking Regu- World Bank. lation." In Capital Markets and Financial Intermediation, Chong, Alberto, and Cesar Calder6n (1998). "Institutional C. Mayer and X. Vives, eds. Cambridge: Cambridge Efficiency and Income Inequality: Cross Country University Press, pp. 12-35. Empirical Evidence." Mimeograph. Washington, D.C.: Dewatripont, Mathias, and Jean Tirole (1994). The Pruden- World Bank. tial Regulation of Banks. Cambridge, Mass.: MIT Press. Coase, Ronald (1937). "The Nature of the Firm." Econom- Diamond, Douglas (1989). "Reputation Acquisition in ica, 4:386-405. Debt Markets."Journal of Political Economy, 97:828-862 Correa Sutil, Jorge (ed.) (1994). Situacidn y Politicas Judi- Diamond, Douglas (1991). "Debt Maturity Structure and ciales en America Latina. Santiago: Universidad Diego Liquidity Risk," Quarterly Journal of Economics, Portalis, Cuadernos de Analisis Jurfdico. 106:1027-1054. Correa Sutil, Jorge (1998). "Judicial Reforms in Latin Dixit, Avinash K. (1996). The Making of Economic Policy: A America: Good News for the Underprivileged?" In The Transaction-Cost Politics Perspective. Cambridge, Mass., (Un)Rule of Law in Latin America. Notre Dame: Notre and London: The MIT Press. Dame University Press. Domfnguez, Jorge (1997). "Latin America's Crisis of Rep- Corsetti, Giancarlo; Paolo Pesenti; and Nouriel Roubini resentation." Foreign Affairs, 76(l):100-113. (1998). "What Caused the Asian Currency and Finan- Drazen, Allan, and Vittorio Grilli (1993). "The Benefit of cial Crisis?" Mimeograph. Yale University, Princeton Crises for Economic Reforms." The American Economic University, and New York University. Review, 83:598-607. Cox, Cristian (1998). "Estrategias de Cambio y Factores de Easterbrook, Frank, and Daniel Fischel (1991). The Eco- Exito: Chile en la Decada de 1990." In Educaci6n: La nomic Structure of Corporate Law. Cambridge, Mass.: Har- Agenda del Siglo XXI. Hacia un Desarrollo Humano. vard University Press. Bogota: Programa de Naciones Unidas para el Desar- Edwards, Sebastian (1984). "The Order of Liberalization of rollo/Tercer Mundo Editores. the External Sector in Developing Countries." Princeton 151 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MfATTER Essays in International Finance, 156. Princeton, N.J.: Foster, John; John Greer; and Erick Thorbecke (1984). "A Princeton University. Class of Decomposable Poverty Measures." Econometrica, Edwards, Sebastian, and Daniel Lederman (1998). "The 52(3):761-765. Political Economy of Unilateral Trade Liberalization: Friedman, Milton (1955). "The Role of Government in The Case of Chile." In Going Alone: The Case for Relaxed Education." In Economics and the Public Interest, edited by Reciprocity, edited by J. Bhagwati. Forthcoming. Cam- R. Solow. New Jersey: Rutgers University Press. bridge, Mass., and Washington, D.C.: MIT Press and Fundacion de Investigaciones Economicas Latinoameri- American Enterprise Institute. Also published as canas [FIELI (1996). La Reforma del PoderJudicial en la National Bureau of Economic Research WTorking Paper Argentina. Buenos Aires: FIEL. Series No. 6510, Cambridge, Mass., April. Garber, Peter M. (1997). "The Transition to a Functional Edwards, Sebastian, and Mois6s Naim, editors (1997). Financial Safety Net." In Liliana Rojas-Suarez, ed. Safe Mexico 1994: Anatomy of an Emerging Market Crash. and Sound Financial Systems: What Worksfor Latin America. Washington, D.C.: Carnegie Endowment for Interna- Washington, D.C.: Inter-American Development Bank. tional Peace. Garber, Peter M., and Steven R. Weisbrod (1992). T'he Eco- Engerman, Stanley L.; Stephen H. Haber; and Kenneth L. nomics of Banking, Liquidity, and Money. Lexington, Sokoloff(1998). "Inequality, Institutions, and Differen- Mass.: D.C. Heath and Company. tial Paths of Growth Among New World Economies." Gavin, Michael, and Ricardo Hausmann (1997). "M[ake or Working Paper. Los Angeles: University of California, Buy? Approaches to Financial Market Integration." In Los Angeles. Liliana Rojas-Suarez, ed., Safe and Sound Financial Sys- Espinola, Viola (1997). Descentralizacidn del Sistema Educativo en tems: What Works for Latin America. Washington D.C.: Chile: Impacto en la Gestion de las Escuelas. Human Capital Inter-American Development Bank. Development Group for Latin America and the Caribbean, Geddes, Barbara (1994). Politician's Dilemma: Building State Informe No. 10. Washington, D.C.: World Bank. Capacity in Latin America. Berkeley, Los Angeles, and Evans, Peter (1992). "The State as Problem and Solution: London: University of California Press. Predation, Embedded Autonomy and Structural Giammarino, Ronald M.; Tracy R. Lewis; and Davicl E. M. Change." In Stephan Haggard and Robert R. Kaufman, Sappington (1993). "An Incentive Approach to Bank- eds., The Politics of Economic Adjustment: International ing Regulation." The Journal of Finance, 48 (4): Constraints, Distributive Conflicts and the State. Princeton, 1523-1539. N.J.: Princeton University Press. Gomes, Armando (1996). "The Dynamics of Stock Prices, Federal Deposit Insurance Corporation (1997). "Resolu- Manager Ownership, and Private Benefits of Control." tions Handbook: Methods for Resolving Troubled Manuscript. Cambridge, Mass.: Harvard University. Financial Institutions in the United States." Graham, Carol, and Moises Naim (1998). "The Political Fernandez, Raquel (1997). "Returns to Regionalism: An Economy of Institutional Reforms in Latin America." In Evaluation of Non-Traditional Gains from RTAs." Beyond Trade-Offs: Market Reforms and Equitable Growth in Mimeograph. Washington, D.C.: International Trade Latin America, edited by N. Birdsall, C. Graham, and R. Division, World Bank, February. Sabot. Washington, D.C.: The Brookings Institution. Fernandez, Raquel, and Dani Rodrik (1991). "Resistance Greene, William H. (1993). Econometric Analysis, Second to Reform: Status Quo Bias in the Presence of Individ- Edition. New Jersey: Prentice Hall. ual-Specific Uncertainty." American Economic Review, Grossman, Sanford, and Oliver Hart (1986). "The Costs 81(5):1146-1155. and Benefits of Ownership: A Theory of Vertical and Fetter, Frank W. (1931). Monetary Inflation in Chile. Prince- Lateral Integration." Journal of Political Economy, ton: Princeton University Press. 94:175-202. Filmer, D.; L. Pritchett; and Tan Jee-Peng (1998). Educa- Guido Bejar, Rafael (1997). "The Dynamics of Education tional Attainment Profiles of the Poor (and Rich): DHS Evi- Reform in El Salvador, 1989-1996." In Paths of Change: dence from around the Globe. Unpublished paper. Wash- Education Reforms Under Way in Latin America and the ington, D.C.: World Bank. Caribbean, edited by Benjamin Alvarez and Monica 152 REFERENCES Ruiz-Casares. Washington, D.C.: U.S. Agency for Irwin, Timothy; Michael Klein; Guillermo E. Perry; and International Development/Agency for Educational Mateen Thobani, eds. (1997). Dealing with Public Risk in Development. Private Infrastructure. Washington, D.C.: World Bank, Hanson, E. Mark (1995). "Best (and Worst) Practices in Latin American and Caribbean Studies Viewpoints. Education Decentralization." Washington, D.C.: World Jensen, Michael, and William Meckling (1976). "Theory of Bank. the Firm: Managerial Behavior, Agency Costs, and Hart, Oliver; Rafael La Porta; Florencio L6pez-de-Silanes; Ownership Structure." Journal of Financial Economics, and John Moore (1997). "A New Bankruptcy Procedure 3:305-360 (October). that Uses Multiple Auctions." European Economic Review, Kane, Edward J. (1977). "Good Intentions and Unin- 41:461-473. tended Evil: The Case Against Selective Credit Alloca- Hausmann, Ricardo, and Liliana Rojas-Suarez (1996). tion." Journal of Money, Credit, and Banking, 9: 55-69. Banking Crises in Latin America. Washington, D.C.: Kane, Edward J. (1989). "How Incentive-Incompatible Inter-American Development Bank. Deposit Insurance Funds Fail." National Bureau of Eco- Hirschman, Albert 0. (1970). Exit, Voice, and Loyalty: nomic Research Working Paper No. 2836. Cambridge, Responses to Decline in Firms, Organizations, and States. Mass., February. Cambridge, Mass., and London: Harvard University Keefer, Philip (1995). "Reforming the State: The Sustain- Press. ability and Replicability of Peruvian Reforms of its Hobbs, Cynthia, and Kathryn Rentel (1997). "The Public Administration." Paper presented at the Confer- Paraguayan Education Reform: A Collective Effort," In ence on the Reform of the State, El Pueblo Hotel, Peru, Paths of Change: Education Reforms Under Way in Latin June; second draft, September. America and the Caribbean, edited by Benjamin Alvarez Kellaghan, Thomas; Kathryn Sloane; Benjamin Alvarez; and Monica Ruiz-Casares. Washington, D.C.: U.S. and Benjamin S. Bloom (eds.) (1993). The Home Envi- Agency for International Development/Agency for Edu- ronment & School Learning: Promoting Parental Involvement cational Development. in the Education of Children. San Francisco: Jossey-Bass Holmes, Malcolm, and David Shand (1995). Management Publishers. Reform: Some Practitioner Perspectives on the Past Ten Years. Klein, B., and K. Leffler (1981). "The Role of Market The SOG Ten-Year Reunion "Ten Years of Change," Forces in Assuring Contractual Performance."Journal of draft not for quotation, May 9. Political Economy, 81:815-841. Hopkins, Jack W. (1991). "Evolution and Revolution: Klitgaard, Robert, and Heather Baser (1997). "Working Enduring Patterns and the Transformation of Latin Together to Fight Corruption: State, Society and the American Bureaucracies." In Ali Farazmand, ed., Hand- Private Sector in Partnership," in Suzanne Taschereau book of Comparative and Development Public Administration. and Jose Edgardo L. Campos, eds., Governance Innova- New York: M. Dekker. tions (Lessons from Experience: Building Government-Citizen- Inglehart, Ronald (1994). Codebook for World Values Surveys. Business Partnerships). Ottawa: Institute of Governance. Ann Arbor, Mich.: Institute for Social Research. Klitgaard, Robert (1998). "Healing Sick Institutions." In Institutional Shareholder Services, Inc. (1994). Proxy Voting Silvio Borner and Martin Paldam, eds., The Political Guidelines. Several countries, ISS Global Proxy Services. Dimension of Economic Growth. Basingstoke: Macmillan. Inter-American Development Bank (1996). Economic and Knack, Stephen, and Philip Keefer (1995). "Institutions Social Progress in Latin America, 1996. Washington, and Economic Performance: Cross-Country Tests Using D.C.: Inter-American Development Bank. Alternative Institutional Measures." Economics and Poli- International Monetary Fund (1998). "The Asian Crisis: tics, 7(3):207-227. Causes and Cures." Finance and Development, 35(1): Knack, Stephen, and Philip Keefer (1997a). "Why Don't Poor 18-21. Countries Catch Up? A Cross-National Test of an Institu- Investor Responsibility Research Center (1994). Proxy tional Explanation." Economic Inquiry, 35:590-602 (July). Voting Guide. Global Shareholder Service, Washington, Knack, Stephen, and Philip Keefer (1997b). "Does Social D.C.: I.R.R.C. Capital Have an Economic Payoff? A Cross-Country 153 BEYOND THE WASHINGTON CONSENSUS; INSTITUTIONS MATTER Investigation." Quarterly Journal of Economics, Levine, Ross (1997b). "Napoleon, Bourses, and Growth in 112:1251-1288 (November). Latin America." Conference on The Development of Koch, Timothy W. (1992). Bank Management. Second Edi- Securities Markets in Emerging Economies: Obstacles tion. Florida: The Dydren Press. and Preconditions for Success. Washington: October Krugman, Paul (1998). "What Happened to Asia." 28-29. Mimeograph. Cambridge, Mass.: Massachusetts Insti- Levine, Ross (1997c). "Law, Finance, and Economic tute of Technology. Growth." Mimeograph. Washington, D.C.: Worlcl Bank. Kydland, Finn, and Edward Prescott (1977). "Rules Lindgren, Carl-Johan; Gillian Garcia; and Matthew Saal Rather than Discretion: The Inconsistency of Optimal (1996). Bank Soundness and Macroeconomic Policy. Wash- Plans."Journal of Political Economy, 85:473-491. ington, D.C.: International Monetary Fund. La Porta, Rafael; Florencio L6pez-de-Silanes; Andrei Lowden, Pamela (1996). "The Escuelas Integrales Reform Shleifer; and Robert W. Vishny (1997a). "Legal Deter- Program in Venezuela." In Implementing Policy Inno- minants of External Finance." Journal of Finance, vations in Latin America: Politics, Economics and Tech- 52(3):1131-1150. niques, edited by Antonia Silva. Washington, D.C.: La Porta, Rafael; Florencio L6pez-de-Silanes; Andrei Shleifer; Inter-American Development Bank, Social Agenda and Robert W. Vishny (1997b). "Trust in Large Organiza- Policy Group. tions." AEA Papers and Proceedings, 87(2):333-338. Machado, Ana Luiza (1998). "La Politica Educacional de La Porta, Rafael; Florencio L6pez-de-Silanes; Andrei Minas Gerais, Brasil," Educaci6n: La Agenda del Siglo XXI, Shleifer; and Robert W. Vishny (1998a). "Law and Hacia un Desarrollo Humano. Bogotb: Program de Naciones Finance. "Journal of Political Economy, December. An ear- Unidas para el Desarrollo/Tercer Mundo Editores. lier version was published as National Bureau of Eco- Mainwaring, Scott, and Timothy R. Scully (1995). "Intro- nomic Research Working Paper No. 5661. Cambridge, duction: Party Systems in Latin America." In Building Mass., July 1996. Democratic Institutions: Party Systems in Latin America, La Porta, Rafael; Florencio L6pez-de-Silanes; Andrei edited by S. Mainwaring and T. Scully. Stanford, Calif.: Shleifer; and Robert W. Vishny (1998b). "Corporate Stanford University Press. Ownership Around the World." Mimeograph. Harvard Mainwaring, Scott, and Mathew S. Shugart, eds. (1997). University, March. Presidentialism and Democracy in Latin America. New La Porta, Rafael; Florencio L6pez-de-Silanes; Andrei York: Cambridge University Press. Shleifer; and Robert W. Vishny (1998c). "Agency Prob- Manning, Nick (1998). "Unbundling the State: lems and Dividend Policies Around the World." Autonomous Agencies and Service Delivery." Draft. National Bureau of Economic Research Working Paper Washington, D.C.: Economic Development Institute, No. 6594. Cambridge, Mass., June. World Bank, April 29. Laking, Rob (1996). "Public Management Lessons for Devel- Marcel, Mario (1997). "Effectiveness of the State and oping Countries and the World Bank." Mimeograph. Development Lessons from the Chilean Experience." Lane, Jan-Erik, and Torgeir Nyen (1995). "On the Rele- Paper prepared for World Bank Seminar, "Chile Devel- vance of Neo-Institutionalism to the Public Sector." In opment Lessons and Challenges." Pal Foss, ed., Economic Approaches to Organizations and Martinez, Nestor Humberto (1996). "Judicial Councils in Institutions: An Introduction. Vermont: Dartmouth Pub- Latin America." Chapter 1 in Lessons Learned, Madeleine lishing Co. Crohn, and William E. Davis, eds. Williamsburg, Va.: Larraniaga, Oswaldo (1997). "Chile: A Hybrid Approach." National Center for State Courts. In The Public-Private Mix in Social Services, edited by Mauro, Paolo (1995). "Corruption and Growth." Quarterly Elaine Zuckerman and Emanuel de Kadt. Washington, Journal of Economics, 110:681-712. D.C.: Inter-American Development Bank. McCulloch, Brian W., and Ian Ball (1992). "Accounting in Levine, Ross (1997a). "Financial Development and Eco- the Context of Public Sector Management Reform." nomic Growth: Views and Agenda." Journal of Economic Financial Accountability & Management, 8(l):7-12 Literature, 35 (2):688-726 (June). (Spring). 154 REFERENCES McKinnon, Ronald (1973). Money and Capital in Economic New York State Committee to Review Audio-Visual Cov- Development. Washington, D.C.: Brookings Institution. erage of Court Proceedings (1997). An Open Courtroom: McKinnon, Ronald (1991). The Order of Economic Liberal- Cameras in New York Courts. New York: Fordham Uni- ization: Financial Control in the Transition to a Market versity Press. Economy. Baltimore, Md.: Johns Hopkins University North, Douglass (1990). Institutions, Institutional Change Press. and Economic Performance. New York: Cambridge Uni- Merryman, John H. (1996). "The French Deviation," versity Press. AmericanJournal of Comparative Law, 44(1):109-119. North, Douglass C. (1994). "Economic Performance Meza, Darlyn (1997). Descentralizaci6n Educativa, Organi- through Time." The American Economic Review, zacion y Manejo de las Escuelas a Nivel Local-el Caso de El 84(3):359-368. Salvador: EDUCO. Human Development Group for Nunberg, Barbara (1995). Managing the Civil Service: Reform Latin America and the Caribbean Informe No. 9 (May). Lessons from Advanced Industrialized Countries. World Bank Washington, D.C: World Bank. Discussion Papers 161. Washington, D.C.: World Bank. Mishkin, Frederic S. (1991). "Asymmetric Information and Olson, Mancur (1965). The Logic of Collective Action: Public Financial Crises: A Historical Perspective." In Financial Goods and the Theory of Groups. Cambridge, Mass., and Markets and Financial Crises, edited by R. H. Hubbard. London: Harvard University Press. Chicago: University of Chicago Press. Olson, Mancur (1993). "Dictatorship, Democracy, and Mishkin, Frederic S. (1996). "Understanding Financial Development." American Political Science Review, Crises: A Developing Country Perspective." National 87(3):567-576 (September). Bureau of Economic Research Working Paper No. Organizaci6n de los Estados Americanos (1998). Educaci6n 5600. Cambridge, Mass. en las Am&ricas: Calidady Equidad en el Proceso de Global- Mishkin, Frederic S. (1998). "Systemic Risk, Moral Haz- izaci6n. Washington, D.C.: Organizaci6n de los Estados ard, and the International Lender of Last Resort." Pre- Americanos Secretarfa General. pared for the World Bank conference, Reinventing a Pastor Prieto, Santos (1993). ;Ah de la Justicia! Politica Bretton Woods Committee Conference on the After- Judicial y Economia. Madrid: Editorial Civitas, S.A. math of the Asian Financial Crisis, March 13-14, Paul, Samuel (1991). "Accountability in Public Services: Washington, D.C. Exit, Voice, and Capture." PRE Working Paper 614. Modigliani, Franco, and Merton Miller (1958). "The Cost Paul, Samuel (1998). "Making Voice Work: What Hap- of Capital, Corporation Finance, and the Theory of pened to the Report Card on Bangalore?" PREM Semi- Investment." American Economic Review, 48:261-297 nar Series, March 4. (une). Pearce, David W., ed. (1992). The MIT Dictionary of Mod- Moe, Ronald C. (1994). "The 'Reinventing Government' ern Economics. Cambridge, Mass.: The MIT Press. Exercise: Misinterpreting the Problem, Misjudging the Perris, Lyall (1998). Implementing Education Reforms in New Consequences." Public Administration Review, Zealand: 1987-1997. Education Reform and Manage- 54(2):111-122 (March/April). ment Series 1(1). Washington, D.C.: World Bank. Moe, Terry M. (1984). "The New Economics of Organiza- Perry, Guillermo (1997). "The Political Economy of Finan- tion." AmericanJournal of Political Science, 28:739-775. cial Reforms." Paper presented at the Conference on Moe, Terry M. (1990). "Political Institutions: The Building Robust Banking Systems, World Bank Neglected Side of the Story." Journal of Law, Economics, Annual Meetings, Hong Kong, China. and Organization, 6:213-253. Perry, Guillermo, and Daniel Lederman (1998). Financial Montenegro, Armando (1995). An Incomplete Educational Vulnerability, Spillover Effects, and Contagion: Lessons from Reform: The Case of Colombia. Human Capital Develop- the Asian Crises for Latin America. Washington, D.C.: ment and Operations Policy Working Papers No. 60 World Bank, Latin American and Caribbean Studies (August). Washington, D.C.: World Bank. Viewpoint Series. Morgan, Gareth (1986). Images of Organization. Beverly Plank, David (1996). The Means of Our Salvation: Public Hills, Calif.: Sage. Education in Brazil. Boulder, Colo.: Westview Press. 155 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Posner, Richard (1995). "What Do Judges Maximize?" Rojas, Carlos, and Juan Esquivel (1998). "Las Sistemas de Chapter 3 in Overcoming Law. Cambridge, Mass.: Har- Medici6n de la Calidad de la Educaci6n Latinoameri- vard University Press. cana." Washington, D.C.: World Bank. Posner, Richard A. (1998). "Creating a Legal Framework Rojas-Suarez, Liliana (1997). Safe and Sound Financial Sys- for Economic Development." The World Bank Research tems: What Works for Latin America. Washington, D.C.: Observer, 13(1):1-12. Inter-American Development Bank. Pratt, John, and Richard Zeckhauser, eds. (1985). Princi- Rojas-Suarez, Liliana (1998). "Early Warnings Indicators for pals and Agents: The Structure of Business. Boston: Harvard Banking Problems: What Works for Emerging Markets?" Business School Press. Working Paper. Inter-American Development Bank. Programa de Naciones Unidas para el Desarrollo [PNUDI Rothschild, M. and Joseph E. Stiglitz (1976). "Equilib- (1998). Educacion: La Agenda del Siglo XXI, Hacia un rium in Competitive Insurance Markets." Quarterly Desarrollo Humano. Bogota: Programa de Naciones Journal of Economics, 90:629-50. Unidas para el Desarrollo/Tercer Mundo Editores. Salas, Denis (1998). Le Tiers Pouvoir: Vers une Autre Justice. Putnam, Robert D. (1993). Making Democracy Work: Civic Paris: Hachette Litteratures. Traditions in Modern Italy. Princeton, N.J.: Princeton Salmen, Lawrence F. (1992). Reducing Poverty: An Institu- University Press. tional Perspective. Poverty and Social Policy Series, Paper Rajan, Raghuram, and Luigi Zingales (1995). "What Do No. 1. Washington, D.C.: World Bank. We Know about Capital Structure: Some Evidence from Savoie, Donald J. (1995). "What Is Wrong with the New International Data." Journal of Finance, 50:1421-1460 Public Management?" Canadian Public Administration, (December). 38(1):112-121 (Spring). Rama, German (1992). cAprenden los Estudiantes en el Ciclo Schick, Allen (1996). The Spirit of Reform: Managaing the Bdsico de Educaci6n Media? Montevideo, Uruguay: New Zealand State Sector in a Time of Change. Report pre- CEPAL. pared for the State Services Commission and the Trea- Reid, Gary J., and Graham Scott (1994). Public Sector sury, New Zealand, August. Human Resource Management: Experience in Latin Schick, Allen (1998). "Why Most Developing Countries America and the Caribbean and Strategies for Reform. Should Not Try New Zealand's Reforms." The World Washington, D.C.: World Bank, Green Cover, Report Bank Research Observer, 13(1):123-132. No. 12839, March 14. Schiefelbein, Ernesto (1995). "La Reforma Educativa en Reid, Gary J. (1998). "Performance-Oriented Public Sector America Latina y el Caribe: Un Programa de Acci6n." Modernization in Developing Countries: Meeting the In Proyecto Principal de Educaci6n en America Latina y el Implementation Challenge." Forthcoming in Research in Caribe. Santiago, Chile: UNESCO/OREALC, Boletin Public Administration, edited by Joseph C. M. Raad- 37 (August): 3-34. schelders and James L. Perry. Greenwich, Conn.: JAI Scobie, Tanya D. (1998). "Agenda del Sindicato cle Mae- Press. stros." Educacion: La Agenda del Siglo XXI, Hacia un Reynolds, Thomas, and Arturo Flores (1989). Foreign Law: Desarrollo Humano. Bogota: Programa de Naciones Current Sources of Basic Legislation in Jurisdictions of the Unidas para el Desarrollo/Tercer Mundo Editores. World. Littleton, Colo.: Rothman and Co. Scott, Graham, and Peter Gorringe (1989). "Reforrn of the Riggs, Fred W. (1964). Administration in Developing Coun- Core Public Sector: The New Zealand Experience." Aus- tries: the Theory of Prismatic Society. Boston: Houghton tralian Journal of Public Administration, 48(1):81-92 Mifflin. (March). Rodrik, Dani (1989). "Promises, Promises: Credible Policy Shapiro, Martin (1981). Courts: A Comparative and Political Reform via Signaling." The Economic Journal, Analysis. Chicago: The University of Chicago Press. 99:756-772. Sheng, Andrew (1996). Bank Restructuring. Washington, Rojas, Carlos (forthcoming). "Las Sistemas de Medici6n de D.C.: World Bank. la Calidad de la Educaci6n Latfnoamericana." Washing- Shepherd, Geoffrey, and Sofia Valencia (1996). "Moderniz- ton, D.C.: World Bank. ing the Public Administration in Latin America: Com- 156 REFERENCES mon Problems, No Easy Solutions." Paper prepared for America and the Middle East, edited by R. Dornbusch the conference, "Reforma del estado en America Latina and S. Edwards. Chicago and London: The University of y Caribe." Madrid, October 14-17, 1996. Chicago Press. Shleifer, Andrei, and Robert W. Vishny (1986). "Large United Nations, Economic Commission for Latin America Shareholders and Corporate Control." Journal of Political and the Caribbean (1998). The Fiscal Covenant: Strengths, Economy, 94:461-488 (June). Weaknesses, Challenges. April 23, Santiago, Chile. Simon, Herbert A. (1991). "Organization and Markets." United Nations Educational, Scientific and Cultural Orga- Journal of Economic Perspectives, 5(2):25-44 (Spring). nization (1994). World Education Report 1994. Paris: Steering Committee for the Review of Commonwealth/ United Nations Educational, Scientific and Cultural State Service Provision (1998), Report on Government Ser- Organization. vices: 1998. Melbourne: Government of Australia, United Nations Educational, Scientific and Cultural Orga- Industry Commission. nization (1995). Proyecto Principal de Educaci6n en Armzrica Stigler, George J. (1971). "The Theory of Economic Regu- Latina y el Carihe. Boletin 37 (August). Santiago, Chile: lation." BellJournal of Economic and Management Science, United Nations Educational, Scientific and Cultural 2:3-21. Organization. Stiglitz, Joseph E., and Andrew Weiss (1981). "Credit Valencia, Sofia (1996). "Public Sector Reforms in a Devel- Rationing in Markets with Imperfect Information." oping Country: an Analysis of the Chilean Experience." American Economic Review, 71:393-410 (June). Mimeograph. Stiglitz, Joseph E. (1986). Economics of the PublicSector. New Villalobos R., Sergio (1983). Breve Historia de Chile. Santi- York and London: W W. Norton & Company. ago: Editorial Universitaria, S.A. Stiglitz, Joseph E. (1993). "The Role of the State in Finan- Villegas-Reimers, Eleonora (1998). "The Preparation of cial Markets." Washington D.C.: World Bank's Annual Teachers in Latin America: Challenges and Trends." Conference on Development Economics. LCSHD Paper Series No. 15. Washington, D.C.: World Stiglitz, Joseph E. (1994). Whither Socialism? Cambridge, Bank. Mass., and London: The MIT Press. Vishny, Paul (1994). Guide to International Commerce Law. Summit of the Americas (1998). Action Plan. Santiago, New York: McGraw-Hill. Chile: Organization of American States. Wade, Robert (1997). "How Infrastructure Agencies Moti- Swope, John, and Marcela Latorre (1998). El aporte de Fe y vate Staff: Canal Irrigation in India and the Republic of Alegria a la educaci6n primaria en America Latina: un estu- Korea." In Ashoka Mody, ed., Infrastructure Strategies in dio comparativo longitudinal de la educaci6n primaria impar- East Asia: The Untold Story. EDI Learning Resource tida por Fe y alegrla y las escuzelas piiblicas en nueve palses en Series, Washington, D.C.: World Bank. Ame'rica Latina. Santiago, Chile: Centro de Investigaci6n Wallis, John J., and Douglass C. North (1986). "Measur- y Desarrollo Educativo/Agency for Educational Devel- ing the Transaction Sector in the American Economy, opment/Inter-American Development Bank. 1870-1970." In Long-Term Factors in American Economic Talley, Samuel, and Ignacio Mas (1992). "The Role of Growth, edited by S. L. Engerman and R. E. Gallman. Deposit Insurance." In Dimitri Vittas, ed. Financial Chicago: University of Chicago Press. Regulation: Changing the Rules of the Game. Washington, Weber, Max (1968). Economy and Society: An Outline of Inter- D.C.: Economic Development Institute, World Bank. pretive Sociology. New York: Bedminster Press. Tendler, Judith (1997). Good Government in the Tropics. Bal- West, Edwin G. (1997). "Education Vouchers in Principle timore, Md.: Johns Hopkins University Press. and Practice: A Survey." The World Bank Research Tommasi, Mariano, and Andres Velasco (1996). "Where Observer, 12(l):83-104. Are We in the Political Economy of Reform?"Journal of White, Michelle (1993). "The Costs of Corporate Bankruptcy: Policy Reform, 1:187-238. the U.S.-European Comparison." Manuscript. Ann Arbor: Tornell, Aaron (1995). "Are Economic Crises Necessary for University of Michigan, Department of Economics. Trade Liberalization and Fiscal Reform? The Mexican Williamson, John (1990). "What Washington Means by Experience." In Reform, Recovery, and Growth: Latin Policy Reform." In Latin American Adjustment: How 157 BEYOND THE WASHINGTON CONSENSUS: INSTITUTIONS MATTER Much Has Happened, edited by J. Williamson. Washing- World Bank by Oxford University Press. ton, D.C.: The Institute for International Economics. World Bank (1996). World Development Report 1996. Wash- Williamson, Oliver (1981). "The Modern Corporation." ington, D.C.: Oxford University Press. Journal of Economic Literature, 19:1537-1568. World Bank (1997a). World Development Report 1997: The Williamson, Oliver (1985). The Economic Institutions of Cap- State in a Changing World. New York: Oxford University italism: Firms. Markets, Relational Contracting. New Press. York: Free Press. World Bank (1997b). Colombia: Paving the Way for a Williamson, Oliver (1989). "Transaction Cost Economics." Results-Oriented Public Sector. A World Bank Country Chapter 3 in Handbook of Industrial Organization, Volume Study, Washington, D.C.: World Bank. I, edited by R. Schmalensee and R. D. Willig. New World Bank (1997c). Global Economic Prospects and the York: Elsevier Science Publishers. Developing Countries. Washington, D.C.: World Bank. Williamson, Oliver (1994). "The Institutions and Gover- World Bank (1997d). Private Capital Flows to Developing nance of Economic Development and Reform." Proceed- Countries: The Road to Financial Integration. Washington, ings of the World Bank Annual Conference on Development D.C.: World Bank. Economics, 1994. Washington, D.C.: World Bank. World Bank (1998a). Global Development Finance. Washing- Winkler, Donald (1997). Descentralizacion de la Educaci6n: ton, D.C.: World Bank. Participacion en el Manejo de las Escuelas al Nivel Local. World Bank (1998b). World Development Report 1998. Human Capital Development Group for Latin America "Information and the Financial Markets: The Need for and the Caribbean, Informe No. 8 (May). Washington, Support and Restraint from Policy." Chapter 6 (work in D.C.: World Bank. progress). Washington, D.C.: World Bank. World Bank (1990). World Development Report 1990. Wash- Zaiter, Josefina (1997). "Civil Society and Educational ington, D.C.: World Bank. Reform: The Dominican Experience." In Paths of World Bank (1992). Governance and Development. Washing- Change: Education Reforms Under Way in Latin America ton, D.C.: World Bank. and the Caribbean, edited by Benjamin Alvarez and Mon- World Bank (1993). Social Indicators of Development ica Ruiz-Casares. Washington, D.C.: U.S. Agency for 1991-1992. Baltimore: Johns Hopkins University Press. International Development/Agency for Educational World Bank (1995a). World Development Report 1995. Development. Washington, D.C.: Oxford University Press. Zuckerman, Elaine and Emanuel de Kadt (editors) (1997). World Bank (1995b). Bureaucrats in Business: the Economics The Public-Private Mix in Social Services. Washington, and Politics of Government Ownership, published for the D.C.: Inter-American Development Bank. 158 WORLD BANK LATIN AMERICAN AND CARIBBEAN STUDIES VIEWPOINTS SERIES Determinants of Crime Rates in Latin America and the World: An Empirical Assessment Latin America after Mexico: Quickenintg the Pace by Pablo Fajnzylber, Daniel Lederman, and Norman by Shahid Javed Burki and Sebastian Edwards Loayza forthcoming Poverty, Inequality, and Human Capital Development in Latin America, 1950-2025 Beyond the Washington Consensus: Institutions Matter by Juan Luis Londofio by Shahid javed Burki and Guillermo E. Perry available in English and Spanish available in English and Spanish Dismantling the Populist State: the Unfinished Revolution in Latin America and the Caribbean PROCEEDINGS SERIES by Shahid Javed Burki and Sebastian Edwards Cuirrency Boards and External Shocks: Hou' M'Iuch Pain, Decentralization in Latin America: Learning through How Much Gain? Experience Edited by Guillermo Perry by George E. Peterson Annual World Bank Conference on Development in Latin Urban Poverty and Violence inJamaica Aimerica and the Caribbean 1995: The Challenges of by Caroline Moser and Jeremy Holland Development available in English and Spanish Edited by Shahid Javed Burki, Sebastian Edwards, and Sri-Ram Aiyer Prospects and Challengesfor the Caribbean by Steven B. Webb Anntual World Bank Conference on Development in Latin America and the Caribbean 1996: Poverty and Inequality Black December: Banking Instability, the Mexican Crisis Edited by Shahid Javed Burki, Sri-Ram Aiyer, and and Its Effect on Argentina Rudolf Hommes by Valeriano Garcia Annual World Batnk Cotnference on Development in Latin The Long March:A Reform Agenda for Latin America and America and the Caribbean 1997: Trade: Towvards Open the Caribbean in the Next Decade Regionalism by Shahid Javed Burki and Guillermo E. Perry Edited by Shahid Javed Burki, Guillermo E. Perry, forthcoming in Spanish and Sara Calvo forthcoming Dealing tvith Public Risk in Private Infrastnucture Edited by Timothy Irwin, Michael Klein, Guillermo E. Perry, and Mateen Thobani Crime and Violence as Development Issues in Latin America and the Caribbean by Robert L. Ayres Financial Vulnerability: Spillover Effects, and Contagion: Lessonsfrom the Asian Crisesfor Latin America by Guillermo E. Perry and Daniel Lederman U', THE WORLD BANK 1818 H Street, N.W. Washington, D).C. 20433 USA T'elephone: 202-477-1234 Facsimile: 202-477-6391 Telex: MCI 64145 WORLDBANK MCI 248423 WORLDBANK World Wide Web: http://www.worldbank.org/ E-mail: hooks@qivworldhank.org ISBN 0-8213-4282-7 DESIGN: JEFFREY KIBLERITHE MAGAZINE GROUP/ILLUSTRATION; DAVID Mc~LIIMANS