72803 GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 Rethinking the Role of the State in Finance Rethinking the Role of the State in Finance GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 Rethinking the Role of the State in Finance Washington, D.C. © 2012 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 15 14 13 12 This work is a product of the staff of The World Bank with external contributions. Note that The World Bank does not necessarily own each component of the content included in the work. The World Bank therefore does not warrant that the use of the content contained in the work will not infringe on the rights of third parties. The risk of claims resulting from such infringement rests solely with you. The �ndings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are speci�cally reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 Unported license (CC BY 3.0) http://creativecommons.org/licenses/by/3.0. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank. 2012. Global Financial Development Report 2013: Rethinking the Role of the State in Finance. Washington, DC: World Bank. doi:10.1596/978- 0-8213-9503-5. License: Creative Commons Attribution CC BY 3.0 Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an of�cial World Bank translation. The World Bank shall not be liable for any content or error in this translation. All queries on rights and licenses should be addressed to the Of�ce of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank .org. ISBN (paper): 978-0-8213-9503-5 ISBN (electronic): 978-0-8213-9504-2 DOI: 10.1596/978-0-8213-9503-5 ISSN: 2304-957X Cover photos: Shutterstock Cover design: Naylor Design Contents Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Abbreviations and Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 Benchmarking Financial Systems around the World . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2 The State as Regulator and Supervisor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3 The Role of the State in Promoting Bank Competition . . . . . . . . . . . . . . . . . . . . . . . . 81 4 Direct State Interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 5 The Role of the State in Financial Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Statistical Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 v vi CONTENTS GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOXES O.1 Main Messages of This Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 O.2 Views from Some of the World Bank Clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 O.3 Navigating This Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 1.1 Selecting the Representative Variables for Individual Characteristics. . . . . . . . . . . . .24 1.2 To Aggregate or Not . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29 1.3 China Case Study: Large Banks and the Need to Diversify to Markets . . . . . . . . . . .38 1.4 Romania Case Study: Rapid Growth Enabled by Foreign Funding . . . . . . . . . . . . . .40 2.1 Distorted Incentives: Subprime Crisis and Cross-Border Supervision . . . . . . . . . . . .50 2.2 What Is in the World Bank’s Bank Regulation and Supervision Survey? . . . . . . . . . .56 2.3 Reforming Credit Rating Agencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 2.4 Institutional Structures for Regulation and Supervision. . . . . . . . . . . . . . . . . . . . . . .64 2.5 Impact of the Basel III Implementation in Developing Economies . . . . . . . . . . . . . . .67 2.6 Accounting Standards (Viewpoint by Nicolas Véron) . . . . . . . . . . . . . . . . . . . . . . . .73 2.7 Incentive Audits (Viewpoint by Martin Čihák, Asli Demirgüç-Kunt, and R. Barry Johnston) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75 2.8 Regulatory Discipline and Market Discipline: Opposites or Complements? . . . . . . .77 3.1 Two Views on the Link between Competition and Stability . . . . . . . . . . . . . . . . . . .82 3.2 Decomposing Bank Spreads to Make Inferences about Bank Competition . . . . . . . .84 3.3 Measuring Banking Sector Concentration and Competition . . . . . . . . . . . . . . . . . . .85 3.4 Analyzing Bank Competition Using Disaggregated Business Line Data: Evidence from Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86 3.5 Banking Competition in the Middle East and North Africa . . . . . . . . . . . . . . . . . . .90 3.6 An Econometric Analysis of Drivers of Bank Competition . . . . . . . . . . . . . . . . . . . .95 3.7 Consumer Protection and Competition in South Africa. . . . . . . . . . . . . . . . . . . . . . .97 4.1 Intervention Using State-Owned Banks in Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . .106 4.2 The Recent Global Crisis and Government Bank Lending in Mexico . . . . . . . . . . .108 4.3 State Commercial Banks in Action during the Crisis: The Case of Poland . . . . . . . .109 4.4 Bank Ownership and Credit Growth during the 2008–09 Crisis: Evidence from Eastern Europe and Latin America . . . . . . . . . . . . . . . . . . . . . . . . .110 GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 CONTENTS vii 4.5 Macroeconomic Evidence on the Impact of Government Banks on Credit and Output Cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113 4.6 Two Views on the Role of State-Owned Banks . . . . . . . . . . . . . . . . . . . . . . . . . . . .116 4.7 Development Banks: What Do We Know? What Do We Need to Know? . . . . . . . .120 5.1 Argentina: Using Credit Registry Information for Prudential Supervision . . . . . . . .138 5.2 Egypt: Removing Regulatory Barriers to the Development of a Private Credit Bureau . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .140 5.3 Monopoly Rents, Bank Concentration, and Private Credit Reporting . . . . . . . . . . .141 5.4 Mexico: State Interventions to Prevent Market Fragmentation and Closed User Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143 5.5 Morocco: Public Support for the Development of a Private Credit Bureau . . . . . . .145 5.6 Reforming Large-Value Payment Systems to Mitigate Systemic Risk . . . . . . . . . . . .151 5.7 Italy: Reviving Interbank Money Markets through Collateralized Transactions . . .156 FIGURES O.1 Benchmarking Financial Development, 2008–10. . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 O.2 Selected Features That Distinguish Crisis-Hit Countries . . . . . . . . . . . . . . . . . . . . . . .9 O.3 Market Power and Systemic Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 O.4 Change in Bank Lending Associated with a 1% Increase in GDP Per Capita . . . . . .12 O.5 Credit Reporting vs. Banking System Concentration . . . . . . . . . . . . . . . . . . . . . . . . .14 1.1 Financial Depth and Income Inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20 1.2 Socioeconomic Development, Financial Development, and Enabling Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 1.3 Correlations between Characteristics in Same Category (example) . . . . . . . . . . . . . .25 1.4 Correlations among Financial System Characteristics . . . . . . . . . . . . . . . . . . . . . . . .31 1.5 Financial System Characteristics, by Income Group, 2010 . . . . . . . . . . . . . . . . . . . .34 1.6 The Uneven Nature of Financial Systems (Illustration) . . . . . . . . . . . . . . . . . . . . . . .35 1.7 Financial Systems: 2008–10 versus 2000–07 (Financial Institutions). . . . . . . . . . . . .36 1.8 Financial Systems: 2008–10 versus 2000–07 (Financial Markets) . . . . . . . . . . . . . . .37 B1.3.1 The Chinese Financial Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 B1.4.1 Romania’s Financial Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40 viii CONTENTS GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 2.1 Introduction of Bank Governance Frameworks . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62 2.2 New Insolvency Frameworks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62 2.3 Introduction of Deposit Protection Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63 2.4 Financial Stability Reporting and Stress Test Publication, 1995–2011. . . . . . . . . . . .65 2.5 Push to Implement New Basel Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65 2.6 Impact of the Move to Basel II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66 2.7 Quality of Capital . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66 2.8 Capital Adequacy Ratios: Minimum and Actual . . . . . . . . . . . . . . . . . . . . . . . . . . . .66 B2.5.1 EMDEs: The Impact of Basel III Capital and Liquidity Requirements . . . . . . . . . . . .67 3.1 Five Bank Concentration Ratio (CR5): Developed and Developing Economies. . . . .86 3.2 Five Bank Concentration Ratio (CR5): Developing Regions, Median Values, 1996–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87 3.3 Regulatory Indicators of Market Contestability . . . . . . . . . . . . . . . . . . . . . . . . . . . .88 3.4 Bank Competition: Developed vs. Developing Economies . . . . . . . . . . . . . . . . . . . . .89 3.5 Bank Competition across Developing Regions, 1996–2007 . . . . . . . . . . . . . . . . . . .89 3.6 Bank Competition: Developed vs. Developing Economies . . . . . . . . . . . . . . . . . . . . .91 3.7 Bank Competition across Developing Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92 4.1 Trends in Government Ownership of Banks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103 4.2 Government Ownership across Developing Regions, 1970–2009 . . . . . . . . . . . . . .104 B4.1.1 Ownership and Credit in Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .106 B4.1.2 BNDES: Sources of Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107 B4.1.3 Distribution of BNDES Disbursements by Size . . . . . . . . . . . . . . . . . . . . . . . . . . . .107 B4.2.1 Gross Loan Portfolio Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .108 B4.2.2 Partial Credit Guarantees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .108 B4.3.1 PKO BP’s Loan Share, 2008–11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109 B4.3.2 Nonperforming Loans for PKO BP, 2008–11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109 B4.4.1 Growth of Gross Loans and Bank Ownership in Latin America and Eastern Europe, 2004–2009 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .111 B4.5.1 Evolution of Real GDP and Credit around Recoveries in Economic Activity . . . . .114 B4.5.2 Evolution of Real GDP and Credit around Recoveries in Economic Activity . . . . .115 GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 CONTENTS ix 5.1 The Development of Credit Reporting Institutions, 1980–2012 . . . . . . . . . . . . . . .134 5.2 Prevalence of Credit Reporting by Income Group . . . . . . . . . . . . . . . . . . . . . . . . . .135 5.3 The Reach of Credit Reporting: Who Contributes Information? . . . . . . . . . . . . . . .135 5.4 The Depth of Credit Reporting: What Information Is Collected? . . . . . . . . . . . . . .136 5.5 GDP Turnover of Large-Value Payment Systems by Region, 2009 . . . . . . . . . . . . .149 5.6 The Adoption of Real-Time Gross Settlement Systems over Time, 1990–2010 . . . .150 5.7 Sources of Intraday Liquidity for Participants of Real-Time Gross Settlement Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .153 5.8 Interbank Money Market Rates in the United States and United Kingdom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .154 5.9 Interbank Money Market Rates in Emerging Markets . . . . . . . . . . . . . . . . . . . . . .155 B5.7.1 Interbank Rates in the Italian Collateralized Money Market (MIC) and Other Segments of the Euro Money Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .156 B5.7.2 Outstanding Volumes and Average Maturity Trend on the MIC . . . . . . . . . . . . . . .157 MAPS B2.2.1 Coverage of the 2011 Bank Regulation and Supervision Survey . . . . . . . . . . . . . . . .56 5.1 Credit Information Systems around the World . . . . . . . . . . . . . . . . . . . . . . . . . . . .133 A.1 Depth—Financial Institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .167 A.2 Access—Financial Institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .168 A.3 Ef�ciency—Financial Institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .169 A.4 Stability—Financial Institutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170 A.5 Depth—Financial Markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .171 A.6 Access—Financial Markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .172 A.7 Ef�ciency—Financial Markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173 A.8 Stability—Financial Markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .174 TABLES 1.1 Stylized 4x2 Matrix of Financial System Characteristics (with examples of candidate variables in each category). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 1.2 Financial System Characteristics: Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33 x CONTENTS GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 2.1 Examples of Weak Supervisory Capacity Identi�ed in the FSAP . . . . . . . . . . . . . . . .52 2.2 Differences between Crisis and Noncrisis Countries . . . . . . . . . . . . . . . . . . . . . . . . .57 2.3 Summary of the Basel III Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59 2.4 Summary of Selected Proposals for Regulatory Reform. . . . . . . . . . . . . . . . . . . . . . .69 B3.5.1 Competition in MENA and across Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .90 B3.6.1 Cross-Country Determinants of Banking Competition . . . . . . . . . . . . . . . . . . . . . . .96 B4.4.1 Determinants of the Growth of Total Gross Loans . . . . . . . . . . . . . . . . . . . . . . . . .110 B4.5.1 Credit Cycles and Government Ownership of Banks . . . . . . . . . . . . . . . . . . . . . . . .113 5.1 Credit Reporting, Coverage by Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .134 B5.3.1 Bank Concentration and Credit Reporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .141 A.1 Countries and Their Financial System Characteristics, Averages, 2008–2010 . . . . .161 Foreword he Global Financial Development The World Bank Group has been actively T Report comes at a time when the worldwide � nancial crisis has starkly highlighted the importance of �nancial sys- engaged in financial sector work for some time, aiming to help various parts of the insti- tutional mosaic—including regulation and tems and their role in supporting economic supervision, corporate governance, and �nan- development, ensuring stability, and reducing cial infrastructure—ensure that the �nancial poverty. sector contributes meaningfully to strong and Finance matters, both when it functions inclusive growth. This report seeks to advance well and when it functions poorly. Sup- the global financial sector policy debate, ported by robust policies and systems, �nance highlighting the important perspective of works quietly in the background, contribut- emerging markets and developing economies. ing to economic growth and poverty reduc- It contains a rich array of new �nancial sector tion. However, impaired by poor sector data that are also publicly available as part of policies, unsound markets, and imprudent our Open Data Agenda. institutions, � nance can lay the foundation Sharpening the focus on the central role of for � nancial crises, destabilizing economies, finance in socioeconomic development and hindering economic growth, and jeopardizing understanding how � nancial systems can be hard-won development gains among the most strengthened are crucial if we are to realize vulnerable. our goal of boosting prosperity and eradi- Fostering sustainable financial develop- cating poverty. The Global Financial Devel- ment and improving the performance of opment Report is an important step in this �nancial systems depends on numerous insti- process. tutional factors and stakeholders. The policy maker, the regulator, the banker, and the Jim Yong Kim �nancial consumer must all play their part. President The World Bank Group GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 xi Preface he goal of this inaugural Global Finan- ideological views, instead aiming to develop T cial Development Report is to contrib- ute to the evolving debate on the role of the state in the �nancial sector, highlighted a more nuanced approach to financial sec- tor policy based on a synthesis of new data, research, and operational experiences. from the perspective of development. The The report emphasizes that the state has a report is aimed at a broad range of stakehold- crucial role in the �nancial sector—it needs to ers, including governments, international provide strong prudential supervision, ensure �nancial institutions, nongovernmental orga- healthy competition, and enhance � nancial nizations, think tanks, academics, private sec- infrastructure. Regarding more direct inter- tor participants, donors, and the wider devel- ventions, such as state ownership of banks, opment community. The report offers policy the report presents new evidence that state advice based on research and lessons from involvement can help in mitigating adverse operational work. effects of a crisis. However, the report cau- This marriage of research and operational tions that over longer periods, direct state work was possible thanks to the engagement involvement can have important negative of a diverse set of experts inside and outside effects on the � nancial sector and the econ- the World Bank Group. The report reflects omy. Therefore, as crisis conditions recede, inputs from Bank staff in a broad range of the evidence suggests that it is advisable for units and collaboration with leading research- governments to shift from direct to indirect ers on � nance and development. Reflecting interventions. the close links between financial develop- Because the financial system is dynamic ment and stability, counterparts at the Inter- and conditions are constantly changing, regu- national Monetary Fund have also provided lar updates are essential. Hence, this report valuable contributions. should be seen as part of an ongoing project The report benchmarks � nancial institu- aimed at supporting systematic evaluation, tions and markets around the world, rec- improving data, and fostering broader part- ognizing the diversity of modern financial nerships. Future reports might address �nan- systems. In its analysis of the state’s role in cial inclusion, the development of local cur- � nance, the report seeks to avoid simplistic, rency capital markets, the � nancial sector’s GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 xiii xiv PREFACE GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 role in long-term � nancing, and the state’s and sound � nancial systems for robust eco- role in � nancing health care and pensions. nomic performance. We hope that this new series of analytical reports will prove useful to all stakeholders in Mahmoud Mohieldin promoting evidence-based decision making Managing Director The World Bank Group Acknowledgments he 2013 Global Financial Develop- Christova, Margaret Miller, Leora Klapper, T ment Report reflects the efforts of a broad and diverse group of experts both inside and outside the World Bank. The Shalini Sankaranarayan, Alban Pruthi, and Thilasoni Benjamin Musuku (chapter 5). The report was prepared under the over- report was cosponsored by the World Bank’s sight of Janamitra Devan, Vice President Financial and Private Sector Development (FPD and IFC); Justin Yifu Lin, Chief Econo- Vice Presidency (FPD) and the Development mist and Senior Vice President (DEC); and Economics Vice Presidency (DEC). It reflects Martin Ravallion, Acting Chief Economist inputs from World Bank Group staff across a and Senior Vice President (DEC). World range of units, including all the regional vice Bank Presidents Robert B. Zoellick and Jim presidencies, the Poverty Reduction and Eco- Yong Kim and Managing Director Mahmoud nomic Management Network, and External Mohieldin provided overall guidance. The Affairs, as well as staff of the International authors received invaluable advice from the Finance Corporation (IFC). FPD Council (Aslı Demirgüç-Kunt, Augusto Aslı Demirgüç-Kunt was the director of Lopez-Claros, Gaiv Tata, Gerardo Corro- this project. Martin C ˇ ihák led the core team, chano, Janamitra Devan, Klaus Tilmes, Loic which included Cesar Calderón, Martin Chiquier, Marialisa Motta, Pierre Guislain, Kanz, Subika Farazi, and Mauricio Pinzon Sujata Lamba, Tilman Ehrbeck, and Tunc Latorre. Other key contributors were Erik Uyanik) as well as the World Bank–Interna- Feyen (chapter 1); Maria Soledad Martínez tional Monetary Fund Financial Sector Liai- Pería (chapters 2, 3, and 4); I˙nci Ötker-Robe, son Committee. Martín Vázquez Suárez, Miquel Dijkman, Peer reviewers of the report were Stijn Valeria Salomao Garcia, R. Barry Johnston, Claessens, Augusto de la Torre, Ross Levine, and Nicolas Véron (chapter 2); Thorsten Beck Norman Loayza, Roberto Rocha, and Tunc and Klaus Schaeck (chapter 3); Marcin Piat- Uyanik. Luis Servén also reviewed the con- kowski, Eva Gutierrez, José De Luna Mar- cept note. Comments on individual chapters tinez, Carlos Leonardo Vicente (chapter 4); were also received from Aart Kraay, Ross Ouarda Merrouche, Miriam Bruhn, Mas- Levine, Roberto Rocha, and Sergio Schmuk- simo Cirasino, Marco Nicoli, Maria Teresa ler (chapter 1); Gerard Caprio, Patrick Hono- Chimienti, Froukelien Wendt, Luchia Marius han, Alain Ize, Ross Levine, and Damodaran GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 xv xvi ACKNOWLEDGMENTS GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 Krishnamurti (chapter 2); Franklin Allen, assistance was provided by Hedia Arbi, Gra- Thorsten Beck, Michael Fuchs, and Martha cia Sorensen, and Agnes Yaptenco. Other Martinez Licetti (chapter 3); and Viral Acha- valuable assistance was provided by Benja- rya, Charles Calomiris, Heinz Rudolph, and min Levine and Vin Nie Ong. Sergio Schmukler (chapter 4). Aart Kraay Mauricio Pinzon Latorre and Subika reviewed all chapters for consistency and Farazi were instrumental in compiling and quality multiple times. updating the databases underlying the report. The authors also received valuable sug- In so doing, they bene�ted from the work of gestions and other contributions at various the current FinStats database team, which stages of the project from Hormoz Aghadey, includes Katie Kibuuka and Diego Sour- Shamshad Akhtar, Deniz Anginer, Mad- rouille, who in turn relied on key efforts from elyn Antoncic, Zso�a Arvai, Steen Byskov, previous FinStats team members, including Kevin Carey, Jeffrey Chelsky, Loic Chiquier, Ed Al-Hussainy, Haocong Ren, and Andrea Gerardo Corrochano, Mariano Cortes, Rob- Coppola. Joanna Nasr, Mariana Carvalho, ert Cull, Stefano Curto, Mansoor Dailami, and Zarina Odinaeva helped with the data Katia D’Hulster, Maya Eden, Tilman Ehr- on the credit information systems used in beck, Matthias Feldmann, Aurora Ferrari, chapter 5. Manuela Ferro, Jose Antonio Garcia, Egbert The work on the 2011 update of the Gerken, Swati Ghosh, David Gould, Neil Banking Regulation and Supervision Survey Gregory, Mario Guadamillas, Pankaj Gupta, started with the collaboration of Maria Sole- Mary Hallward-Driemeier, Darrin Hartzler, dad Martínez Pería, Roberto Rocha, Con- Richard Hinz, Mustafa Zakir Hussain, Sujit stantinos Stephanou, and Haocong Ren. The Kapadia, Isfandyar Khan, Thomas Kirch- survey benefited from contributions from meier, Kalpana Kochhar, Rachel Kyte, Jeffrey numerous banking regulation experts in the Lewis, Samuel Maimbo, Mariem Malouche, World Bank, including David Scott, Krish- Cledan Mandri-Perrott, Claire Louise namurti Damodaran, Katia D’Hulster, Ced- McGuire, Martin Melecky, Dino Merotto, ric Mousset, and others outside the World Sebastian Molineus, Fredesvinda Montes, Bank, in particular, Michael Andrews and Cedric Mousset, Nataliya Mylenko, Makoto Jan-Willem van der Vossen. Insights and Nakagawa, Harish Natarajan, Aloysius Uche encouragement from Gerard Caprio, Ross Ordu, Jorge Patiño, Jean Pesme, Tigran Pog- Levine, and James Barth, who organized hosyan, John Pollner, Daniel Pulido, Hao- the previous rounds of the survey, are grate- cong Ren, Ivan Rossignol, Heinz Rudolph, fully acknowledged. PKF (UK) and Auxilium Consolate Rusagara, Andre Ryba, David helped with compiling and following up on Scott, James Seward, Sophie Sirtaine, Con- the survey responses. Amin Mohseni pro- stantinos Stephanou, Mark Stone, Vijay Tata, vided excellent research assistance on the Marilou Uy, S. Kal Wajid, Juan Zalduendo, survey. Catiana Garcia-Killroy (FPD), Dilek Laura Zoratto, and participants in seminars Aykut and Eung Ju Kim (both DEC), and and brie�ngs organized at the World Bank. Isabella Reuttner (World Economic Forum) The report would not be possible with- provided helpful consultations on data. Tariq out the production team, including Merrell Khokhar, Neil Fantom, Ibrahim Levent, and Tuck-Primdahl and Nicole Frost, as well as William Prince were instrumental in integrat- Stephen McGroarty, Santiago Pombo, Jose ing the report’s data with the World Bank’s De Buerba, Jane Zhang, Ryan Hahn, Mary Open Data Initiative. Donaldson, and Xenia Zia Morales. Aziz The authors would also like to thank the Gokdemir was the production editor, with many country of�cials and other experts who Debra Naylor as the graphic designer. Roula participated in the surveys underlying this Yazigi assisted the team with the website report, including the Bank Regulation and and communications. Paul Holtz was the Supervision Survey and the Financial Devel- language editor. Excellent administrative opment Barometer. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 ACKNOWLEDGMENTS xvii Financial support from State Secretariat Change program and the Research Support for Economic Affairs (Switzerland) is grate- Budget provided funding for the underlying fully acknowledged. The latest update of the research program in DEC. Frank Sader had Bank Regulation and Supervision Survey and a key role in FPD’s fundraising efforts for the related research was � nanced with � nancial Global Financial Development Report. support from the U.K. Department for Inter- national Development. The Knowledge for EXTERNAL ADVISERS Viral Acharya CV Starr Professor of Economics, New York University Stern School of Business; Program Director for Financial Economics, Centre for Economic Policy Research Franklin Allen Nippon Life Professor of Finance and Professor of Economics at the Wharton School of the University of Pennsylvania Thorsten Beck Professor of Economics and Chairman of the European Banking Center, Tilburg University, Netherlands Charles Calomiris Henry Kaufmann Professor of Financial Institutions, Graduate School of Business, Columbia University Gerard Caprio William Brough Professor of Economics and Chair, Center for Development Economics, Williams College Stijn Claessens Assistant Director, Research Department, International Monetary Fund Patrick Honohan Governor, Central Bank of Ireland R. Barry Johnston Former Assistant Director, Monetary and Capital Markets Department, International Monetary Fund Ross Levine James and Merryl Tisch Professor of Economics; Director, William R. Rhodes Center for International Economics and Finance, Department of Economics, Brown University Monica Rubiolo Head of Macroeconomic Support, State Secretariat for Economic Affairs, Switzerland Klaus Schaeck Professor of Empirical Banking, Bangor University Nicolas Véron Senior Fellow, Bruegel Institute; Visiting Fellow, The Peterson Institute for International Economics The report also bene�ted from suggestions and insights from country of�cials and other experts participating in the Financial Development Barometer and the other surveys and dis- cussions underlying this report. The �ndings, interpretations, and conclusions expressed in this report do not necessarily reflect the views of the advisers or institutions with which they are af�liated. xviii ACKNOWLEDGMENTS GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 PEER REVIEWERS Stijn Claessens Assistant Director, Research Department, International Monetary Fund Augusto de la Torre Chief Economist, Latin America and the Caribbean Vice Presidency, World Bank Ross Levine James and Merryl Tisch Professor of Economics; Director, William R. Rhodes Center for International Economics and Finance, Department of Economics, Brown University Norman Loayza Lead Economist and Director, 2014 World Development Report: Risks, Vulnerabilities, and the Crisis, World Bank Roberto Rocha Senior Adviser, Financial and Private Sector Vice Presidency, World Bank Tunc Uyanik Director, Financial Systems Global Practice and East Asia and Paci�c Region, Financial and Private Sector Vice Presidency, World Bank Abbreviations and Glossary ATP/TA after-tax pro�ts to assets e-MID Electronic Market for Interbank BANSEFI Banca de Ahorro Nacional y Deposit Servicios Financieros FIRA Fideicomisos Instituidos en BB Banco do Brasil Relación con la Agricultura, BCB Banco Central do Brasil Mexico BCBS Basel Committee for Banking FIRST Financial Sector Reform and Supervision Strengthening Initiative BIS Bank for International FOGAPE State-Owned Guarantee Fund Settlements for Small Entepreneurs, Chile BNDES Banco Nacional de FSA Financial Sector Assessment Desenvolvimento Econômico e FSAP Financial Sector Assessment Social (state-owned development Program bank, Brazil) FSB Financial Stability Board BTP/TA before-tax pro�ts to assets FSSA Financial System Stability CCP central counterparty Assessment CEF Caixa Econômica Federal GCC Gulf Cooperation Council CoCo contingent capital GDP gross domestic product CPSIPS Core Principles for Systemically GOB government-owned bank Important Payment Systems GTS global trading system CPSS Committee on Payment and HHI Her�ndahl-Hirschman index (of Settlement Systems market concentration) CR5 concentration ratio (share of IDB Inter-American Development the �ve largest banks in total Bank banking system assets) IFC International Finance DB development bank Corporation DNS deferred net settlement IFRS International Financial DTAs deferred tax assets Reporting Standards EAP East Asia and Paci�c IMF International Monetary Fund ECA Europe and Central Asia IOSCO International Organization of EMDEs emerging markets and Securities Commissions developing economies IRB international ratings-based GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 xix xx A B B R E V I AT I O N S A N D G L O S S A R Y GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 KfW Kreditanstalt für Wiederaufbau, PKO BP PKO Bank Polski Germany PRISM Pakistan Real Time Interbank KOTEC Korean government guarantor Settlement Mechanism LAC Latin America and the PSEFT Payment System and Electronic Caribbean Fund Transfer LIBOR London interbank offered rate PwC Pricewaterhouse Coopers LLP loan loss provisioning RCCP Recommendations for Central M2 M2 measure of money supply Counterparties MENA Middle East and North Africa ROA return on assets MFI micro�nance institution RSSS Recommendations for Securities MIC Collateralized Interbank Settlement Systems Market (Italy) RTGS real-time gross settlement MSR mortgage servicing rights RWA risk-weighted assets NAFIN Nacional Financiera, Mexico SAR Special Administrative Region NBFI nonbank �nancial institution SBP State Bank of Pakistan NBP National Bank of Poland SECO State Secretariat for Economic NI net interest income Affairs, Switzerland NII non-interest income SELIC Sistema Especial de Liquidação NPL nonperforming loan e de Custódia NPS national payment system SIFIs systemically important �nancial NSFR net stable funding ratio institutions OECD Organisation for Economic SME small and medium enterprise Co-operation and Development SSA Sub-Saharan Africa OLS ordinary least squares STR Sistema de Transferência de OTC over the counter Reservas OV overhead costs TA/A taxes to assets P/E price-to-earnings ratio GLOSSARY OF KEY TERMS USED THROUGHOUT THE REPORT The �nancial The �nancial system in a country is de�ned to include �nancial insti- system tutions (banks, insurance companies, and other nonbank � nancial institutions) and � nancial markets (such as those in stocks, bonds, and �nancial derivatives). It also includes the �nancial infrastructure (which includes, for example, credit information–sharing systems and payment and settlement systems). Financial Conceptually, �nancial development is a process of reducing the costs development of acquiring information, enforcing contracts, and making transac- tions. Empirically, measuring �nancial development directly is chal- lenging. Instead, the report measures four �nancial system character- istics (depth, access, ef�ciency, and stability) for �nancial institutions and �nancial markets (“4x2 framework�). The state The state is de�ned in a broad economic sense, to include not only the country’s government but also autonomous or semiautonomous agen- cies such as a central bank or a �nancial supervision agency. The roles of the The roles of the state in the � nancial sector include those of a pro- state moter, owner, regulator, and overseer. The report focuses on areas that were highlighted by the crisis and are of particular relevance for �nancial development. Country A territorial entity for which statistical data are maintained and pro- vided internationally on a separate and independent basis (not neces- sarily a state as understood by international law and practice). Overview O n September 15, 2008, the failure of the U.S. investment banking giant Lehman Brothers marked the onset of the larg- Which lessons about the connections between finance and economic development should shape policies in coming decades? est global economic meltdown since the Great On the surface, the main contrast between Depression. The aftershocks have severely this global crisis and those in recent decades is affected the livelihoods of millions of people that developed economies were affected much around the world. The crisis triggered policy more strongly and more directly than were steps and reforms designed to contain the cri- developing economies. But some developed sis and to prevent repetition of these events. �nancial systems (such as those of Australia, Four years later, with banking woes ongo- Canada, and Singapore) have shown remark- ing in various parts of the world (most nota- able resilience so far, while some developing bly in the euro area), it is a good time to ones have been brought to the brink of col- evaluate these reforms and their likely con- lapse. The bigger point is that the quality of tribution to long-run � nancial development. a state’s policy for the � nancial sector mat- The crisis experience is thus an important ters more than the economy’s level of devel- part of the motivation for this inaugural opment. This report reassesses the role of the Global Financial Development Report. The state in �nance, based on updated data, ongo- crisis has prompted many people to reassess ing research, and World Bank Group experi- various official interventions in financial ences from around the world. systems, from regulation and supervision of Two building blocks underlie the report’s financial institutions and markets, to com- view of the role of the state in � nance. First, petition policy, to state guarantees and state there are sound economic reasons for the ownership of banks, and to enhancements in state to play an active role in � nancial sys- �nancial infrastructure. tems. Second, there are practical reasons to But the crisis does not necessarily negate be wary of the state playing too active a role the considerable body of evidence on these in �nancial systems. The tensions inherent in topics accumulated over the past few decades. these two building blocks emphasize the com- It is important to use the crisis experience to plexity of �nancial policies. Though econom- examine what went wrong and how to � x it. ics identi�es the social welfare advantages of GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 1 2 OVERVIEW GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 certain government interventions, practical Nevertheless, with ample reservations and experience suggests that the state often does cautions, this report teases out broad lessons not intervene successfully. Furthermore, since for policy makers from a variety of experi- economies and the state’s capacity to regu- ences and analyses (see box O.1 for a sum- late differ across countries and over time, mary of the main messages). the appropriate involvement of the state in The state tends to play a major role in the �nancial system also varies case by case. the modern financial sector, as promoter, BOX O.1 Main Messages of This Report The report’s overall message is cautionary. The global bility. However, research presented in this report � nancial crisis has given greater credence to the idea suggests that, for the most part, factors such as poor that active state involvement in the � nancial sector regulatory environment and distorted risk-taking can help maintain economic stability, drive growth, incentives promote instability, rather than competi- and create jobs. There is evidence that some interven- tion itself. With good regulation and supervision, tions may have had an impact, at least in the short bank competition can help improve efficiency and run. But there is also evidence on potential longer- enhance access to � nancial services, without neces- term negative effects. The evidence also suggests that, sarily undermining systemic stability. Rather than as the crisis subsides, there may be a need to adjust restricting competition, it is necessary to address the role of the state from direct interventions to less distorted competition, improve the flow of informa- direct involvement. This does not mean that the state tion, and strengthen the contractual environment. should withdraw from overseeing �nance. To the con- Lending by state-owned banks can play a positive trary, the state has a very important role, especially in role in stabilizing aggregate credit in a downturn, but providing supervision, ensuring healthy competition, it also can lead to resource misallocation and dete- and strengthening �nancial infrastructure. rioration of the quality of intermediation. The report Incentives are crucial in the � nancial sector. The presents some evidence that lending by state-owned main challenge of �nancial sector policies is to better banks tends to be less procyclical and that some align private incentives with public interest without state-owned banks even played a countercyclical role taxing or subsidizing private risk-taking. Design of during the global �nancial crisis. However, the track public policy needs to strike the right balance—pro- record of state banks in credit allocation remains gen- moting development, yet in a sustainable way. This erally unimpressive, undermining the bene�ts of using approach leads to challenges and trade-offs. state banks as a countercyclical tool. Policy makers In regulation and supervision, one of the crisis les- can limit the inef�ciencies associated with state bank sons is the importance of getting the “basics� right credit by paying special attention to the governance � rst. That means solid and transparent institutional of these institutions and schemes and ensuring that frameworks to promote financial stability. Specifi- adequate risk management processes are in place. cally, it means strong, timely, and anticipatory super- However, this oversight is challenging, particularly in visory action, complemented with market discipline. weak institutional environments. In many developing economies, that combination of Experience points to a useful role for the state in basic ingredients implies a priority on building up promoting transparency of information and reducing supervisory capacity. Here, less can mean more: less counterparty risk. For example, the state can facili- complex regulations, for instance, can mean more tate the inclusion of a broader set of lenders in credit effective enforcement by supervisors and better moni- reporting systems and promote the provision of high- toring by stakeholders. quality credit information, particularly when there The evidence also suggests that the state needs to are signi�cant monopoly rents that discourage infor- encourage contestability through healthy entry of mation sharing. Also, to reduce the risk of freeze-ups well-capitalized institutions and timely exit of insol- in interbank markets, the state can create the condi- vent ones. The crisis fueled criticisms of “too much tions for the evolution of markets in collateralized competition� in the �nancial sector, leading to insta- liabilities. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 OVERVIEW 3 owner, regulator, and overseer. Indeed, eco- that pay off, bank owners reap the pro�ts. nomics provides several good motivations But when such gambles fail, the bank may for an active role for the state in finance. not bear the full cost. For example, bail- These motivations reflect the effects of “mar- outs of troubled banks spread the cost of ket imperfections,� such as the costs and failed bets broadly among others in society uncertainties associated with (a) acquiring who had no connection to the original risky and processing information, (b) writing and investment decision. This potential for cas- enforcing contracts, and (c) conducting trans- cading events can be a reason for the state to actions. These market imperfections often intervene by imposing “speed limits� on risk create situations in which the actions of a few taking by banks. people or institutions can adversely influence Third, limitations on the ability of people many other people throughout society. These to process information, and the tendency of externalities provide the economic rationale some people to follow the crowd, can moti- for the government to intervene to improve vate governments to take an active role in the functioning of the �nancial system. �nancial markets. For example, when people A few examples demonstrate how market have dif�culty fully understanding complex imperfections motivate government action. investments or do not appreciate the possibil- First, when one bank fails, this can cause ity of rare but extreme events, this can lead depositors and creditors of other banks to investors to make systematic mistakes, which become nervous and start a run on these can jeopardize the stability of the economy, other banks. This “contagion�—whereby the with potentially adverse ramifications for weakness in one bank can cause stress for people who neither make those investments otherwise healthy �nancial institutions—can nor have any influence over those that do. reverberate through the economy, causing Governments can limit the adverse reper- problems for the individuals and � rms that cussions of these market failures. For exam- rely on those otherwise healthy institutions. ple, regulation and supervision can limit risk This is the classic bank run. taking by � nancial institutions to avoid the A second example stresses the externali- potential externalities associated with �nan- ties associated with risk taking, especially cial fragility. Also, authorities can regulate for large � nancial institutions. For the sake information disclosure to facilitate sound of this illustration, imagine a busy road with decisions, and even regulate � nancial prod- cars and trucks. If a car or truck goes faster, it ucts, similar to how governments regulate can get to its destination sooner, but there is a the sale of food and drugs. Thus, economics chance that it will be involved in a crash. The provides many reasons for an active role of likelihood of a crash is small but it increases the state in �nance. with speed. Crashes involving large vehicles But just because the state can ameliorate are particularly costly to others involved in market imperfections and improve the oper- the crash and very disruptive to traf�c in gen- ation of �nancial systems does not mean that eral. Nobody wants to be involved in a crash, it will. Designing and enforcing appropriate of course. But when deciding on how fast to policy can be tricky. Returning to the previ- go, a car or truck driver may not fully con- ous analogy with speed limits for cars and sider the costs that a crash might have on oth- trucks, having a single speed limit may not ers in terms of injuries, damages, time lost in seem very effective, because some vehicles traf�c jams, and so on. The state can play a have better safety features, such as braking role, for example by imposing and enforcing systems, and therefore are less likely to end speed limits, and perhaps imposing stricter up in a crash. If vehicles with better brakes regulation of vehicles that pose bigger risks, were allowed to go faster, they could spend such as large trucks. less time on the road, and traf�c could ease Similarly, financial institutions often do up. But brake quality is dif�cult to monitor not bear the full risks of their portfolios. in real time. So, differentiated speed lim- When a large bank makes risky investments its can be difficult to design and enforce, 4 OVERVIEW GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 resulting in more speeding and crashes. The objectives, including less altruistic ones, such state could also intervene directly by pro- as helping friends, family, cronies, and politi- viding government-approved drivers for all cal constituents. When this happens, the gov- cars and trucks. That way, the state can have ernment can do serious harm in the �nancial more control over safety and soundness, but system. These arguments suggest a sober it can become quite expensive for taxpay- wariness concerning the role of the state in ers. Alternatively, the state could build large �nance that will vary according to con�dence speed bumps on the road, so that there are in the political system’s ability to promote the almost no crashes; however, traffic would public good. slow down to a crawl. Determining the proper role of the state in The analogy underscores that correct- �nance is thus as complex as it is important: ing market imperfections is a complicated one size does not �t all when it comes to pol- task, requiring considerable information and icy intervention. In less developed economies, expertise to design, implement, and enforce there may seem to be more scope for the gov- sound policies. State interventions in �nance ernment’s involvement in spearheading �nan- need to be risk-sensitive, but measuring risk cial development. However, less development properly and enforcing risk-based regulations is often accompanied by a less effective insti- is far from straightforward. The state can try tutional framework, which in turn increases to run parts of the � nancial system directly, the risk of inappropriate interventions. And but evidence shows that approach to be very the role of the state naturally changes as the costly. And if the state required banks to hold �nancial system creates new products, some capital as large as their loans, the risk of fail- of which obviate the need for particular poli- ures would be minimal, but � nancial inter- cies while others motivate new government mediation would grind to a halt since banks interventions. Reflecting this complexity, would not be able to lend. country of�cials and other � nancial sector An important complicating factor is that experts often hold opposing views and opin- the same government policies that ameliorate ions on the pros and cons of various state one market imperfection can create other— interventions—a point illustrated by a recent sometimes even more problematic—distor- informal global opinion poll carried out by tions. For example, when the government the Global Financial Development Report insures the liabilities of banks to reduce the team (box O.2). possibility of bank runs, the insured credi- The Global Financial Development tors of the bank may not diligently monitor Report provides new insights on financial the bank and scrutinize its management. development and the role of the state in �nan- This can facilitate excessive risk taking by cial systems, building on the experience from banks. The state can try to limit risk tak- the global �nancial crisis. Varying economic ing by large, interconnected � nancial insti- and political circumstances across countries tutions. However, such interventions might imply that financial sector policies require reduce the incentives of private shareholders customization: appropriate policies will dif- to exert strong corporate control over these fer across countries and over time. But there institutions, because they think the govern- are common lessons and guidelines. While ment is already doing it. Thus, state interven- recognizing the complexity of the issue and tions can create even more reliance on the the limits of existing knowledge, this report state. contributes new data and analysis to the pol- An even deeper issue is whether the state icy discussion. always has suf�cient incentives to correct for market imperfections. Governments do not BENCHMARKING FINANCIAL always use their powers to address market SYSTEMS imperfections and promote the public inter- est. Sometimes, government officials use A growing body of evidence shows that the power of the state to achieve different � nancial institutions and � nancial markets GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 OVERVIEW 5 BOX O.2 Views from Some of the World Bank Clients As part of its effort to find out more about client firmed various areas of agreement. For example, country views, the Global Financial Development there is a widespread notion that state-owned Report team carried out an informal global poll— � nancial institutions and government-backed credit the 2011/12 Financial Development Barometer. This guarantees can in principle play a useful role. The poll, which covered country of� cials and � nancial poll also shows many respondents seeing potential sector experts from 78 countries (23 developed and bene�ts in more stringent supervision of new � nan- 55 developing), provides interesting insights into cial instruments in light of the crisis. A majority views about � nancial development and the role of also see a scope for a more active role of the state the state in � nance. in promoting technological innovations in � nancial Despite the crisis experience, 90 percent of the infrastructure. country officials and experts surveyed in the poll Perhaps more interestingly, the poll also indi- perceive that positive effects of � nance (in particular cated many key policy areas where the views for and those on economic growth and poverty reduction) against are almost evenly split. This split includes, outweigh its potential negative effects. A majority for example, opinions on the need for stringency of the respondents therefore see that their country’s and greater scope of regulation and supervision, the � nancial sector needs to grow, especially in terms pros and cons of greater competition in countries’ of � nancial markets and nonbank � nancial institu- � nancial systems, the possible countercyclical role tions, to better serve its clients and expand to new of state-owned � nancial institutions, and the role of ones. the state in promoting information sharing—all top- As regards the role of the state in the � nancial ics that are examined in the current Global Finan- sector, the Financial Development Barometer con- cial Development Report. Selected Responses from the 2011/12 Financial Development Barometer Views were split on important aspects of the state’s role . . . Agree? (%) “In view of the global �nancial crisis, more stringent �nancial sector regulation and supervision is needed.� 49 “In view of the global �nancial crisis, there is a need for broadening the scope of �nancial sector regulation and supervision.� 54 “More �nancial sector competition would help �nancial stability in my home country.� 58 “State-owned �nancial institutions played an effective countercyclical role during the recent global �nancial crisis.� 48 “Government-backed credit guarantee schemes do play an important role in promoting �nancial stability.� 64 “The development of collateral registries can be left, fully or mostly, to the private sector.� 42 Note: The Financial Development Barometer is an informal global poll covering country of�cials and � nancial sector experts from 78 economies (23 developed and 55 developing). The response rate was 65 percent. Results are percentages of total responses received. exert a powerful influence on economic the banking industry as a proxy for �nancial development, poverty alleviation, and the development. However, size is not a measure stability of economies around the world. Yet of quality, ef�ciency, or stability. Moreover, measuring the functioning of the financial the banking sector is only one component system has important shortcomings. Indeed, of �nancial systems. This report, along with empirical work has largely—though not the accompanying public database, assembles exclusively—relied on measures of the size of and improves cross-country data that can be 6 OVERVIEW GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 used to benchmark �nancial systems. Chap- less deep and also somewhat less ef�cient and ter 1 addresses questions such as: How can to provide less access, their stability has been one empirically describe different charac- comparable to developed-country � nancial teristics of � nancial systems? How can one systems. These measures are then used to compare � nancial systems across countries characterize and compare � nancial systems and regions and through time? How have across countries and over time, highlight- �nancial systems been affected by the global ing the multidimensional nature of � nancial � nancial crisis, and what are the key recent development. Country-by-country informa- trends? tion on the key �nancial system characteris- To measure and benchmark �nancial sys- tics is presented in the Statistical Appendix, tems, the report develops several measures with more data available through the report’s of four characteristics of financial institu- website. tions (banks, insurance companies, and so on) and �nancial markets (stock markets and RETHINKING THE ROLE OF bond markets): (a) the size of �nancial insti- THE STATE IN THE FINANCIAL tutions and markets (�nancial depth), (b) the SECTOR degree to which individuals can and do use financial institutions and markets (access), The report addresses the following key pol- (c) the efficiency of financial institutions icy questions: (a) What is the early postcrisis and markets in providing � nancial services thinking on transforming regulatory prac- (ef�ciency), and (d) the stability of � nancial tices around the world? (b) How should gov- institutions and markets (stability). These ernments promote competition in the � nan- four characteristics are measured both for cial sector without planting the seeds of the � nancial institutions and � nancial markets, next crisis? (c) When do direct government leading to a 4x2 matrix of the characteristics interventions—such as state ownership and of �nancial systems. A basic comparison (�g- guarantees—help in developing the � nancial ure O.1) con� rms that although developing- sector, and when do they fail? and (d) What economy � nancial systems tend to be much should states do to support robust � nancial FIGURE O.1 Benchmarking Financial Development, 2008–10 a. Financial institutions b. Financial markets Depth Depth 100 100 75 75 50 50 25 25 0 0 Stability Access Stability Access Developed economies (%) Efficiency Developing economies (%) Efficiency Source: Calculations based on Cˇ ihák, Demirgüç-Kunt, Feyen, and Levine 2012. Note: Average values are shown for 2008–10 with simple (unweighted) averages across country groups. The 0 corresponds to a historical low of the proxy variable, and 100 corresponds to a historical high calculated for all countries over the period 1960–2010. For the explanation of individual proxy variables for �nancial depth, access, stability, and ef�ciency, see chapter 1. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 OVERVIEW 7 BOX O.3 Navigating This Report In addition to this Overview, the report has two Chapter 4 examines direct state interventions, main parts. The first part (chapter 1) introduces particularly the experience with state-owned banks measures of different characteristics of � nancial sys- during the � nancial crisis. It reviews existing and tems that are useful in benchmarking � nancial sys- new research and reexamines the performance of tems around the world. The second part (chapters 2 state-owned banks during crises. A large part of through 5) examines various aspects of the state’s the discussion focuses on state-owned commer- role in � nance. cial banks as opposed to state-owned development Chapter 1 describes � nancial depth, access, ef�- banks; nonetheless, the chapter also presents a new ciency, and stability across countries and regions, data set based on a recent survey of development especially in developing economies. Chapter 1 intro- banks. It also examines the role of credit guarantees. duces a major new database, the Global Financial Chapter 5 relates to the role of the state in � nan- Development Database, and discusses how subse- cial infrastructure, with a focus on two topics high- quent editions of the report will revisit the analysis lighted by the crisis: (a) information sharing in credit and benchmarking of � nancial systems with updated markets, and (b) the role of the state in reducing and expanded data. counterparty risk in payments and securities settle- Chapter 2 examines the role of the state as reg- ment systems. ulator and supervisor. It presents results from a The accompanying website (http://www.world recently updated and substantially expanded World bank.org/� nancialdevelopment) contains a wealth Bank survey of regulation and supervision around of underlying research, additional evidence includ- the world, explores how crisis countries were differ- ing country examples, and an extensive database on ent from noncrisis countries, and tracks changes that � nancial development, providing users with interac- governments made after the crisis. The chapter also tive access to information on � nancial systems. The reviews international regulatory and supervisory website is also a place where users participate in an reforms and discusses proposals for further reforms. online version of the Financial Development Barom- Chapter 3 focuses on the role of the state in com- eter, provide feedback on this Global Financial petition policy. After discussing various measures of Development Report, and submit their suggestions competition, and presenting trends across countries for future issues of the report. and over time based on a new worldwide data set, it The report concentrates on banks. There are reviews the evidence on the implications of banking some references to and data on financial markets competition for bank ef� ciency, access to � nance, and nonbank � nancial institutions (for example, in a and � nancial stability. The chapter then analyzes the discussion on the regulatory perimeter and on access policy drivers of competition and highlights the role by nonbank institutions to � nancial infrastructure). of the state in (a) promoting a contestable banking But to keep the report focused, much of the discus- system and (b) enabling a market-friendly informa- sion is devoted to banks. Future issues of the report tional and institutional environment. It also ana- will cover � nancial markets and nonbank � nancial lyzes the impact of government actions during crises institutions in more depth. on bank competition. infrastructure? Box O.3 provides an over- factors, including a country’s level of devel- view of the report’s chapters. opment and the government’s capacity. Two How should public policy be designed themes emerge throughout this report. to address these four key questions? The The first relates to direct and indirect issue of concern in this report is how best interventions. During the recent crisis, direct to balance the various roles of the state as state interventions have increased, and early promoter, owner, regulator, and overseer. evidence reveals that some of these inter- The right balance depends on a number of ventions worked, at least in the short run. 8 OVERVIEW GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 However, there is also evidence on potential Overall, there is broad agreement to longer-term negative effects. Therefore, as the address the “basics� � rst. This means hav- crisis subsides, there may be a need to rebal- ing in place a coherent institutional and legal ance toward less direct state involvement. framework that establishes market discipline The second important theme is the criti- complemented by strong, timely, and antici- cal role that incentives play in the � nancial patory supervisory action. In many develop- sector. The challenge for the state’s involve- ing economies, this also means that building ment is to better align private incentives with up supervisory capacity needs to be a top public interest, without taxing or subsidizing priority. Among the important lessons of private risk taking. The design of public pol- the global � nancial crisis are renewed focus icy needs to strike the right balance in order on systemic risk and the need to pay greater to promote sustainable development. This attention to incentives in the design of regula- leads to different challenges and trade-offs in tion and supervision. answering each of the four questions below. Using a new survey of regulation and supervision around the world (figure O.2), chapter 2 confirms that countries where What are the best ways to reform the global financial crisis originated had regulation and supervision? weaker regulation and supervisory practices The global financial crisis that intensified (for example, less stringent definitions of with the collapse of Lehman Brothers in Sep- capital, less stringent provisioning require- tember 2008 presented a major test of the ments, and greater reliance on banks’ own international architecture developed over risk assessment), as well as less scope for many years to safeguard the stability of the market incentives (for example, lower qual- global �nancial system. Although the causes ity of financial information made publicly of the crisis are still being debated, there is available, more generous deposit insurance agreement that the crisis revealed major coverage). Tracking changes during the cri- shortcomings in market discipline, regula- sis reveals that countries have stepped up tion, and supervision. The financial crisis efforts in the area of macroprudential pol- therefore has reopened important policy icy, as well as on issues such as resolution debates on financial regulation. After the regimes and consumer protection. However, onset of the meltdown, there was much talk it is not clear whether incentives for market about not wasting the crisis, and using it to discipline have improved. Some elements of push through necessary reforms. Indeed, disclosure and quality of information have many reforms have been enacted or are in improved, but deposit insurance coverage has process. Much has been done, but the system increased during the crisis. This increased was tested further by the more recent euro coverage, together with generous support for area crisis, leading to the questions: Are the weak banks, did not improve incentives for reforms adequate and will they be suf�cient monitoring. The survey suggests that there is to reduce the likelihood and severity of future further scope for improving disclosures and �nancial crises? monitoring incentives. Regulation and supervision represent one Despite the progress made on regulatory area in which the role of the state is not in reform, there are still important areas of dis- dispute. The crucial role of the state is widely agreement. Hence, chapter 2 also presents acknowledged and is well established in the a number of reform proposals that call for economic and �nancial literature. Hence, the greater emphasis on simplicity and transpar- debate is not about whether the state should ency, as well as a focus on incentive-compat- regulate and supervise the � nancial sector, ible regulations. Importantly, these proposals but about how best to go about ensuring that warn against growing complexity of regula- regulation and supervision support sound tion, which may reduce transparency and �nancial development. accountability, increase regulatory arbitrage GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 OVERVIEW 9 FIGURE O.2 Selected Features That Distinguish Crisis-Hit Countries Broader capital definition (Is Tier 3 allowed in regulatory capital?) More sophisticated modeling (Is an advanced internal ratings-based approach offered to banks?) Less strict provisioning I (Are minimum levels of specific provisions for loans and advances set by the regulator?) Less strict provisioning II (Is there a regulatory requirement for general provisions on loans and advances?) Less oversight of external auditors (Are external auditors subject to independent oversight by the supervisor?) Lower standards for public data quality (Do laws or regulations require auditors to conduct their audits in accordance with international standards?) 0 20 40 60 80 100 Crisis Non-Crisis ˇ ihák, Demirgüç-Kunt, Martínez Pería, and Mohseni 2012. Source: C Note: Percentage of countries that responded “yes� to the question in parentheses. Based on the World Bank’s 2011 Bank Regulation and Supervision Survey. “Crisis� countries are de�ned as those that had a banking crisis between 2007 and 2011, as identi�ed in Laeven and Valencia (2012). opportunities, and signi�cantly strain regu- other risk-mitigating features. However, if latory resources and capacity. The propos- the state does not have the capacity to moni- als suggest a regulatory approach that is tor and police such complex rules, the likely more focused on proactively identifying and result is more speeding and more crashes. addressing incentive problems and making Similarly, complex approaches to calculat- regulations incentive-compatible. This can ing capital requirements are not appropriate help to end the continuous need to elimi- if there is limited capacity to verify the cal- nate de�ciencies and close loopholes that are culations, do robustness checks, and police inevitably present in ever more complex sets implementation. of regulations. Other proposals address the One of the positive developments triggered incentives that the regulators face and either by the crisis is much greater debate and com- propose alternative institutional structures or munication among regulators, policy mak- suggest tools to identify incentive issues on an ers, and academics, who are striving to reach ongoing basis. the common goal of designing regulations to In implementing supervisory best prac- minimize the occurrence and cost of future tices, emerging markets and developing econ- crises. The diverse views and multiple reform omies should focus on establishing a basic proposals in this debate (presented in chapter robust supervisory framework that reflects 2) are likely to inform the regulatory reform local � nancial systems’ characteristics, and process and improve future outcomes. refraining from incorporating unnecessary (and in several cases inapplicable) complex How should the state promote elements. Referring back to the earlier anal- competition in the �nancial sector? ogy with speed limits for cars and trucks, it may be appealing to have a complex rule The global � nancial crisis also reignited the in which each car has its own speed limit, interest of policy makers and academics in depending on the quality of its brakes and the impact of bank competition and the role 10 OVERVIEW GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 of the state in shaping competition policies. bank regulatory agencies: survey data reveal Some believe that increasing � nancial inno- that the majority now have explicit responsi- vation and competition in certain markets, bilities in the areas of competition policy. such as subprime mortgage lending, con- The Global Financial Development tributed to the global � nancial turmoil, and Report ’s analysis (chapter 3) provides guid- they are calling for policies to restrict com- ance on this important issue. Research sug- petition. Others worry that, as a result of gests that bank competition brings about the crisis and the actions of governments in improvements in ef�ciency across banks and support of the largest banks, concentration in enhances access to �nancial services, without banking increased, reducing the competitive- necessarily undermining systemic stability. ness of the sector and access to � nance, and A cursory look at trends in average systemic potentially also contributing to future insta- risk and bank market power (figure O.3) bility as a result of moral hazard problems indicates that greater market power (that is, associated with “too big to fail� institutions. less competition) is associated with more sys- Hence, the design of competition policy is temic risk (chapter 3 examines this in more challenging because it again involves a pos- detail). Hence, the evidence of a real trade-off sible trade-off between ef�ciency and growth is weak at best. on one hand and stability concerns on the This analysis suggests that policies to other hand. Another reason why rethinking address the causes of the recent crisis should competition policies is important relates to not unduly restrict competition. The appro- the changing mandate of central banks and priate public policy is (a) to establish a regu- latory framework that does not subsidize risk taking through poorly designed exit poli- cies and too-big-to-fail subsidies and (b) to FIGURE O.3 Market Power and Systemic Risk remove barriers to entry of “�t and proper� bankers with well-capitalized financial R-squared, log (systemic risk) Lerner index (market power) institutions. –0.5 For competition to improve access to 0.24 finance, the state has an important role to –0.6 play in enabling a market-friendly informa- –0.7 0.22 tional and institutional environment. Policies –0.8 0.20 that guarantee market contestability, timely –0.9 flow of adequate credit information, and –1.0 0.18 contract enforceability will enhance compe- tition among banks and improve access. For –1.1 0.16 instance, evidence across business line data in –1.2 0.14 Brazil shows that competition in the corpo- –1.3 rate segment is higher than in the retail seg- –1.4 0.12 ment. This reflects the existence of a larger pool of credit providers and easier access to –1.5 0.10 2002 2003 2004 2005 2006 2007 2008 information for large corporations. Com- Systemic risk (R-squared, log, left axis) petition in the retail sector can be fostered Market power (Lerner index, right axis) by promoting portability of bank accounts, expanding credit information sharing, and Source: Calculations based on Anginer, Demirgüç-Kunt, and Zhu 2012. increasing payment system interconnection. Note: The systemic risk measure follows Anginer and Demirgüç-Kunt (2001) and builds on In this context, consumer protection laws Merton’s (1974) contingent claim pricing. Systemic risk is de�ned as the correlation in the risk- taking behavior of banks and is captured by the R-squared from a regression of a bank’s weekly have been at the forefront of competition change in distance to default on country average weekly change in distance to default (excluding policies in many countries. One example the bank itself). Higher R-squared means higher systemic risk. Lerner index is a proxy for pro�ts that accrue to a bank as a result of its pricing power, so higher values mean less competition. is South Africa, where new legislation pro- The calculations cover 1,872 publicly traded banks in 63 economies (developed and developing). vided a framework to bolster competition by GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 OVERVIEW 11 providing a sound information environment often than not serving political interests to customers and protecting consumers from instead. Nevertheless, the global financial unfair credit and credit marketing practices. crisis underscored the potential countercycli- It established a National Credit Regulator to cal role of state-owned banks in offsetting act as a knowledge platform for credit prac- the contraction of credit from private banks, tices and to ensure compliance with the law. leading to arguments that this is an impor- Competition agencies also play a crucial tant function that can perhaps better justify advocacy role in promoting competition. their existence. One example in this regard is Romania’s The crisis and the actions adopted by Competition Council, which has extended different countries reignited the debate on the European Union Consumer Credit Direc- the need for direct government intervention tive of 2008. The directive establishes com- in the � nancial sector. Supporters of state- mon rules on consumer credit over mort- owned banks argue that they provide the gage or real estate guaranteed loans and state an additional tool for crisis manage- eliminates (or sets a low threshold for) early ment and, relative to central banks, may be repayment fees. more capable of providing a safe haven for Finally, state interventions during crises retail and interbank deposits, creating a � re may constitute a barrier to exit that permits break in contagion, and stabilizing aggregate insolvent and inefficient banks to survive credit. On the other hand, those opposing and generate unhealthy competition. Gov- government bank ownership point out that ernments should be aware that their inter- agency problems and politically motivated ventions during crises may have potentially lending render state-owned banks inef�cient negative long-term consequences on bank and prone to cronyism. Furthermore, past competition and may distort risk-taking experiences of numerous countries suggest incentives. that cronyism in lending may build up large �scal liabilities and threaten public sector sol- vency and � nancial stability, as well as mis- When do direct government allocate resources and retard development in interventions help? the long run. During the global � nancial crisis, countries During the recent crisis, several coun- pursued a variety of strategies to restart tries used their public bank infrastructure their � nancial and real sectors. As the bal- to prop up the �nancial sector. For instance, ance sheets of private banks deteriorated and the Brazilian government injected capital they curtailed their lending activities, many into its state-owned development bank and countries used state-owned banks to step up authorized state-owned banks to acquire their financing to the private sector. Most equity stakes from private banks and loan countries relied heavily on the use of credit portfolios from financial institutions with guarantee programs. Others adopted a num- liquidity problems. In China, state-owned ber of unconventional monetary and fiscal banks were instructed to boost credit to measures to prop up credit markets. speci�c sectors in order to promote growth. Historically, many state-owned banks In the Russian Federation, Vnesheconom- were created to fulfill long-term develop- bank, the country’s state-owned develop- ment roles by filling market gaps in long- ment bank, received new capital to assist term credit, infrastructure, and agriculture troubled smaller financial institutions and finance, and to promote access to finance to invest in Russian � nancial instruments. to underserved segments of the economy— It also injected money into large state- notably, small and medium enterprises. In controlled banks to increase their loans to practice, however, there is widespread evi- Russian companies. In Mexico, state-owned dence that state banks have generally been development banks extended credit to large very inefficient in allocating credit, more companies, participated in loan programs 12 OVERVIEW GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 for fragile sectors, and extended guarantees FIGURE O.4 Change in Bank Lending Associated on commercial paper and credit instruments with a 1% Increase in GDP Per Capita Growth issued by specialized nonbank �nancial insti- tutions. Similar actions were also taken by Percent some developed economies. For example, 2.5 Procyclical lending Germany’s state-owned development bank, 0.021*** 0.020*** Kreditanstalt für Wiederaufbau, increased 2.0 lending to larger companies with short- term liquidity problems, provided additional 1.5 0.013** financing for infrastructure, and helped 0.010*** 1.0 recapitalize regional state banks. And in Finland, the government raised the limits on 0.006*** 0.5 domestic and export �nancing for the coun- try’s state-owned bank to boost lending to 0.0 small and medium enterprises. Countercyclical lending –0.004*** Chapter 4 highlights that not all state- –1.5 owned banks are alike. They can be classi�ed Full Developing Developed sample economies economies as state commercial banks, state development banks, and development financial institu- Private State tions, depending on whether they aim to maximize pro�ts, are deposit takers, or have Source: Bertay, Demirgüç-Kunt, and Huizinga 2012. Note: The �gure shows marginal effects from a regression of bank lend- a clear developmental mandate. State-owned ing on GDP per capita growth and a number of control variables, esti- development banks and �nancial institutions, mated using a sample of 1,633 banks from 111 countries for the period 1999–2010. in turn, can lend to the public either directly Signi�cance level: ** 5 percent, *** 1 percent. or indirectly through private banks. Most of the evidence discussed on the short-term and long-term effects of state-owned banks The mitigating short-term effect of state- focuses on commercial banks or does not dis- owned banks is illustrated in �gure O.4. The tinguish between commercial and develop- �gure shows the relationship between lending ment banks. patterns of banks with private and state own- Chapter 4 reviews the historical and new ership and economic growth, measured by research evidence and concludes that lending real GDP per capita growth. Globally, bank by state-owned banks tends to be less procy- lending is procyclical, growing during booms clical than that of their private counterparts. and falling during downturns. Yet the lend- During the global �nancial crisis, some state- ing pattern of private banks is more procycli- owned banks have indeed played a counter- cal compared with their state-owned counter- cyclical role by expanding their lending port- parts. In high-income countries, state-owned folio and restoring favorable conditions in banks even behave in a clearly countercycli- key markets. For instance, the chapter high- cal fashion, increasing in downturns. lights the expansion of the lending portfolio However, because in many cases lend- of state-owned commercial banks (for exam- ing growth continued even after economic ple, PKO Bank Polski in Poland) and state- recovery was under way, and loans were not owned development banks (for example, directed to the most constrained borrowers, BNDES in Brazil) in mitigating the effects the countercyclical bene�ts of state-owned from the global credit crunch and �lling the banks came at the cost of resource misal- gap of lower credit from the private sector. location and worsened intermediation. This Also, Mexican development banks supported mixed view is supported by evidence from the credit channel through the extension previous crises as well. In other words, a tem- of credit guarantees and lending to private porary boom in state bank lending has long- �nancial intermediaries. term adverse effects by creating a portfolio of GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 OVERVIEW 13 bad loans in crises that take a long time to stable systems for large-value �nancial trans- sort out. actions. Reflecting the focus on the aftermath Ideally, focusing on the governance of of the financial crisis, the report does not these institutions may help policy makers examine other components of �nancial infra- address the inefficiencies associated with structure, such as retail payment systems and state-owned banks. State banks need a clear collateral regimes; it leaves these important mandate to complement (rather than sub- issues to be covered in future editions. stitute for) private banks, and adopt risk Chapter 5 emphasizes that the transpar- management practices that allow them to ent exchange of credit information reduces guarantee a �nancially sustainable business. information asymmetries between borrowers However, these governance reforms are par- and lenders and is an essential requisite of a ticularly challenging in weak institutional well-functioning credit market. However, the environments, further emphasizing that the �nancial crisis has shown that there is much trade-off is a serious one for policy makers. room for improvement in this area, especially Credit guarantee schemes have also been in the use of existing credit reporting systems a popular intervention tool during the recent for prudential oversight and regulation. crisis. However, given their limited scale, they Information sharing in credit markets acts are used not to stabilize aggregate credit but as a public good that improves credit market to alleviate the impact of the credit crunch on efficiency, access to finance, and financial segments that are most severely affected, such stability. Nonetheless, for an individual com- as small and medium enterprises. Unfortu- mercial bank, proprietary credit information nately, rigorous evaluations of these schemes is valuable, so it has incentives to collect the are very few, and existing studies suggest information and keep it away from others. that the bene�ts of these programs tend to be Information sharing among private lend- rather modest, particularly in institutionally ers thus may not arise naturally, especially underdeveloped settings, and they tend to where banking systems are concentrated (�g- incur �scal and economic costs. Nevertheless, ure O.5). This creates an important rationale best practices can be identi�ed. These include for state involvement. In addition, the report leaving credit assessments and decision mak- highlights that information sharing in credit ing to the private sector; capping coverage markets has increasing returns to scale: the ratios and delaying the payout of the guar- bene�ts of credit reporting for �nancial access antee until recovery actions are taken by the and stability are greatest when participation lender, so as to minimize moral hazard prob- is as wide as possible and includes banks as lems; having pricing guarantees that take into well as nonbank �nancial institutions. There- account the need for � nancial sustainability fore, another important role for the state is to and risk minimization; and encouraging the create a level playing �eld for the provision use of risk management tools. Success again and exchange of credit information, and to hinges on overcoming the challenges of get- facilitate the inclusion of nonregulated lend- ting the design right, particularly in underde- ers into existing credit reporting systems. In veloped institutional and legal settings. many emerging markets, such as China and South Africa, major initiatives are under way to integrate the rapidly growing micro�nance What is the role for the state in and consumer loan markets into the existing promoting �nancial infrastructure? credit reporting infrastructure. The global �nancial crisis has highlighted the Liquidity provision by central banks dur- importance of a resilient �nancial infrastruc- ing the crisis helped prevent major payment ture for � nancial stability. It also has led to system disruptions. However, stress emerged a discussion about the role of the state, par- in interbank and over-the-counter derivatives ticularly in promoting the provision of high- markets. The state can play an important role quality credit information and in ensuring in mitigating counterparty risks in interbank 14 OVERVIEW GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE O.5 Credit Reporting vs. Banking System money markets by providing robust and Concentration secure infrastructure and, potentially, by promoting the development of collateralized Probability of existence of a credit reporting system interbank markets. The state can also con- 1.0 tribute in the development of a robust infra- 0.92 0.9 structure for security settlement systems and 0.80 the oversight of securities transactions, par- 0.8 ticularly for over-the-counter transactions. 0.7 Increased standardization and transparency 0.6 0.56 0.53 of transactions is needed and can be achieved 0.5 by (a) trading on exchanges or electronic 0.4 0.37 0.39 trading platforms; (b) clearing transactions through central counterparties, that is, enti- 0.3 ties that interpose themselves as counterpart 0.2 to each trade (examples include the Chicago 0.1 Mercantile Exchange’s CME Clearing in the 0.0 United States, Eurex Clearing in Germany, Credit Credit Any credit and London Clearing House’s LCH.Clear- registry bureau reporting net in the United Kingdom); and (c) report- Low bank concentration High bank concentration ing transactions to trade repositories, which are entities that store centralized records of Source: Bruhn, Farazi, and Kanz 2012. transaction data. These policy prescriptions Note: The �gure reports the percentage of countries with private (credit bureau), public (credit registry), or any credit reporting institutions for are especially important in many emerging countries with high and low degrees of bank concentration (above and markets, where the development of a modern below the sample mean), respectively. It shows that bank concentration (the asset share of a country’s three largest banks) is negatively asso- settlement infrastructure has lagged behind ciated with the development of credit reporting. This relationship is also the rapid growth of emerging equity and conditional on the level of economic development. securities markets. 1 Benchmarking Financial Systems around the World • Financial systems are multidimensional. Four characteristics are of particular interest for benchmarking �nancial systems: �nancial depth, access, ef�ciency, and stability. These characteristics need to be measured for �nancial institutions and markets. • Financial systems come in all shapes and sizes, and differ widely in terms of the four characteristics. As economies develop, services provided by �nancial markets tend to become more important than those provided by banks. • The global �nancial crisis was not only about �nancial instability. In some economies, the crisis was associated with important changes in �nancial depth and access. A growing body of evidence suggests that � nancial institutions—such as banks and insurance companies—and �nan- managerial performance, this boosts the ef�- ciency of corporations and reduces waste and fraud by corporate insiders. But that is not all. cial markets—stock markets, bond markets, When equity, bond, and derivative markets derivative markets, and so on—exert a pow- enable the diversi�cation of risk, this encour- erful influence on economic development, ages investment in higher-return projects poverty alleviation, and economic stability that might otherwise be shunned. And, when (Levine 2005). For example, when banks � nancial systems lower transaction costs, screen borrowers and identify �rms with the this facilitates trade and specialization—fun- most promising prospects, this is a key step damental inputs to technological innovation that helps allocate resources, expand economic (Smith 1776). opportunities, and foster growth. When banks When financial systems perform these and securities markets mobilize savings from functions poorly, they hinder economic households to invest in promising projects, growth, curtail economic opportunities, and this is another crucial step in fostering eco- destabilize economies. For example, if �nan- nomic development. When �nancial institu- cial systems collect funds and pass them along tions monitor their investments and scrutinize to cronies, the wealthy, and the politically GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 15 16 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 connected, it slows economic growth and across countries and regions and through blocks potential entrepreneurs. And if �nan- time; and how � nancial systems have been cial institutions fail to exert sound corporate affected by the global �nancial crisis. governance over firms that they fund, that To benchmark financial systems, the failure makes it easier for managers to pursue report measures the following four charac- projects that bene�t themselves rather than teristics of financial institutions and mar- the � rms and the economy. When � nancial kets: (a) the size of �nancial institutions and institutions create complex � nancial instru- markets (� nancial depth), (b) the degree to ments and sell them to unsophisticated inves- which individuals can and do use � nancial tors, it might generate more income for �nan- institutions and markets (access), (c) the ef�- cial engineers and executives associated with ciency of � nancial institutions and markets marketing the new instruments, distorting in providing financial services (efficiency), the allocation of society’s savings and imped- and (d) the stability of � nancial institutions ing economic prosperity. and markets (stability). These characteristics Evidence on the � nancial system’s role in are measured separately for �nancial institu- shaping economic development is substantial tions and �nancial markets (both equity and and varied. But there are shortcomings asso- bond markets), leading to a 4x2 matrix of ciated with assessing �nancial systems. There � nancial system characteristics. The report are no good cross-country, cross-time mea- uses these measures to characterize and sures of how they (a) enhance information compare � nancial systems across economies about �rms and hence the ef�ciency of resource and over time and to assess the relationships allocation; (b) exert sound corporate gover- between these measures and �nancial sector nance over �rms to which they channel those policies. resources; (c) manage, pool, and diversify risk; In focusing on these four characteristics of (d) mobilize savings from savers so that these �nancial institutions and markets, the report resources can be allocated to the most prom- gives empirical shape and substance to the ising projects in the economy; and (e) facili- complex, multifaceted, and sometimes amor- tate trade. Instead, researchers have largely phous concept of the functioning of �nancial focused on the size of the banking industry as systems. Financial depth, access, ef�ciency, a proxy for �nancial development. But size is and stability might not capture all features of not a measure of quality, ef�ciency, or stabil- �nancial systems, and the report does not try ity. And the banking sector is only one part of to construct a composite index of � nancial �nancial systems. development. Instead, it uses these four char- Accordingly, a key contribution of this acteristics to describe, compare, and analyze chapter involves data. In recent years, substan- �nancial systems and their evolution in recent tial efforts have been made to improve these decades. data, which this chapter uses. This report is This chapter, together with the underly- accompanied by the new Global Financial ing data and analysis, highlights the multi- Development Database, an extensive world- dimensional nature of financial systems. wide database that combines and updates Deep financial systems do not necessarily several �nancial data sets (Čihák, Demirgüç- provide broad financial access, highly effi- Kunt, Feyen, and Levine 2012). The data- cient financial systems are not necessarily base is available on the Global Financial more stable than the less ef�cient ones, and Development Report Web page (http://www so on. Each of these characteristics is asso- .worldbank.org/�nancialdevelopment). ciated with socioeconomic development, But this chapter goes beyond compiling � nancial sector policies, and other parts of data. It answers some substantive questions the enabling environment for �nance. Finan- using the data, such as how to empirically cial systems differ widely in terms of the 4x2 describe different characteristics of �nancial characteristics, so it is crucial to measure and systems; how to compare � nancial systems evaluate each one. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 17 The chapter also suggests that the global the background, contributing to economic �nancial crisis resulted in more than �nancial growth and poverty reduction. But when instability: in some countries, it also caused things go wrong, the malfunctioning of the problems along the other dimensions, such as financial system can slow growth, throw making people’s and �rms’ access to �nancial more people into poverty, and destabilize services more dif�cult. Finance is about more entire economies. Indeed, financial crises than just stability. Having � nancial systems hurt not only those who work in � nance or channel society’s savings to those with the those who access financial systems. When most promising investment opportunities is the government undertakes costly bailouts essential for fostering economic growth, alle- of bankrupt � nancial institutions, this can viating poverty, and enabling people to pur- lead to increases in public indebtedness, thus sue their economic goals. undermining governments’ ability to support Finally, this chapter is linked to future key social objectives, including the fund- editions of the Global Financial Develop- ing of education, health, and infrastructure ment Report. The report is envisaged as programs. As a result, malfunctioning �nan- part of a series, with future reports return- cial systems can also lay the foundations for ing to the analysis of �nancial systems using enduring economic crises, as illustrated quite updated and extended data. They will use the dramatically by recent events. measurement framework introduced here to With so much attention focused on sta- examine new topics, such as � nancial inclu- bility issues following the recent crisis, the sion, capital market development, and oth- powerful linkages between the functioning ers. Future editions might expand or improve of the � nancial system and economic devel- on the framework, which is designed to be opment have been somewhat underempha- flexible to accommodate such adjustments if sized. Although the focus on stability has needed—for example, if new types of �nan- been understandable, sound � nancial sector cial data become available. policies are not only about avoiding crises. Finance is also about the ef�cient allocation of capital, economic growth, and expanding THE IMPORTANCE OF economic horizons. Therefore, an impor- FINANCIAL SYSTEMS tant goal is to raise awareness of policies to TO DEVELOPMENT enhance the operation of � nancial systems, Finance is central to development. This develop a better understanding, and foster may seem obvious to � nancial development debate. To help in framing the debate, this experts. It may also seem obvious to bank section clarifies the definition of financial depositors who just had their entire life sav- development and provides a review of the ings wiped out by a �nancial crisis. But �nan- literature on the linkages between � nancial cial crises get forgotten after a period of time. sector development, economic growth, and And when compared with other factors that poverty reduction. are also important—health, the environ- ment, and so on—the case for � nance may What is �nancial development? appear less obvious. Indeed, when panels of the world’s leading economists tried to iden- Financial markets are imperfect. Acquiring tify “the 10 great global challenges� in both and processing information about poten- 2004 and 2008 as part of the Copenhagen tial investments is costly. There are costs Consensus Project, the list did not include and uncertainties associated with writing, any �nancial issues.1 interpreting, and enforcing contracts. And This section argues that finance indeed there are costs associated with transacting matters. It matters both when it functions goods, services, and � nancial instruments. well and when it malfunctions. When oper- These market imperfections inhibit the flow ating effectively, finance works quietly in of society’s savings to those with the best 18 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 ideas and projects, thus curtailing economic At a broader level, � nancial development development. can be de�ned as improvements in the quality It is the existence of these costs—these of �ve key �nancial functions: (a) producing market imperfections—that creates incen- and processing information about possible tives for the emergence of �nancial contracts, investments and allocating capital based on markets, and intermediaries. Motivated by these assessments; (b) monitoring individuals pro�ts, people create � nancial products and and �rms and exerting corporate governance institutions to ameliorate the effects of these after allocating capital; (c) facilitating the market imperfections. And governments trading, diversi�cation, and management of often provide an array of services—rang- risk; (d) mobilizing and pooling savings; and ing from legal and accounting systems to (e) easing the exchange of goods, services, government-owned banks—with the stated and financial instruments. Financial insti- goals of reducing these imperfections and tutions and markets around the world dif- enhancing resource allocation. Some econo- fer markedly in how well they provide these mies are comparatively successful at develop- key services. Although this report sometimes ing �nancial systems that reduce these costs. focuses on the role of the �nancial systems in Other economies are considerably less suc- reducing information, contracting, and trans- cessful, with potentially large effects on eco- action costs, it primarily adopts a broader nomic development. view of �nance and stresses the key functions At the most basic level, therefore, � nan- provided by the �nancial system to the over- cial development occurs when financial all economy. instruments, markets, and intermediaries mitigate—though do not necessarily elimi- Financial development and nate—the effects of imperfect information, economic growth limited enforcement, and transaction costs. For example, the creation of credit registries Economists have long debated the finan- tends to improve acquisition and dissemina- cial sector’s role in economic growth. Lucas tion of information about potential borrow- (1988), for example, dismissed finance as ers, improving the allocation of resources an overstressed determinant of economic with positive effects on economic develop- growth, and Robinson (1952, 86) quipped ment. As another example, countries with that “where enterprise leads �nance follows.� effective legal and regulatory systems have From this perspective, finance responds to facilitated the development of equity and demands from the nonfinancial sector: it bond markets that allow investors to hold does not cause economic growth. At the more diversi�ed portfolios than they could other extreme, Miller (1998, 14) argued that without efficient securities markets. This the idea that �nancial markets contribute to greater risk diversification can facilitate economic growth “is a proposition too obvi- the flow of capital to higher return proj- ous for serious discussion.� Bagehot (1873) ects, boosting growth and enhancing living and others rejected the idea that the �nance- standards. growth nexus can be ignored without limit- De� ning � nancial development in terms ing understanding of economic growth. of the degree to which the � nancial system Recent literature reviews (such as Levine eases market imperfections, however, is too 2005) conclude that evidence suggests a posi- narrow and does not provide much infor- tive, � rst-order relationship between � nan- mation on the actual functions provided by cial development and economic growth. In the � nancial system to the overall economy. other words, well-functioning � nancial sys- Thus, Levine (2005) and others have devel- tems play an independent role in promoting oped broader de� nitions that focus on what long-run economic growth: countries with the �nancial system actually does.2 better-developed financial systems tend to GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 19 grow faster over long periods of time, and business and who cannot, who can pay for a large body of evidence suggests that this education and who cannot, who can attempt effect is causal (Demirgüç-Kunt and Levine to realize his or her economic aspirations and 2008).3 who cannot. Furthermore, by affecting the Moreover, research sheds light on the allocation of capital, � nance can alter both mechanisms through which � nance affects the rate of economic growth and the demand growth. The financial system influences for labor, with potentially profound implica- growth primarily by affecting the alloca- tions for poverty and income distribution. tion of society’s savings, not by affecting the Potentially, � nance can have rather com- aggregate savings rate. Thus, when �nancial plex effects on the income distribution. It systems do a good job of identifying and could boost returns to high-skilled work- funding those � rms with the best prospects, ers or to low-skilled workers. The mecha- not those firms simply with the strongest nisms are complex and could be good or bad political connections, this improves the capi- for the poor and reduce or increase income tal allocation and fosters economic growth. inequality. Such � nancial systems promote the entry of There is an emerging body of empirical new, promising firms and force the exit of research, however, suggesting that in prac- less ef�cient enterprises. Such � nancial sys- tice, improvements in financial contracts, tems also expand economic opportunities, markets, and intermediaries actually do so that the allocation of credit—and hence tend to expand economic opportunities and opportunity—is less closely tied to accumu- reduce persistent income inequality. Figure lated wealth and more closely connected to 1.1 provides a basic empirical illustration the social value of the project. Furthermore, of the link between financial development by improving the governance of � rms, well- (approximated here in a simpli�ed way by the functioning financial markets and institu- ratio of private sector credit to gross domes- tions reduce waste and fraud, boosting the tic product) and income inequality (approxi- ef�cient use of scarce resources. By facilitat- mated by changes in the Gini coef�cient). The ing risk management, � nancial systems can graph illustrates that higher levels of �nancial ease the �nancing of higher return endeavors development are associated with declines in with positive reverberations on living stan- inequality. dards. And, by pooling society’s savings, More in-depth empirical research is con- � nancial systems make it possible to exploit sistent with this basic observation. For exam- economies of scale—getting the biggest ple, evidence suggests that access to credit development boost from available resources. markets increases parental investment in the education of their children and reduces the substitution of children out of schooling Financial development and and into labor markets when adverse shocks poverty reduction reduce family income (Belley and Lochner Beyond long-run growth, � nance can also 2007). Better-functioning � nancial systems shape the gap between the rich and the poor stimulate new �rm formation and help small, and the degree to which that gap persists promising � rms expand as a wider array of across generations (Demirgüç-Kunt and firms gain access to the financial system. Levine 2009). Financial development may Moreover, better-functioning � nancial sys- affect to what extent a person’s economic tems will identify and fund better projects, opportunities are determined by individual with less emphasis on collateral and incum- skill and initiative, or whether parental bency. Not only do they allow new, ef�cient wealth, social status, and political connec- �rms to enter, they also force old, inef�cient tions largely shape economic horizons. The �rms to leave, as evidenced by data (Kerr and � nancial system influences who can start a Nanda 2009). 20 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 1.1 Financial Depth and Income Inequality and accounting systems influence the costs associated with evaluating � rms and writ- ing and enforcing contracts and, hence, in Change in Gini coefficient identifying and �nancing an economy’s most 0.03 UGA promising endeavors. Regulatory, supervi- sory, and tax systems all affect the incentives NGA facing the executives of �nancial institutions 0.02 and participants in securities markets. Thus, ZMB these components of the enabling environ- ROM ECU GTM CHL 0.01 GHA NER LSO HKG ment for finance also shape the allocation ARG USA and use of capital. And the state often plays SLE a more direct role in shaping the operation 0 IND of financial systems, running state-owned NPL JPN banks, subsidizing agriculture or housing, TZA ITA NOR –0.01 ETH TUR IRL SWE CHE or issuing government securities. Thus, the FIN NLD entire legal, accounting, regulatory, and SEN FRA AUT policy apparatus influences the operation of –0.02 MUS �nancial systems. Given the importance of � nance for eco- –0.03 nomic development and poverty alleviation, –3 –2 –1 0 1 2 it is natural to ask: Why does this chapter Private credit to GDP focus on measuring the functioning of the � nancial system rather than on examining Source: Update of Beck, Demirgüç-Kunt, and Levine 2007. the direct impact of � nancial sector policy, Note: The Gini coef�cient is on a scale from 0 (total equality) to 1 (maximum inequality). The chart regulations, and the rest of the enabling is a partial scatter plot, visually representing the regression of changes in the Gini coef�cient between 1960 and 2005 on the private sector credit–to-GDP ratio (logarithm, 1960–2005 aver- environment on economic growth, poverty age), controlling for the initial (1960) Gini coef�cient. Variables on both axes are residuals. The alleviation, and the availability of economic abbreviations next to some of the observations are the three-letter country codes as de�ned by the International Organization for Standardization. opportunities? The answer is that to provide guidance to Besides the direct benefits of enhanced policy makers, one needs a detailed under- access to financial services, finance also standing of the mechanisms through which reduces inequality, particularly through the enabling environment for � nance influ- indirect labor market mechanisms. Spe- ence the functioning of �nancial systems. It is ci�cally, accumulating evidence shows that not enough to assess the associations between � nancial development accelerates economic financial sector policies and development growth, intensi�es competition, and boosts outcomes because these correlations might the demand for labor. Importantly, it usually reflect reverse causality—in which economic brings relatively bigger bene�ts to those at the development shapes the types of financial lower end of the income distribution (Beck, sector policies that a country adopts—or the Demirgüç-Kunt, and Levine 2007; Beck, correlations might simply reflect the impact Levine, and Levkov 2010). Hence, � nance, of some other factor on both economic devel- with good policies, can be both pro-growth opment and �nancial sector policies. To pro- and pro–poverty reduction. vide more accurate assessments about the enabling environment for � nance, it is vital to trace through the channels from particular Financial development and the enabling policies and regulations to the operation of environment for �nance � nancial systems and on to particular eco- Many factors shape the functioning of �nan- nomic development outcomes. cial systems and hence their impact on eco- This report contributes to this goal of pro- nomic growth and poverty alleviation. Legal viding more sound advice to policy makers by GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 21 FIGURE 1.2 Socioeconomic Development, Financial Development, and Enabling Environment Socioeconomic Development Social welfare (sustainable long-term growth, poverty reduction) Financial Development Financial sector functions Financial development outcomes Producing information about investments and (empirical proxies, measured separately for allocate capital; monitoring investments and �nancial institutions and markets) exerting corporate governance; managing risks; – Depth pooling savings; and easing the exchange of – Access goods and services – Ef�ciency – Stability Enabling Environment Financial sector policies (examples) Other policies and features (examples) – Regulation (micro- and macro-prudential, – Macroeconomic policy framework (e.g., business conduct, etc.) exchange rate regime, monetary policy, – Direct interventions (state ownership, tax policy, capital controls) guarantees, subsidies, liquidity provision) – Legal framework, social capital, etc. – Competition policy in �nance (level playing – Concentration in the system �eld, entry/exit, etc.) – Internationalization, dollarization – Promotion of �nancial infrastructure/ technology ˇ ihák, Demirgüç-Kunt, Feyen, and Levine 2012. Source: Based on the review of literature in C (a) developing and analyzing measures of the THE GLOBAL FINANCIAL functioning of �nancial institutions and mar- DEVELOPMENT DATABASE kets (chapter 1) and (b) assembling databases AND THE 4X2 MEASUREMENT on regulations, supervision, and institutional MATRIX structures that shape �nancial system opera- tions (chapters 2 to 5). Introducing the Global Financial To summarize the discussion in this sec- Development Database tion, �gure 1.2 presents in a visual form the relationships between socioeconomic devel- To measure the functioning of � nancial sys- opment, financial development, and the tems, country of�cials, researchers, and oth- enabling environment. It is important to care ers would ideally like to have direct measures about the process of � nancial development of how well �nancial institutions and �nan- because it has a well-documented association cial markets (a) produce information ex ante with economic and social development more about possible investments and allocate capi- generally. It improves sustainable long-term tal; (b) monitor investments and exert cor- growth and reduces poverty, thereby improv- porate governance after providing finance; ing social welfare. One can think about these (c) facilitate the trading, diversi�cation, and as the ultimate developmental objectives. Fig- management of risk; (d) mobilize and pool ure 1.2 also highlights that �nancial systems savings; and (e) ease the exchange of goods do not exist in a vacuum. Financial system and services. So if data were not an issue, the characteristics depend on the enabling envi- ideal approach to measurement would involve ronment, which consists of financial sec- the following determinations: in terms of tor policies and other relevant policies and producing information about possible invest- features. ments and allocating capital, the financial 22 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 sector in Country A, for example, scores 60 �gure 1.2). For completeness, the accompa- on a scale from 0 to 100, while Country B’s nying database includes some variables that � nancial sector scores 75; in terms of moni- measure social welfare (the upper part of toring investments and exerting corporate �gure 1.2) as well as �nancial sector policies governance after providing �nance, Country and the other factors that de�ne the enabling A scores 90, while Country B scores only 20 environment for � nance (the bottom of �g- on a scale from 0 to 100, and so on. ure 1.2). The following subsections introduce So, instead of the direct measures, empiri- each dimension of this measurement frame- cal studies have focused on proxy variables, work. The annex to this chapter and C ˇ ihák, such as various measures of � nancial depth Demirgüç-Kunt, Feyen, and Levine (2012) and access. And despite evidence of the cru- provide more detailed information on each cial role of finance for economic develop- component of the measures of the four � nan- ment, there is a surprising lack of comprehen- cial system traits in the matrix. sive data on basic aspects of �nancial systems To obtain a comprehensive characteriza- across countries and over time. For example, tion of � nancial systems, one must measure there are major gaps in data on trading vol- the four categories for the two key compo- umes in securities markets. Even data on nents of the �nancial sector, namely �nancial � nancial institutions become rather patchy institutions (banks and nonbank financial when one looks beyond the world’s major, institutions) and financial markets (stock publicly listed banks. market, bond market, and other markets). Against this background, one of the key Therefore, to be comprehensive, one needs contributions of the Global Financial Devel- to assemble a 4x2 matrix: four characteris- opment Report is the launch of a new, com- tics for two components. Table 1.1 provides prehensive online database on � nancial sys- a summary representation of such a 4x2 tems—the Global Financial Development matrix, with examples of variables that can Database, which is made available online be used to �ll in each cell of the matrix. The together with the report. The database, same structure is used to organize the under- which will be updated on a regular basis, lying database. The following subsections go compiles and disseminates data on the char- through the individual characteristics in turn. acteristics of � nancial systems in 205 juris- Box 1.1 focuses on the selection of represen- dictions around world. The database has tative variables within the individual charac- data going back some 50 years (to 1960), teristics. Box 1.2 discusses the challenges of although some measures of �nancial system aggregating across the four dimensions. traits do not go back that far.4 The data from Critically, this chapter looks beyond the the Global Financial Development Database size of banks and stock markets. Many fac- are integrated with the World Bank’s Open tors shape the mixture of � nancial interme- Data initiative. Some of the data are new, and diaries and markets operating in an economy. this is the �rst time such comprehensive data Different types and combinations of infor- are available. The data are made available mation, enforcement, and transaction costs in a Web-friendly form, allowing the users in conjunction with different legal, regula- of the database to interact with the data, for tory, and tax systems have motivated distinct example, by creating their own country peer � nancial contracts, markets, and intermedi- groups and their own tables and charts. aries across countries and throughout his- tory. Thus, financial institutions and mar- kets can and do look very different across The 4x2 measurement framework countries and over time, but these structural This chapter develops and presents four mea- differences do not necessarily translate into sures of the characteristics of � nancial sys- differences in the quality of the services pro- tems: depth, access, ef�ciency, and stability. vided by the �nancial system to the economy. The focus here is on empirically character- To measure financial systems, this chapter izing financial systems (the middle part of digs deeper into the functioning of �nancial GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 23 TABLE 1.1 Stylized 4x2 Matrix of Financial System Characteristics (with examples of candidate variables in each category) Financial Institutions Financial Markets Private sector credit to GDP Stock market capitalization plus outstanding domestic Financial institutions’ assets to GDP private debt securities to GDP DEPTH Money (M2 aggregate) to GDP Private debt securities to GDP Deposits to GDP Public debt securities to GDP Value-added of the �nancial sector to GDP International debt securities to GDP Stock market capitalization to GDP Stocks traded to GDP Accounts per thousand adults (commercial banks) Percent of market capitalization outside of top Branches per 100,000 adults (commercial banks) 10 largest companies ACCESS Percent of people with a bank account (from user survey) Percent of value traded outside of top 10 traded companies Percent of �rms with line of credit (all �rms) Government bond yields (3 month and 10 year) Percent of �rms with line of credit (small �rms) Ratio of domestic to total debt securities Ratio of private to total debt securities (domestic) Ratio of new corporate bond issues to GDP Net interest margin Turnover ratio (turnover/capitalization) for stock market EFFICIENCY Lending-deposits spread Price synchronicity (co-movement) Noninterest income to total income Price impact Overhead costs (percent of total assets) Liquidity/transaction costs Pro�tability (return on assets, return on equity) Quoted bid-ask spread for government bonds Boone indicator (Her�ndahl, or H-statistic) Turnover of bonds (private, public) on securities exchange Settlement ef�ciency z-score (or distance to default) Volatility (standard deviation/average) of stock price index, Capital adequacy ratios sovereign bond index STABILITY Asset quality ratios Skewness of the index (stock price, sovereign bond) Liquidity ratios Price/earnings (P/E) ratio Other (net foreign exchange position to capital, etc.) Duration Ratio of short-term to total bonds (domestic, international) Correlation with major bond returns (German, United States) Source: Based on the review of literature in C ˇ ihák, Demirgüç-Kunt, Feyen, and Levine 2012. Note: This is a stylized matrix. For details, see C ˇ ihák, Demirgüç-Kunt, Feyen, and Levine (2012). Variables that are highlighted in bold are the ones suggested for the benchmarking exercise. Private sector credit to GDP is domestic private credit to the real sector times deposit money banks to GDP. Accounts per thousand adults (commercial banks) is the num- ber of depositors with commercial banks per 1,000 adults. For each type of institution, this �gure is calculated as the (reported number of depositors)*1,000/adult population in the reporting country. The net interest margin is the accounting value of the bank’s net interest revenue as a share of its average interest-bearing (total earning) assets. The z-score (or distance to default) is (ROA + equity)/assets)/sd(ROA), where ROA is average annual return on end-year assets and sd(ROA) is the standard deviation of ROA. Stock market capitalization plus outstanding domestic private debt securities to GDP is de�ned as the value of listed shares to GDP plus amount of outstanding domestic private debt securities to GDP. Percent of market capitalization outside of top 10 largest companies is the market capitalization out of the top 10 largest companies to total market capitalization. Turnover ratio (turnover/capitalization) for stock market is the ratio of the value of total shares traded to market capitalization. Volatility (standard deviation/average) of stock price index is the standard deviation of the sovereign bond index divided by the annual average of that index. systems and does not just look at the size of literature on � nancial development is private particular institutions and markets. credit, de� ned as credit to the private sector from deposit money banks, as a percentage of GDP. 5 There is a wide literature demon- First characteristic: Financial depth strating the link between financial depth, The most common way to characterize �nan- approximated by private sector credit to cial systems is by measuring the size of �nan- GDP, on one hand, and long-term economic cial institutions or markets relative to the growth and poverty reduction on the other size of the economy. “Financial depth� is an hand (for example, Demirgüç-Kunt and analytically incomplete, though empirically Levine 2008). Private credit varies widely ubiquitous, measure of the functioning of across countries. For example, averaged �nancial systems. over 1980–2010, private credit was less than For �nancial institutions, the variable that 10 percent of GDP in Angola, Cambodia, has received much attention in the empirical and the Republic of Yemen, while exceeding 24 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 1.1 Selecting the Representative Variables for Individual Characteristics For every category in the 4x2 matrix, several vari- How should one pick among such competing vari- ables could be used as proxies. Which combination ables? For the purpose of presenting the raw data of these variables should one choose when trying to in the database, it is not necessary to pick. Indeed, compare � nancial systems? the Global Financial Development Database shows In some cases, the variables in the same dimen- the competing variables, so that users can examine sion are complementary, and some are even additive. the data for themselves. However, for the purpose of For example, the total assets of banks to GDP and characterizing � nancial systems and for comparisons total assets of nonbank � nancial institutions to GDP across the dimensions, it is useful to pick one of the are in the same units and complement each other, so competing variables. they can be added up to obtain a proxy of total assets The general approach is to select indicators that of � nancial institutions to GDP. The result will be are widely available and have a clearly documented a good proxy variable, provided that the underlying link to long-term economic growth or poverty reduc- variables are comprehensive in their coverage and tion in the literature. When two variables capture the that no double counting occurs between them. Other same dimension, and both have a link to economic examples include measures of volatility in the stock development, one would select the variable that— market and volatility in the bond market. If these are even if it is perhaps less sophisticated—has greater measured in a similar way (as standard deviations), country coverage. The more sophisticated variable they can actually be added, using the capitaliza- is still included in the Global Financial Development tions of the two markets, as proxy for their relative Database, and relationships between some of these weights (as well as the covariance between the two), variables are explored in C ˇ ihák, Demirgüç-Kunt, to approximate the general volatility in the � nancial Feyen, and Levine (2012). For most of the variables, markets. the competing indicators tend to be highly (although In other cases, the variables “compete� to mea- not perfectly) correlated. For example, the correla- sure the similar things in slightly different ways. For tion coef�cient for private sector credit to GDP and example, private sector credit to GDP and total assets total banking assets to GDP is 0.98 (�gure 1.3). of � nancial institutions to GDP are both proxies for The chapter’s illustrative comparison of the 4x2 � nancial institutions’ size. The two variables differ characteristics across countries selects one variable in terms of their comprehensiveness and country cov- from each dimension. The selected variables are erage, with private sector credit to GDP covering a highlighted in bold in table 1.1. smaller set of assets but being available for a large number of countries. 85 percent of GDP in Austria, China, and in the literature on �nancial development. In the United Kingdom. The annual average any case, the two variables are rather closely value of private credit across countries was correlated, with a correlation coef�cient of 39 percent, with a standard deviation of 36 about 0.98 (�gure 1.3), so private credit can percent. provide a reasonably close approximation for An alternative to private credit is total total banking assets.6 banking assets to GDP, a variable that is also Despite the literature’s focus on banks, the included in the Global Financial Develop- global �nancial crisis has highlighted issues in ment Database. Compared to private credit, some nonbank �nancial institutions (NBFIs). this variable also includes credit to govern- Data coverage of NBFIs is less comprehensive ment and bank assets other than credit. It is than coverage of banks. Nonetheless, recog- arguably a more comprehensive measure of nizing the importance of NBFIs, the Global size, but it is available for a smaller number of Financial Development Database includes countries and has been used less extensively total assets of NBFIs to GDP, which includes GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 25 pension fund assets to GDP, mutual fund FIGURE 1.3 Correlations between Characteristics in Same Category assets to GDP, insurance company assets to (example) GDP, insurance premiums (life) to GDP, and insurance premiums (non-life) to GDP. Private credit versus domestic bank assets For � nancial markets, the two main seg- ments for which consistent worldwide data Domestic bank assets/GDP (%) can be collected are stock markets and bond 300 CYP markets (both sovereign and corporate). To approximate the size of stock markets, the ESP IRL most common choice in the literature is stock HKG NLD 200 PRT GBR market capitalization to GDP. For the size of JPN CHE LUX the bond markets, the mostly commonly used ITA MLT NZL LBN proxy for size is the outstanding volume of KNA FRA DEU GRC ISL SGPMYS LCA BEL VNM AUT AUS CHN debt securities (private and public) to GDP. 100 JOR MUS ATG GRD SVN THA ISR BHS FIN KOR EST BRA MDV MAR HRVPAN ZAF To measure the depth of stock markets, EGY TUR ALBIND BGD SVK CPV UKR BLZ USA VCT LTU VUT TUN ABW DMA NPL MNE CHL SAU BIH SRB this report primarily uses the stock value SYR SYC GMB PAK DZAGTM MEX LKA MOZ AGO KEN GUY PHL JAM BLR NGA COL PNG BTN MNG MDA NIC STP PRY BOL KSV GEO FJI WSM CRI ROMHND MAC TON NAM MKD RUS SLV KAZ IDN TGO SEN BEN ECU ARM URY SWZ BDI KHM traded indicator, which equals the value of GHA TZA ARG UGA MWI ZMB SLE SDN YEM MDG CMR NER GAB CAF AFG GNB MMR PER DOM SLB BWA VEN AZE CIV MLI BFA HTI LSO TMP stock market transactions as a share of GDP. 0 0 100 200 300 This market development indicator incor- porates information on the size and activity Private credit/GDP (%) of the stock market, not simply on the value Source: Calculations based on the Global Financial Development Database. of listed shares. Earlier work by Levine and Note: Correlation = 0.98. A signi�cant correlation coef�cient at the 5% level or better. Zervos (1998) indicates that the trading of ownership claims on � rms in an economy is closely tied to the rate of economic growth. There is substantial variation across coun- on the mixture of � nancial institutions and tries. Although the mean value of stock value markets operating in a �nancial system.7 The traded is about 29 percent of GDP, the stan- degree to which the � nancial system is rela- dard deviation is about double this value. In tively bank based or market based has been Armenia, Tanzania, and Uruguay, stock value an important topic in the � nancial develop- traded annually averaged less than 0.23 per- ment literature. In a recent contribution to cent over the 1980–2008 sample (10th per- this literature, Demirgüç-Kunt, Feyen, and centile). In contrast, stock value traded aver- Levine (2012) �nd that as economies develop, aged over 75 percent in China (both mainland services provided by � nancial markets tend and Hong Kong SAR, China), Saudi Arabia, to become relatively more important than Switzerland, and the Unites States (90th per- those provided by banks. centile). Also, this report con� rms Levine’s and Zervos’s results using other market devel- Second characteristic: Financial access opment indicators. In particular, it examines (inclusion) stock market capitalization, which simply measures the value of listed shares on a coun- But �nance is not just about the size of �nan- try’s stock exchanges as a share of GDP and cial institutions and securities; �nance is also securities market capitalization, which equals about the ability of individuals and � rms in the capitalization of the stock market plus the an economy to access �nancial services. Mea- capitalization of the private domestic bond sures of � nancial access are indeed strongly markets, divided by GDP. associated with economic development, a The relative size of banks and markets— relationship that is separate from the associa- called the financial structure ratio—mea- tion between � nancial depth and economic sures the ratio of private credit to stock mar- development. Besides the direct benefits of ket capitalization and provides information enhanced access to �nancial services, �nance 26 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 also reduces inequality, particularly through financial systems, these individuals and indirect labor market mechanisms. Speci�- enterprises with promising opportunities are cally, accumulated evidence shows that �nan- limited to their own savings and earnings. cial access accelerates economic growth, Financial access has been overlooked in tra- intensifies competition, and boosts the ditional literature on � nancial system char- demand for labor—and it usually brings big- acteristics, mostly because of serious data ger bene�ts to those at the lower end of the gaps on who has access to which financial income distribution (see, for instance, Beck, services and a lack of systematic information Demirgüç-Kunt, and Levine 2007, and Beck, on the barriers to broader access. The Global Levine, and Levkov 2010). It is important to Financial Development Database contains emphasize that the issue is not only access to both variables that measure the use of �nan- any form of � nance, but also the quality of cial services (which reflects both supply and �nancial services available to people. In other demand) as well as variables that focus more words, having a bank account is nice, but it is closely on the supply of �nancial services. also important to have a competitive interest The main proxy variable in the � nancial rate, reliable payment services, and so on. access category for �nancial institutions is the A well-functioning �nancial system offers number of bank accounts per 1,000 adults. savings, payments, and risk-management Other variables in this category include the products to as large a set of participants number of bank branches per 100,000 adults as possible. It seeks out and finances good (commercial banks), the percentage of � rms growth opportunities wherever they may be. with line of credit (all �rms), and the percent- Without inclusive financial systems, poor age of � rms with line of credit (small �rms). individuals and small enterprises need to rely When using these proxies, one needs to be on their personal wealth or internal resources mindful of their weaknesses. For example, to invest in their education, become entre- the number of bank branches is becom- preneurs, or take advantage of promising ing increasingly misleading with the move growth opportunities. Though still far from toward branchless banking. The number of conclusive, the existing body of evidence sug- bank accounts does not suffer from the same gests that developing the �nancial sector and issue, but it has its own limitations (in par- improving access to finance are likely not ticular, it focuses on banks only). only to accelerate economic growth but also The measure of access in �nancial markets to reduce income inequality and poverty. relies on various measures of concentration in Access to financial services—financial the market, the idea being that a high degree inclusion—implies an absence of obstacles to of concentration reflects dif�culties for access the use of these services, whether the obsta- for newer or smaller issuers. The variables in cles are price or nonprice barriers to �nance. this category include the percentage of mar- It is important to distinguish between access ket capitalization outside of the top 10 larg- to—the possibility to use—and actual use of est companies, the percentage of value traded �nancial services. In some cases, a person or outside of the top 10 traded companies, gov- business has access to services but decides ernment bond yields (3 month and 10 year), not to use them. But in other cases, price ratio of domestic to total debt securities, ratio barriers or discrimination, for example, bar of private to total debt securities (domestic), access. Failure to make this distinction can and ratio of new corporate bond issues to complicate efforts to define and measure GDP. access. Financial market imperfections, such The data for the � nancial access dimen- as information asymmetries and transaction sion of the Global Financial Development costs, are likely to be especially binding on Database came largely from the IMF’s the talented poor and on micro- and small recently established Access to Finance data- enterprises that lack collateral, credit his- base, based on earlier work by Beck, Demir- tories, and connections. Without inclusive güç-Kunt, and Martínez Pería (2007). 8 In GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 27 addition, a part of the �nancial access data is analysis and other more sophisticated mea- based on the Global Financial Inclusion Indi- sures; for example, Angelidis and Lyroudi cators database (Global Findex) that is being (2006) apply data envelopment analysis built at the World Bank (Demirgüç-Kunt and and neural networks to calculate ef�ciency Klapper 2012). The Global Findex is the �rst indexes using bank-by-bank data for the Ital- public database of indicators that consistently ian banking industry. But the data required measures individuals’ usage of �nancial prod- for this type of analysis are available only for ucts across countries and over time. It can be a small subsample of countries, and therefore used to track the effect of financial inclu- much additional data-collection work would sion policies and facilitate a deeper and more be needed to compile a comprehensive cross- nuanced understanding of how adults around country database. The background paper by the world save, borrow, and make payments. Cˇ ihák, Demirgüç-Kunt, Feyen, and Levine The data will be based on interviews with (2012) contains a discussion on data envel- at least 1,000 people per country in up to opment analysis and other examples of more 150 countries about their financial behav- sophisticated measures. ior through the Gallup World Poll survey. For � nancial markets, the basic measure The survey was rolled out in January 2011. of ef�ciency in the stock market is the turn- The � rst data set was made available to the over ratio, that is, the ratio of turnover to public in April 2012, and the full database capitalization in the stock market. The ratio- will be updated every three years, with head- nale of using this variable is that the higher line indicators of the use of bank accounts turnover relative to capitalization means rela- and formal credit, which are collected on an tively higher volumes of trading in the market annual basis. and more liquidity. This in turn means more scope for price discovery, better transmission of information in the price, and greater ef�- Third characteristic: Financial ef�ciency ciency of the market. In the bond market, the To perform its functions well, a �nancial sec- most commonly used variable is the tightness tor should be ef�cient. It should perform its of the bid-ask spread (with the U.S. and West- intermediating functions in the least costly ern European markets showing low spreads, way possible. If intermediation is costly, the and the Dominican Republic, Pakistan, Peru, higher costs may get passed on to households, Qatar, and Vietnam reporting high spreads) � rms, and governments. (In)ef�ciency mea- and the turnover ratio (although the mea- sures for institutions include indicators such surement of the latter often suffers from as overhead costs to total assets, net interest incomplete data). margin, lending-deposits spread, noninterest A range of other proxies for ef�ciency in income to total income, and cost to income � nancial markets have been used in empiri- ratio (table 1.1). Closely related variables cal literature (table 1.1). One of them is price include measures such as return on assets synchronicity, calculated as a degree of co- and return on equity. While ef�cient � nan- movement of individual stock returns in an cial institutions also tend to be more pro�t- equity market. The variable aims to capture able, the relationship is not very close (for the information content of daily stock prices. example, an inef�cient � nancial system can It is based on the notion that a market oper- post relatively high pro�tability if it operates ates ef�ciently when prices are informative in an economic upswing, while an otherwise about the performance of individual � rms. ef�cient system hit by an adverse shock may When their movements are highly synchro- generate losses). nized, they are less likely to provide such As with the other dimensions, these are individualized information (although one relatively crude measures of (in)ef�ciency. For also needs to control for common shocks to a subset of countries, it is possible to calculate economywide fundamentals to establish a ef�ciency indices based on data envelopment benchmark for this variable). Also, ef�ciency 28 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 can be approximated by the real transaction has been used extensively in the empirical cost. Based on daily return data of the listed literature. For other indicators, such as the stocks, this variable attempts to approximate regulatory capital to risk-weighted assets and the transaction costs associated with trad- nonperforming loans to total gross loans, ing a particular security. This variable helps the Global Financial Development Database determine the barriers to efficiency in the cross-references the Financial Soundness market. All these indicators are constructed Indicators database available on the IMF by compiling and statistically processing website (http://fsi.imf.org). Variables such as firm-level data from a variety of market the nonperforming loan ratios may be bet- sources. ter known than the z-score, but they are also known to be lagging indicators of soundness (Cˇ ihák and Schaeck 2010). Fourth characteristic: Financial stability One of the few reliable forward-looking Last, but not least, the degree of �nancial sta- indicators of �nancial instability is excessive bility is an important feature of the �nancial credit growth. The focus here is on excessive sector. There is a vast literature speci�cally credit growth. A well-developing financial on measuring systemic risk. Because of the sector is likely to report expansion in credit importance of �nancial stability for broader growth. Without credit growth, financial macroeconomic stability, the topic is some- sectors would lack depth or would not be times treated as separate from the other three able to provide good access to � nancial ser- dimensions. 9 But financial stability is an vices. Credit growth is important, and indeed important feature of �nancial systems, and it may be necessary, even if it is connected with is closely interlinked with the broader process some instability.10 But a very rapid growth of � nancial development. To illustrate this, in credit is one of the most robust com- imagine a country where banks’ lending stan- mon factors associated with banking cri- dards become very loose, with banks provid- ses (Demirgüç-Kunt and Detragiache 1997; ing loans left and right, without proper risk Kaminsky and Reinhart 1999). IMF (2004), management and loan monitoring. On the for example, estimated that about 75 percent surface, one could observe the rapid growth of credit booms in emerging markets end in as a sign of deepening and increased access banking crises. Typically, credit expansions to �nance. Also on the surface, the �nancial are fueled by overly optimistic expectations sector can seem ef�cient, for some period of of future income and asset prices, often time: without the loan approval process, such combined with capital inflows. Over time, banks would be able to lower their costs, at households and firms accumulate substan- least until the loans turned bad. And this is tial debt while income does not keep pace. A the problem, of course: the system would decline in income or asset prices then leads be unstable and likely would end in a crisis. to an increase in nonperforming loans and For more on the complex linkages between defaults. If the problem is severe, the coun- �nancial development, �nancial fragility, and try experiences a banking crisis. Drehmann, growth, see, for example, Loayza and Ran- Borio, and Tsatsaronis (2011) examine the ciere (2006). performance of different variables as anchors The key variable used here to measure for setting the level of the countercyclical �nancial stability is the z-score, de�ned as the regulatory capital buffer requirements for sum of capital to assets and return on assets, banks, �nding that the gap between the ratio divided by the standard deviation of return of credit to GDP and its long-term backward- on assets. This variable explicitly compares looking trend performs best as an indicator buffers (capitalization and returns) with the for the accumulation of capital, because this potential for risk (volatility of returns). The variable captures the build-up of systemwide z-score has a direct link with the probabil- vulnerabilities that typically lead to banking ity of default, and for this reason the variable crises. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 29 BOX 1.2 To Aggregate or Not To provide a rough sense of how � nancial systems To convert the representative indicator in each of stack up across the 4x2 dimensions, it is helpful to the 4x2 characteristics to a 0–100 scale, each score is convert the individual characteristics to the same rescaled by the maximum for each indicator and the scale. To prepare for this, the 95th and 5th percentile minimum for each indicator. The rescaled indicator for each variable for the entire pooled country-year can be interpreted as the percent distance between data set are calculated, and the top and bottom 5 the worst (0) and the best (100) outcome, de�ned by percent of observations are truncated. Speci�cally, all the 5th and 95th percentile of the original distribu- observations from the 5th percentile to the minimum tion. These winsorized and rescaled variables are the are replaced by the value corresponding to the 5th core of much of the analysis presented in this chapter. percentile, and all observations from the 95th percen- To arrive at a more condensed aggregate indica- tile to the maximum are replaced by the value corre- tor, it may be useful to examine the average across sponding to the 95th percentile. In effect, the 5th and the various characteristics; however, a strong caveat 95th percentile become the minimum and maximum applies. An ongoing and rather active debate on mul- of the new (truncated) data set. The main reason tidimensional indices (such as the Multidimensional for truncating the “tails� of the distribution is that Poverty Index, Human Development Index, and sometimes the best and worst scores are very extreme various Unsatis�ed Basic Needs indices long used in and may reflect some peculiar (idiosyncratic) features many countries) has focused much criticism on the of a single jurisdiction. However, the top and bot- dif�culty of the choice of weights for such an index tom 5 percent of observations are not dropped from (for example, Ravallion 2011). Mindful of the debate the sample completely. If they were dropped, the and the shortcomings associated with creating such calculations would lose too much of the potentially mash-up indices, this report does not explicitly pres- valuable information. Replacing the top and bottom ent such a formal mash-up index. Nonetheless, the 5 percent of observations with the 95th and 5th per- data made available on this report’s website allow centile value, respectively, ensures that much of the interested users to assign different weights to the var- original information is still retained. This so-called ious characteristics and calculate their own aggregate winsorizing is consistent with approaches used in indices. earlier literature. The advantage of the credit growth vari- For �nancial markets, the most commonly able is that it is relatively easy to observe and used proxy variable for (in)stability is mar- monitor. Also, unlike some of the other mea- ket volatility, although other proxies are also sures (for instance, those that include nonper- included in the database (table 1.1). One of forming loan ratios), it is a forward-looking these variables is the skewness, the reason measure of instability. A disadvantage is that being that a market with a more negative the de�nition of “excessive� credit growth is skewed distribution of stock returns is likely not trivial. Also, this measure does not, by to deliver large negative returns, and likely to itself, capture situations where �nancial sec- be prone to instability. tor problems have already crystallized in a Other variables approximating (in)stability full-blown crisis. In such situations, credit is in the stock market are the price-to-earnings declining in real terms rather than growing. ratio (P/E ratio) and duration (a re� ned ver- It is therefore important to amend the exces- sion of the P/E ratio that takes into account sive credit growth indicator, as an ex ante factors such as long-term growth and inter- measure of �nancial instability, by including est rates). These variables are based on the credit declines as ex post proxies for situa- empirical fact that market prices contain tions of �nancial instability. expectations of future cash flows and growth 30 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 instead of current fundamentals only, and interventions. Finally, another group of indi- therefore stock prices may be more volatile cators relates to the features of the underly- and negatively skewed in the future. ing financial infrastructure. This includes basic indicators on information disclosure, contract enforcement, and other quantitative Measuring the enabling environment for characteristics of �nancial infrastructure (for �nance: A start and an important area example, public registry coverage in percent for further data work of adults, private bureau coverage in percent The focus of the 4x2 matrix is on charac- of adults, procedures to enforce contracts, terizing financial systems (the middle part time to enforce contract, and cost to enforce of �gure 1.2). It does not explicitly include contracts). Several other traits of the enabling variables capturing financial sector policy, environment for � nance are included in the such as features of financial sector regula- Global Financial Development Database and tion and supervision (the bottom of figure listed in this chapter’s annex. 1.2). The reason for focusing on measures But this is just a start. For policy evalua- of the functioning of financial systems is tion and policy design purposes, it is impor- that those indicators bridge the gap between tant to start collecting more consistent and policy measures and �nal objectives, such as more comprehensive information on gov- growth, poverty alleviation, and the expan- ernment policies in the � nancial sector (for sion of economic opportunities. Financial example, on supervision of nonbank �nancial depth, access, ef�ciency, and stability func- institutions and �nancial markets). This is an tion as “intermediate� indicators and targets. important gap in the globally available data; To some extent, this is an analogy with mon- future reports hope to go in more depth into etary policy, where intermediate targets have how this gap might be �lled. a relatively clear link to the policy variable (such as a central bank’s interest rate) and an impact on the policy target (such as future SELECTED FINDINGS inflation rate). Financial system multidimensionality This report, however, has started the pro- cess of assembling comprehensive data on One basic, yet important, observation the enabling environment for �nance: �nan- derived from the Global Financial Develop- cial sector policies, regulations, supervisory ment Database is that the four characteris- practices, legal and accounting systems, tics of � nancial systems are far from closely and so forth. As part of the work underly- correlated across countries (�gure 1.4). Each ing chapter 2 of this report, a comprehen- characteristic captures a different, separate sive and updated data set on bank regula- facet of financial systems. Capturing only tion and supervision around the world was � nancial institutions and not � nancial mar- put together, building on earlier work by kets would be insuf�cient. Also, looking only Barth, Caprio, and Levine (2004). The data- at � nancial depth as the only proxy would base also covers policies and issues that go not be sufficient. And similarly, focusing beyond the narrow concept of banking regu- only on �nancial stability or on access or on lation and supervision, such as deposit pro- ef�ciency would be insuf�cient. Stability has tection systems and resolution issues. Also, particularly low correlation with the other the World Bank has recently published a three characteristics. comprehensive update on payment systems and the related policies around the world— Important differences across regions some of these results are featured in chapter and income groups 5. As part of chapter 4, new data are pre- sented on development � nancial institutions A regional comparison shows major differ- and some other forms of direct government ences in the four characteristics of � nancial GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 31 FIGURE 1.4 Correlations among Financial System Characteristics a. Financial Institutions Depth vs. Access Depth vs. Efficiency Correlation = 0.80* Correlation = 0.47* Private credit to GDP (%) Private credit to GDP (%) MYS GRC ESP ITA AUT NLD PRT MLT LCA HKG CAN CHE CHN AUS VNM MYSNZL JPN 100 THA EST SGP JPN KOR 100 EST BRB SGP KOR SVN ISR ISR THA BHS 80 80 MUS GRD ATG PAN KWT MAR ZAF BHR ZAF LBN HUN CHL HRV KNA CHLHUN LBN JOR MDV BLZ MDV VUTLTU 60 BIH UKR 60 ABW CPV BLZ UKR BIH CZE DMA NPL HND VCT MAC SAU OMN BRA CRI SRB BRN OMN FJI NAM TON WSM NAM 40 BGD KAZ TUR MNG 40 MNG TTO ROM BGD ALB RUS MKD BLR STP NGA IRN KEN KSV GEO MDA PRY GEO NICKEN BOL MDA COL EGY DJI PHL GUY MRT PHL KHMPNG JAM DJI PNG GTM MOZ SUR SWZ ARM SYC PER ARM SYC IDN SWZ LKA PAK PER 20 AGO AZE ETH TGO TZA SYR DZA BWA URY MEX 20 MWI GMB SLB TMP AGOAZE TZA BWA PAK DOM URY DZA MEX SYR VEN LAO COM UGA GHA HTI LSO ARG LAO HTI UGA LBR COM LSO ARG MDG CMRSLE ZMB SLE LBY CAF AFG GAB MDG YEM TCDYEM MMR COG 0 0 0 1,000 2,000 3,000 80 85 90 95 100 Accounts per thousands adults, commercial banks 100 minus lending-deposit spread (%) High income Low income Lower middle income Upper middle income High income Low income Lower middle income Upper middle income Depth vs. Stability Access vs. Efficiency Correlation = 0.02 Correlation = 0.46* Private credit to GDP (%) Accounts per thousands adults, commercial banks DEU IRL GRC LCA CHE CYP DNK NZL NLD FRA GBR AUT JPN MLT ITA SWE PRT LUX AUS HKG VNM ESP CHN MYS CAN 100 EST BEL SVN KOR BRB SGP 3,000 UKR JPN KOR FIN ISR THA BHS GRD MUS 80 PAN HUN BHR MAR KWT ZAF MNEKNA LBN SGP CHL JOR LTU UKR BLZ VUT HRV 2,000 EST BGR 60 USA BIH CZE TUN DMA NPL MYS MAC SYC BRA CRI HND SVK SAU SRB BRN OMN MKD NAM IND WSM MNG LVA 40 KAZ BTN RUS BGD ROM MNG BLR TTO ALB SLV TUR MDA MDV THA MEX ISR HUN GEO PRY MDA NIC NGA BOL EGY MRT KEN GUY COL 1,000 BLZ BIH OMN ZAF LBN SUR ECU PHL KHM PNG GTM JAM QAT SYC ARMSWZ BWA MOZ PER LKAIDNSEN GEO ARG 20 URY AZE GMB MEX TZA TGO BEN VEN PAKBDI AGO SYR CIV MLI ETH DOM BFA PER ARM URY WBG BWA SWZ NAM PHL NOR LAO DZA MWI LSO UGA GHA HTI CMR KEN BGD MDG ARG NER HTI DZA AFG CAF SLE GAB SDN ZMB LBY RWA LSO PAK SYR YEM KGZ UGA SLE PNG TZA MMR TCD AGO DJI COM YEM MDG LAO AZE 0 0 0 20 40 60 80 80 85 90 95 100 Z-score-weighted average commercial bank 100 minus lending-deposit spread (%) High income Low income Lower middle income Upper middle income High income Low income Lower middle income Upper middle income Access vs. Stability Efficiency vs. Stability Correlation = 0.02 Correlation = 0.18 Z-score–weighted average commercial bank Z-score–weighted average commercial bank 80 80 HRV CMR MAR NAM LBN THA FSM WSM NAM COL THA LBY LBN HTI HTI QAT QAT 60 60 JAM MYS MAC MYS CHN SGP DOM SGP CAN GUY BRBCZE PHL ESP PHL BHS IDN VUT 40 KEN GHA BIH OMN TUR KOR 40 KEN GTM TTO MOZ JOR BIH LKA NPL OMN VNM AUS KOR PER PER MRT PNG MUS PNG ALB BLR PRT HND SYR LSO NOR BWA UZB ZAF LSO BWA HKG BHR EGY SYRZAF NOR ETH PAK AGO SWZ ITA MLT MWI CRI AGO SWZ BRN PAK UGA ARM BLZ ARM BLZ KWT MKD DZA MNG ZAR UGA BOLMNG DZA PAN GNQ ARG AUT JPN BRA SRBSUR JPN ARG YEM ISR BGR SYC BGR YEM ISR 20 GAB RWA TGO KGZ TZA BGD SAU HUN MDA MEX SYC NLD 20 PRY KGZ NGANIC RWA MDA TZA GRD ROM BGD RUSSVN VEN HUN NZL MEX KNA KHM UKR UKR CHE SLE SLE LCA MDG GEO KAZ GRC MDG GEO ZMB GMB TCD TJK EST AZE EST AZE DMA LAO CAFAFG URY LVA LAO LVA URY MMR LTU 0 0 0 1,000 2,000 3,000 80 85 90 95 100 Accounts per thousands adults, commercial banks 100 minus lending-deposit spread (%) High income Low income Lower middle income Upper middle income High income Low income Lower middle income Upper middle income (�gure continues next page) 32 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 1.4 Correlations among Financial System Characteristics (continued) b. Financial Markets Depth vs. Access Depth vs. Efficiency Correlation = 0.39* Correlation = 0.47* (Stock market capitalization + outstanding domestic private debt securities)/GDP (Stock market capitalization + outstanding domestic private debt securities)/GDP 200 LUX CHE AUS MYS HKG USA ZAF 200 LUX MYS ZAF CHE DNK AUS USA HKG SGP SGP SWE KOR KOR CAN 150 CHL ESP GBR CAN 150 CHL NLD GBR ESP IRL IRL FRA PNG JOR JOR BEL JPN BRB JPN CHN 100 NOR BRA ISR CHN 100 BHR KNA MNE PRT ISRBRA KWT NOR IND ISL IND THA THA SAU SAU AUT AUT DEU MAR DEU ITA MAR ITA FIN RUS RUS COL PHL COL PHL TTOPERMUS MEX PER 50 MEX GRC MUS 50 FJI JAM HRV KAZ IDN EGY GRC EGY IDN KEN CZE TUR MLT POL NZL PAN MLT NPL LBN OMN POL NZL BWA SVN CIV MWI SRB LKA HUN HUN SVN TUR LKA CYP TUN CYP SLV MKD UKRNGA ARG ZMB IRN PAK VNM ARG BOL UGABGRLTU MNGEST IRN SVK NAM ECUGHA BGD GEO LVA CRI VEN ARMURY KGZ 0 0 0 20 40 60 80 80 50 100 150 Percent market capitalization out of the top 10 largest companies (%) Stock market turnover ratio (%) High income Low income Lower middle income Upper middle income High income Low income Lower middle income Upper middle income Depth vs. (in)stability Access vs. Efficiency Correlation = 0.09 Correlation = 0.51* (Stock market capitalization + outstanding domestic private debt securities)/GDP Percent market capitalization out of the top 10 largest companies (%) 200 MYS CHE HKG USA DNK AUS ZAF LUX 80 ZAF CAN CHN SGP SWE SGP IND USA KOR CAN GBR KOR 150 CHL NLD ESP MYS JPN ESP HKG GBR IRL 60 PHL LKA IRN IDN EGY AUS FRA THA DEU CHL TUR JOR BEL MUS POL ISR JPN CHN NZL BRA 100 BHR KWT ISR PRT BRA NOR 40 CHE SAU RUS ITA THA IND GRC AUT NOR ISL PER SAU MEX MAR DEU ITA ARG JOR FIN RUS COL PHL MAR MUS MEX PER 50 JAM EGY IDN HRV GRC 20 SVN COL CYP IRL PAN LBN MLT KEN NZL TUR KAZ POL BWA SVN LKA CZE HUN OMN SRB CYP LUX HUN TUN NGA PAK MKD VNM UKR MLT BGR ARG MNG LTU EST SVK LVA NAM CRI 0 0 10 20 30 40 50 0 50 100 150 Asset price volatility Stock market turnover ratio (%) High income Low income Lower middle income Upper middle income High income Low income Lower middle income Upper middle income Access vs. (in)stability Efficiency vs. (in)stability Correlation = 0.21 Correlation = 0.24* Percent market capitalization out of the top 10 largest companies (%) Stock market turnover ratio (%) 80 HKG USA CHN KOR ITA TUR USA SGP CAN CHN ZAF IND 150 VNM GBR KOR RUS JPN MYS HKG ESP ESP THA HUN 60 PHL LKA IDN AUS JPN GBR DEU IND CHL THA EGY DEU TUR NLD FIN ISR MUS 100 SGP AUS SWE NOR POL KWT NZL BRA SAU FRA GRC BRA 40 CHE SAU ITA PER NOR RUS GRC ISR CHE PAK DNK CAN EGY IDN MEX JOR ARG IRL POL MAR 50 OMN NZL BEL ZAF IRL JOR PRT SVN COL 20 CYP MYS PHL LKA MEX CZE TUN CHL NGA LBN MAR KEN COL EST MLT LUX HUN ISL UKR MNG CYP MUS LTU ARG KAZ BHR JAM BWA CRI SVN SRB BGR HRV LVA PAN MLT SVK MKD LUX PER NAM 0 0 10 20 30 40 50 0 20 30 40 50 Asset price volatility Asset price volatility High income Low income Lower middle income Upper middle income High income Low income Lower middle income Upper middle income *Indicates a significant correlation coefficient at the 5% level or better GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 33 FIGURE 1.4 Correlations among Financial System Africa on access to � nance (table 1.2). This Characteristics (continued) number resonates with the complaints heard during the unrest in the region in 2011.11 c. Financial Sector Depth: Institutions vs. Markets Much of the differences among regions are correlated with differences in income levels. Private credit/GDP (%) 300 Countries that have lower income tend to CYP also show lower values on the 0–100 scale in the 4x2 framework (table 1.2 and �gure IRL 1.5). However, the stability indicator is not ESP 200 PRT NLD GBR very correlated with income level—a point LUX CHE HKG highlighted quite dramatically by the global �nancial crisis. MLT AUS AUT CHN ITA FRA VNM ISL GRC DEU MYS 100 EST SVN FINTHA JPN ISRBEL KOR SGP LTU UKR PAN LBN TUN MUS HRV MAR MNE KNAJOR CHL ZAF Large disparities in �nancial systems USA SVK CRI NAM SRB MKD BGD SLV MNG BOL NGA NPL FJI SAU KAZ RUS TUR EGY KEN COL BRA IND across countries GEO GUY ECU LKA PHL JAM IDN PNG ARM URY VEN BWA PAK PER MEX GHA UGACIV MWI ARG 0 ZMB Behind these regional and peer group aver- 0 100 200 300 400 500 ages are vast differences among individual (Stock market capitalization + outstanding countries, and in some cases also major domestic private debt securities)/GDP differences among different parts of each Source: Calculations based on the Global Financial Development Database. country’s financial sector. The data from Note: See table 1.2. the Global Financial Development Database demonstrate rather strikingly the large differ- ences in � nancial systems around the globe. systems across the key regions as of 2010 For example, the largest � nancial system in (table 1.2). The results are by and large in the sample is more than 34,500 times the line with what one would expect, with Sub- smallest one. Even if the financial systems Saharan Africa scoring the lowest on aver- are rescaled by the size of the corresponding age on most of the characteristics, and high- economies (that is, by their GDP), the larg- income countries scoring the highest on most est (deepest) financial system is still some dimensions. A remarkable number is the rela- 110 times the smallest (least deep) one. And tively low score of Middle East and North even if the top and bottom 5 percent of this TABLE 1.2 Financial System Characteristics: Summary Financial Institutions East Asia Europe and Latin America Middle East and Sub-Saharan (Mean) High income and Paci�c Central Asia and the Caribbean North Africa South Asia Africa Depth 69 43 37 37 33 32 17 Access 43 23 35 30 14 16 10 Ef�ciency 80 70 65 62 83 81 51 Stability 42 52 20 35 57 38 32 Financial Markets East Asia Europe and Latin America Middle East and Sub-Saharan (Mean) High income and Paci�c Central Asia and the Caribbean North Africa South Asia Africa Depth 43 38 12 21 24 17 20 Access 46 80 56 40 50 85 77 Ef�ciency 29 40 17 8 24 49 7 Stability 66 60 43 64 81 56 54 (table continues next page) 34 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 TABLE 1.2 Financial System Characteristics: Summary (continued) Financial Institutions Upper middle Lower middle (Mean) High income income income Low income Depth 84 44 28 13 Access 55 32 19 5 Ef�ciency 86 75 61 42 Stability 35 38 40 35 Financial Markets Upper middle Lower middle (Mean) High income income income Low income Depth 51 27 16 10 Access 53 58 69 29 Ef�ciency 45 19 20 21 Stability 53 60 53 44 Source: Calculations based on the Global Financial Development Database. Note: The summary statistics refer to the winsorized and rescaled variables (0–100), as described in the text. Financial Institutions—Depth: Private Credit/GDP (%); Access: Number of Accounts Per 1,000 Adults, Commercial Banks; Ef�ciency: Net Interest Margin; Stability: z-score. Under Financial Markets—Depth: (Stock Market Capitalization + Outstanding Domestic Private Debt Securities)/GDP; Access: Percent Market Capitalization Out of the Top 10 Largest Companies (%); Ef�ciency: Stock Market Turnover Ratio (%); Stability: Asset Price Volatility. FIGURE 1.5 Financial System Characteristics, by Income Group, 2010 a. Financial institutions b. Financial markets 70 80 60 60 Index (0–100) Index (0–100) 50 40 40 30 20 20 0 Depth Access Efficiency Stability Depth Access Efficiency Stability High income Low income Lower middle income Upper middle income Source: Calculations based on the Global Financial Development Database. Note: The summary statistics refer to the winsorized and rescaled variables (0–100), as described in the text. See also table 1.2. distribution are taken out, the ratio of the examines country-level data, one � nds vast largest to the smallest is about 28—a large differences in � nancial sector depth, as well degree of disparity, considering that these are as in the other characteristics. not raw �gures but ratios relative to the size The cross-country differentiation along of the economy. Similar orders of magnitude the key characteristics of � nancial systems are obtained for the other characteristics of can be seen from the scatter plots in �gure �nancial systems.12 In other words, when one 1.4 as well as from cartograms such as the GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 35 one shown for illustration in �gure 1.6. The liquidity shocks. In addition, � nancial insti- scatter plots and the cartogram underscore tutions on average rebounded faster than the large cross-country differences. The mea- markets, showing improvements in depth surement framework underscores that �nan- and ef�ciency after the crisis. This improve- cial sectors in jurisdictions such as the Repub- ment seems to have been the case so far, for lic of Korea and the United States exhibit a example, for Brazil and other Latin Ameri- relatively great financial market depth, as can countries (de la Torre, Ize, and Schmuk- one would expect. The United States has less ler 2011), China (box 1.3), and many Sub- deep �nancial institutions, reflecting the less Saharan African countries (see, for example, bank-centric (and more market-based) nature World Bank 2012). However, the medium- of the U.S. financial system. Several Euro- term effect of the crisis on � nancial systems pean countries exhibit relatively great �nan- still remains to be seen, and will be examined cial depth. further in future issues of the Global Finan- Financial systems have changed. As illus- cial Development Report. trated in �gures 1.7 and 1.8, the most visible change is the observed declines in stability, Increased importance of securities which in turn reflects the increased volatility markets at higher income levels in returns by � nancial institutions in some countries and in most �nancial markets. The Global Financial Development Database Overall, the data from the Global Finan- allows for an examination of the relative size cial Development Database suggest that the of � nancial institutions and � nancial mar- key disparities among countries in terms of kets around the world. The issue of finan- the nature of their financial systems have cial structure—usually approximated by the somewhat subsided in the aftermath of the relative size of bank credit and stock market recent crisis, as financial sectors in many capitalization—has been an important topic medium- and low-income countries were in the policy debate. relatively more isolated from the global tur- In a recent paper that used data that are moil, and therefore less affected by the global part of the Global Financial Development FIGURE 1.6 The Uneven Nature of Financial Systems (Illustration) Source: Calculations based on the Global Financial Development Database. Note: The map is for illustration purposes only. Country sizes are adjusted to reflect the volume of �nancial sector assets in the jurisdiction, measured in U.S. dollars at the end of 2010. The image was created with the help of the MapWindow 4 and ScapeToad software. 36 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 1.7 Financial Systems: 2008–10 versus 2000–07 (Financial Institutions) a. Depth b. Efficiency Average 2008–10 Average 2008–10 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Average 2000–07 Average 2000–07 c. Access d. Stability Average 2008–10 Average 2008–10 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Average 2000–07 Average 2000–07 Source: Calculations based on the Global Financial Development Database. Database, Demirgüç-Kunt, Feyen, and would need to � nd out if taxes, regulations, Levine (2012) examine empirically the issue legal impediments, or other distortions are of �nancial structure and �nd that, as econ- leading to excessive reliance on banks or omies develop, use of services provided by markets. Using policy to facilitate a shift securities markets increases relative to those from a bank-centric system to a more market- provided by banks. This work suggests that based system is never an easy task. Actively policies and institutions should adapt as intervening to develop markets is likely to be countries develop in order to allow � nancial problematic. Interventions should be more structure to evolve. along the lines of fostering an enabling envi- The existing research and policy work do ronment and reducing impediments. Even not provide enough guidance to justify tar- in systems with a relatively strong state role geting a particular � nancial structure for a in the economy, shifts in the financial sec- particular country. However, if market or tor structure do not occur overnight. China bank development is too skewed compared (box 1.3) is a case in point: despite policy to what one could expect given their level of intentions and reforms aimed at promot- economic development, the above research ing nonbank � nancial institutions and mar- � ndings provide a reason to dig deeper: one kets, the �nancial system remains very much GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 37 FIGURE 1.8 Financial Systems: 2008–10 versus 2000–07 (Financial Markets) a. Depth b. Efficiency Average 2008–10 Average 2008–10 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Average 2000–07 Average 2000–07 c. Access d. Stability Average 2008–10 Average 2008–10 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Average 2000–07 Average 2000–07 Source: Calculations based on the Global Financial Development Database. dominated by large banks, and in some ways in the fourth quarter of 2008, around the has become even more bank-centric during Lehman failure. On the surface, it may seem the recent period of rapid credit growth. as if the U.K. financial sector underwent a “productivity miracle� from the 1980s onward, as finance appeared to rise as a The “bright� and “dark� sides of share of GDP despite a declining labor and �nancial systems capital share. However, a decomposition of The data from the Global Financial Devel- returns to banking suggests that much of opment Database can be used to examine the growth reflected the effects of higher the notion that growth of � nancial systems risk taking (Haldane, Brennan, and Madou- may seem explosive. To some extent, this ros 2010). Leverage, higher trading pro�ts, notion reflects the inadequacy of some of the and investments in deep-out-of-the-money available proxies for � nancial systems. For options were the risk-taking strategies that instance, in the case of the United Kingdom, generated excess returns to bank sharehold- the nominal value-added of the �nancial sec- ers and staff. Subsequently, as these risks tor (as measured in the System of National materialized, the “miracle� turned into a Accounts) grew at the fastest pace on record mirage. 38 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 1.3 China Case Study: Large Banks and the Need to Diversify to Markets The 4x2 measurement framework allows country restructured, and a resolution mechanism and inves- of� cials and analysts to examine � nancial systems tor protection scheme have been set up. Pension sec- across borders and to put them in a broader interna- tor reform has also progressed, with the establish- tional perspective. The rapid and somewhat uneven ment of a National Social Security Fund in 2000. development of China’s � nancial system provides for The � xed income market has grown as an alterna- an interesting case study. tive funding channel, but it remains heavily concen- Over the past three decades, China’s economy trated in public sector securities. The equity market has maintained high growth rates. Since the start of mainly meets the needs of large enterprises, in spite reforms in 1978, productivity growth has been rapid of recent progress in establishing a multilayer equity and capacity has been expanded by very high levels market to facilitate funding to small and medium of investment. enterprises (SMEs). Assets under management by China has made progress in moving toward a more the insurance sector corresponded to less than 11 commercially oriented � nancial system and toward percent of household bank deposits. Trust, � nancial strengthening of its banks (World Bank 2011c). leasing, and � nance companies have all been grow- This progress has been underpinned by reforms that ing rapidly but remain small relative to banks. China included recapitalizing the banking system, upgrad- also has a flourishing informal � nancial sector, parts ing the prudential regulatory regime, opening the of which provide funding to SMEs and small retail financial system following accession to the World investors. Nonetheless, the large commercial banks Trade Organization, and taking steps to reform inter- make up almost two-thirds of commercial bank est rate and exchange rate policies. Reform of the assets, with the assets of the four largest banks each joint-stock banks has boosted the commercial orien- exceeding 25 percent of GDP (and ranking among tation of the banking system, and reform of the rural the largest banks in the world). credit cooperatives has yielded some initial results. The commercial banking sector has grown very FIGURE B1.3.1 The Chinese Financial Sector rapidly in the past decade. In terms of the 4x2 frame- 100 work, the Chinese banking sector was already rather large, being close to or at the 100 score in terms of 80 the depth indicator (see �gure B1.3.1). In this sense, China may seem already “developed� in terms of the Index (0–100) 60 size of its � nancial institutions. However, the rapid credit growth of the 2000s may have been too rapid 40 and contributed to a somewhat reduced stability score. Perhaps greater increase in depth of the � nan- 20 cial markets may have been more warranted. The 4x2 framework underscores that one of 0 the challenges for the Chinese � nancial sector is to 2000 2002 2004 2006 2008 2010 increase its diversi� cation. Banks, particularly the Depth (institutions) Stability (institutions) largest ones, still dominate � nancial intermediation. Depth (markets) Stability (markets) Recognizing this challenge, the country authorities Source: Calculations based on World Bank 2011c. have taken steps to diversify the financial sector. Note: For simplicity, the �gure shows only four out of the eight variables In the securities sector, key companies have been in the 4x2 framework. See also the note to table 1.2. More examples of the explosive growth They observe that �nancial systems develop- of � nancial systems using the data from the ment paths exhibit “convexities,� as rising Global Financial Development Database can participation and interconnectedness gener- be found in de la Torre, Ize, and Schmukler ate positive externalities that promote further (2011) and de la Torre, Feyen, and Ize (2011). participation and interconnectedness. Thus, GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 39 much of financial system growth may be to the crisis. For example, the �nancial stabil- explosive. According to the authors, a counter- ity indicators for many countries show dete- part of such explosiveness is that the associa- rioration several years prior to the crisis (see tion between �nancial development (approxi- figure B1.3.1 in box 1.3 for an illustration mated, for example, as private credit to GDP) for China). This finding is consistent with and real development (output growth) exhibits the observation by Anginer and Demirgüç- decreasing returns. In other words, the asso- Kunt (2011), who construct a default risk ciation between �nance and growth levels off measure for publicly traded banks using the at some point. This result is consistent with Merton contingent claim model, and exam- �ndings of recent papers that regress output ine the evolution of the correlation struc- growth against �nancial depth indicators.13 ture of default risk for some 1,800 banks in A different aspect of this convexity has over 60 countries. Based on their measure, been brought up recently by C ˇ ihák, Muñoz, which is a more sophisticated analogue of and Scuzzarella (2011). Using a subset of data the z-score used in this chapter, they � nd a from the Global Financial Development Data- signi�cant increase in default risk codepen- base, and building on an earlier theoretical dence over the three-year period leading to paper by Nier and others (2007), they exam- the �nancial crisis. They also �nd that coun- ine the “bright� and “dark� sides of cross- tries that are more integrated, and that have border financial interlinkages. They ask liberalized �nancial systems and weak bank- whether making a country’s banking sector ing supervision, have higher codependence in more linked to the global banking network their banking sector. The results support an renders that country more or less prone to increase in scope for international supervi- banking crises. Their answer, interestingly, is sory cooperation, as well as capital charges that it depends on how connected the coun- for “too-connected-to-fail� institutions that try’s banking sector already is. For banking can impose signi�cant externalities. sectors that are not very connected to the The 4x2 framework also allows examin- global banking network, increases in inter- ing the effects of the global � nancial crisis. connectedness are associated with a reduced Box 1.4 illustrates this in the case of Roma- probability of a banking crisis. Once intercon- nia, a country whose �nancial sector seemed nectedness reaches a certain value (estimated relatively sound based on conventional ratios to be at about the 95th percentile of the distri- (such as capital adequacy and nonperforming bution of countries in terms of interconnected- loan ratios) but that was subjected to rather ness), further increases in interconnectedness large shocks during the crisis. Figures 1.7 and can increase the probability of a banking cri- 1.8 examine the crisis effect in a cross-section sis. Also, the analysis suggests that it is impor- of countries. tant to distinguish whether the cross-border interlinkages are stemming primarily from CONCLUSION banks’ asset side or from their liabilities side: increasing interconnectedness on the liabilities The 4x2 framework presented in this chapter (borrowing) side is more likely to become det- puts a spotlight on the multifaceted nature of rimental to banking stability than increasing modern � nancial systems. Focusing only on interconnectedness on the asset (creditor) side. one dimension—say, �nancial depth or �nan- cial stability—would be shortsighted. Also, focusing only on �nancial institutions, or just Analysis of the crisis: Increased on banks, is too much of a simpli�cation and instability in the run-up, decreased can lead to distorted results and biased policy access in the aftermath conclusions. The rich data set in the Global Financial Devel- This chapter illustrates that financial opment Database allows one to examine in sectors come in different shapes and sizes, more depth the developments in the run-up and they differ widely in terms of the 4x2 40 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 1.4 Romania Case Study: Rapid Growth Enabled by Foreign Funding In the run-up to the global � nancial crisis, Romania’s FIGURE B1.4.1 Romania’s Financial Sector � nancial sector has gone through a period of rapid 100 growth, reflected in an increase in the measured � nancial depth. Similarly to many other countries in the region, the rapid growth of domestic credit 80 was fueled by ample funding provided by parents of foreign-owned banks to their subsidiaries in Roma- Index (0–100) 60 nia. In terms of the 4x2 framework, Romania’s score for � nancial institutions’ depth grew from only 3 in 40 2000 to 28 in 2007, and its score for � nancial mar- kets’ depth grew from 1 to 13 over the same period. The Romanian banking system, which dominates 20 the � nancial sector, entered the crisis with relatively high reported capitalization and liquidity ratios 0 (IMF 2009). Also, the ratios of nonperforming loans 2000 2002 2004 2006 2008 2010 to total loans reported before the crisis were rather Depth (institutions) Stability (institutions) low; however, this � nding was mostly just a reflec- Depth (markets) Stability (markets) tion of the high credit growth that masked to some Source: Calculations based on World Bank; IMF 2009. extent the underlying weaknesses in the system. The Note: For simplicity, the �gure shows only four of the eight variables z-score, that is, the proxy for stability used in the in the 4x2 framework. See also the note to table 1.2 4x2 framework, suggested that the soundness of Romanian banks was far from perfect in the run-up becoming undercapitalized as the downturn contin- to the crisis. ues. The FSAP therefore called for strengthening of A rapid deterioration in market confidence in capital positions of some banks and for maintain- the Romanian economy has led to bouts of down- ing by parents of foreign-owned banks those lines of ward pressure on the exchange rate, upward pres- credit to their subsidiaries and corporate borrowers sure on interest rates, and a large decline in equity in Romania. In terms of the 4x2 framework, these values (some 80 percent between 2008 and 2009). stability challenges are refl ected in major declines These effects led to sharp increases in nonperform- of the stability indicators, both for � nancial institu- ing loans, putting strains on bank capital positions. tions and � nancial markets, in 2008 and 2009. Also, Stress-testing analysis performed during the recent the framework highlights that the crisis has halted, Financial Sector Assessment Program (FSAP) (IMF at least temporarily, Romania’s increases in � nancial 2009) suggested that some banks were at risk of depth. dimensions. More specifically, the chap- with this report, should help country offi- ter also documents developments during cials, researchers, and anybody else with the global � nancial crisis, not only in terms interest in the matter to better benchmark of � nancial instability, but also in terms of � nancial systems. The Statistical Appendix �nancial depth, access, and ef�ciency. to the report includes country tables with Despite the remarkable progress in gather- select indicators, as well as aggregates across ing data and information on �nancial systems regions and income groups. A pocket edi- around the world in recent years, researchers’ tion of the database is also made available as and practitioners’ ability to properly mea- Little Data Book on Financial Development. sure � nancial systems has been constrained Finally, readers are encouraged to go online by lack of comprehensive data. The data that and explore this large and interesting source are being made publicly available, together of data by themselves. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 41 Future versions of the Global Financial new trends or observations, and they will Development Report will revisit issues of focus on the relevant theme at hand, such as measurement of �nancial systems around the �nancial inclusion or capital market develop- world. They will also report on substantial ment, or other issues of policy relevance. Chapter 1 Annex: Overview of the Data Sources Underlying the Global Financial Development Database This annex is a summary. For more on the The Doing Business database (http:// Global Financial Development Database, www.doingbusiness.org/data), a part of the including the individual country data and Doing Business project, offers an expansive metadata, see this report’s Statistical Appen- array of economic data in 183 countries, dix and the Global Financial Development covering the period from 2003 to the pres- Report website at http://www.worldbank ent. The data cover various aspects of busi- .org/�nancialdevelopment. ness regulations, including those relevant to Database on Financial Development and � nancial sector development issues, such as Structure (updated November 2010). This enforcing contracts and obtaining credit. database was used a starting point for many IMF’s Access to Finance database (http:// of the basic indicators of size, activity, and fas.imf.org/) aims to systematically measure efficiency of financial intermediaries and access to and use of � nancial services. Fol- markets. Beck, Demirgüç-Kunt, and Levine lowing Beck, Demirgüç-Kunt, and Martínez (2010) describe the sources and construction Pería (2007), the database measures the reach of, and the intuition behind, different indica- of �nancial services by bank branch network, tors and present descriptive statistics. and availability of automated teller machines, Bankscope (Bureau van Dijk, http:// and does so by using four key financial w w w.bvdinfo.com / Products/Company- instruments: deposits, loans, debt securities Information/International/BANKSCOPE issued, and insurance. The website contains .aspx) was used to obtain and update data on annual data from about 140 respondents for banks. Bankscope combines widely sourced the six-year period, including data for all data with flexible software for searching and G-20 countries. analyzing banks. Bankscope contains com- The Global Financial Inclusion Index prehensive information on banks across the (Global Findex) is a new database of demand- globe. It can be used to research individual side data on �nancial inclusion, which docu- banks and � nd banks with speci�c pro� les ments financial usage across gender, age, and analyze them. Bankscope has up to 16 education, geographic regions, and national years of detailed accounts for each bank. income levels. The core set of indicators and Bloomberg (http://www.bloomberg.com/), subindicators of �nancial inclusion, based on Dealogic (http://www.dealogic.com/), and the Global Findex database, includes Use of Thomson Reuters Datastream (http://thom bank accounts (% of adults with an account sonreuters.com/products_services/� nancial/ at a formal � nancial institution, purpose of financial_products/a-z/datastream/) were accounts, frequency of transactions; % of used to obtain higher frequency data on adults with an active account at a formal stock exchange and bond markets that were �nancial institution, mode of access); Savings aggregate on a country level. (% of adults who saved in the past 12 months 42 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 using a formal financial institution, % of Bank for International Settlements (BIS) adults who saved in the past 12 months using (http://www.bis.org/) statistics were used for an informal savings club or a person outside the aggregate data on bond statistics, includ- the family, % of adults who otherwise saved ing domestic debt securities by residence in the past 12 months); Borrowing (% of and type of instrument (bonds and notes vs. adults who borrowed in the past 12 months money market instruments, issued by � nan- from a formal financial institution, % of cial and non�nancial corporations; based on adults who borrowed in the past 12 months publicly available or country-reported data). from informal sources, % of adults with an Domestic debt securities (Quarterly Review outstanding loan to purchase a home or an Table 16) for a given country comprise issues apartment); Payments (% of adults who used by residents in domestic currency targeted at a formal account to receive wages or govern- resident investors, whereas international debt ment payments in the past 12 months, % of securities (Quarterly Review Table 11) are the adults who used a formal account to receive ones targeted at nonresidents (a) in domes- or send money to family members living else- tic currency on the domestic market, (b) in where in the past 12 months, % of adults domestic and foreign currency on the inter- who used a mobile phone to pay bills or send national market, plus (c) the issues in foreign or receive money in the past 12 months); currency in the domestic market (further Insurance (% of adults who personally pur- information can be found in the Guide to chased private health insurance, % of adults the International Financial Statistics, http:// who work in farming, forestry, or �shing and www.bis.org/publ/bppdf/bispap14.htm). personally paid for crop, rainfall, or livestock Two different collection systems are used insurance). (s-b-s for international debt securities and Financial Soundness Indicators database aggregated data for domestic debt securities), (http://fsi.imf.org/), hosted by the IMF, dis- resulting in some possible overlap (between seminates data and metadata on selected domestic debt securities and international � nancial soundness indicators provided by debt securities) and inconsistencies (classi�- participating countries. cation of issuers). World Development Indicators (http:// Country authorities’ websites were used to data.worldbank.org/data-catalog/world- recon� rm and � ll in some of the gaps in the development-indicators) is the primary World data. Bank collection of development indicators, compiled from of�cially recognized interna- tional sources. It presents the most current NOTES and accurate global development data avail- 1. See http://www.copenhagenconsensus.com. able, and includes national, regional, and Among the top 30 solutions, micro�nance global estimates. was considered as a way to improve liveli- International Financial Statistics (http:// hoods of poor women, but this topic did not elibrary-data.imf.org / FindDataReports. make it to the top 10. aspx?d=33061&e=169393), from the IMF, 2. This is not the only approach to classifying the provides is a standard source of international functions provided by the �nancial system, statistics on all aspects of international and but it is not dramatically different from other approaches (such as Merton 1992; Merton domestic � nance. It reports, for most coun- and Bodie 2004), and it is an approach that tries of the world, basic financial and eco- �ts rather well with the large �nance litera- nomic data on international banking, money ture, including recent research. and banking, interest rates, prices, produc- 3. In the empirical literature, identifying the tion, international transactions, international impact of �nance has sometimes proved liquidity, government accounts, exchange challenging. Some of the early literature on rates, and national accounts. the subject requires the problematic iden- GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BENCHMARKING FINANCIAL SYSTEMS AROUND THE WORLD 43 tifying assumption that legal origins matter 8. See annex. for development only through their impacts 9. For example, many central banks around the on �nance. But subsequent papers have world publish reports focused almost exclu- tried more nuanced and more persuasive sively on �nancial stability (Čihák, Muñoz, approaches to identi�cation (such as Rajan Teh Sharifuddin, and Tintchev 2012). Simi- and Zingales 1998). larly, the IMF’s Global Financial Stability 4. The database builds on previous work within Report has a clear stability focus. There are, the World Bank Group, in particular the rel- however, many complementarities between evant papers by Beck, Demirgüç-Kunt, and �nancial stability, depth, access, and ef�- Levine (2000, 2010) and Beck and others ciency, as emphasized for instance in the (2006). The database also builds on Finan- World Bank–IMF’s Financial Sector Assess- cial Soundness Indicators database (http:// ment Program. fsi.imf.org/) and the Financial Access Survey 10. Ranciere, Tornell, and Westermann (2008), (http://fas.imf.org/). There are several major for example, �nd that countries that have sets of data. Chapter 1 annex provides a basic experienced occasional �nancial crises have, description of the data sources. The Statisti- on average, grown faster than countries with cal Appendix at the end of the report shows stable �nancial conditions. country-by-country data for 2008–10. 11. In contrast, ef�ciency seems surprisingly 5. The data source is IMF’s International Finan- relatively high in Middle East and North cial Statistics (see annex). Private credit iso- Africa, as well as in South Asia. This is in lates credit issued to the private sector and part because an important part of bank lend- therefore excludes credit issued to govern- ing goes to large companies and to the public ments, government agencies, and public sector, leading to relatively lower reported enterprises. Private credit also excludes credit margins. issued by central banks. 12. To put this in a more anthropomorphic per- 6. This report includes other measures as well. spective, the tallest adult person on earth is Also relevant are indicators of structure less than 5 times taller than the smallest per- within the individual �nancial segments, son (http://www.guinessworldrecords.com). such as the concentration ratios (Her�ndahl 13. For example, Rioja and Valev (2004) �nd index, shares of various types of �nancial (a) no statistically signi�cant relationship institutions in total assets and in GDP, and between �nance and growth at low levels of shares of individual markets in total market �nancial development, (b) a strong positive capitalization). Some of these measures (for relationship at intermediate levels of �nancial example, the percentage of assets of the three development, and (c) a weaker but still posi- or �ve largest �nancial institutions in GDP) tive effect at higher levels of �nancial devel- are important for the stability dimension, opment. Arcand, Berkes, and Panizza (2011) because they provide a rough approximation �nd that �nance actually starts having a nega- for the potential for impact in the case of a tive effect on output growth when credit to major �nancial disruption. the private sector exceeds 110 percent of GDP. 7. Financial structure differs markedly across Similarly, Cecchetti and Kharroubi (2012) economies. Over the full sample period, the �nd that the aggregate productivity growth annual average value of the �nancial struc- in an economy increases with private sector ture ratio is 279. Countries such as Austra- credit to GDP, but only up to a point; after lia, India, Singapore, and Sweden have this that point, increases in private sector credit ratio at or below 2.35 (10th percentile), while to GDP are associated with lower aggregate Bolivia, Bulgaria, Serbia, and Uganda are productivity growth. examples of countries where this ratio is over 356 (90th percentile). 2 The State as Regulator and Supervisor • Financial sector regulation and supervision are areas where the role of the state is not in dispute; the debate is about how to ensure that the role is carried out well. • A key challenge of regulation is to better align private incentives with public interest, without taxing or subsidizing private risk taking. Supervision is meant to ensure the implementation of rules and regulations. It needs to harness the power of market dis- cipline and address its limitations. • The �nancial crisis underscored limitations in supervisory enforcement and market dis- cipline. It emphasized the importance of combining strong, timely, anticipatory super- visory enforcement with better use of market discipline. It also highlighted the impor- tance of basics—solid and transparent legal and institutional frameworks to promote �nancial stability. In many developing economies that means that building supervisory capacity needs to be a priority. • Useful lessons can be learned by analyzing regulation and supervision in economies that were at the epicenter of the global �nancial crisis and those that were not. A new World Bank global survey, presented in this chapter, suggests that economies that suf- fered from the crisis had weaker regulation and supervision practices as well as less scope for market incentives than the rest. • This chapter reviews progress on regulatory reforms at the global and national levels, and identi�es advances made so far. Tracking changes during the crisis reveals that countries have stepped up efforts in the area of macroprudential policy, as well as on issues such as resolution regimes and consumer protection. However, the survey sug- gests that there is further scope for improving market discipline, namely disclosures and monitoring incentives. • The �nancial crisis has triggered a healthy debate on approaches to regulation and supervision among regulators, policy makers, and academics, leading to multiple pro- posals for further reforms. These proposals aim to limit regulatory arbitrage and make better use of regulatory resources. Common themes of these proposals are calls for more transparency and simpler regulation to enhance accountability, as well as for more proactive efforts to identify and address incentive problems. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 45 46 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 or years, developed economies enjoyed analyses can provide insight to policy mak- F stable macroeconomic conditions— often referred to as the Great Modera- tion—where developments in the � nancial ers and regulators designing reforms aimed at making � nancial systems more resilient and ef�cient. sector were often considered major contribu- After the onset of the crisis, there was tors to �nancial stability and thus to economic much talk about using the crisis to push growth.1 This has been followed by a phase through needed reforms. At the global level, of deep instability, in which major �nancial the G-20 has mandated the Financial Stabil- institutions collapsed and �nancial markets ity Board (FSB), after its transformation in malfunctioned. 2009, to promote the coordinated develop- Because of the central role that � nancial ment and implementation of effective regula- sectors play in market economies, govern- tory, supervisory, and other � nancial sector ments and major central banks intervened policies. 2 As part of this regulatory reform to avoid the collapse of the economic sys- agenda, the Basel Committee has prepared tem. Massive rescue packages and unortho- new capital and liquidity requirements for dox monetary measures were used to lower banks under the third Basel framework, market participants’ risk aversion to tolerable Basel III. On the national level, many econ- levels and to avert the worst scenarios. This omies have enacted or are considering new frantic activity presented a striking contrast laws and regulations in response to the les- with the sanguine attitude of investors and sons from the crisis. The crisis has also led supervisors in the years before the crisis, to an active policy debate among regulators, when the excesses of financial institutions policy makers, and academics, giving rise to were allowed to grow unhampered. multiple reform proposals. The global � nancial crisis that began in Much has been done, but is it appropri- 2007 and intensified with the collapse of ate? And will it be sufficient to reduce the Lehman Brothers in 2008 presented a major likelihood and severity of future financial test of the international architecture devel- crises? The key questions addressed in this oped to safeguard the stability of the global chapter are: What is the early thinking on financial system—and the architecture transforming regulatory practices around largely failed. the world? What are the speci�c issues for Although there is some consensus on emerging markets and developing economies attributing to financial markets an impor- (EMDEs)? What should be the role of the tant component of procyclicality, the reasons state as regulator and supervisor of the �nan- behind the absence of decisive preemptive cial sector?3 This chapter reviews some of supervisory action in the run–up to the cri- the lessons from the crisis and the responses sis are still a subject of debate. Some analysts proposed to address them, including those emphasize the weaknesses in policy making. reflected in the World Bank’s updated survey Others blame the trend toward deregulation. of bank regulation and supervision practices Yet others emphasize problems with incen- around the globe. The chapter summarizes tives in the �nancial markets and the regula- the progress made through recent regulatory tory and supervisory framework. reforms, as well as some promising new ideas Whatever the relative importance of these and reform proposals. factors, the crisis has thrust into the spotlight There is no major debate on whether the major shortcomings in regulation and super- state should be in regulation and supervision. vision, in the capacity of market discipline to Though the bene�ts of, for example, direct promote �nancial stability, and in the sound- state ownership of � nancial institutions are ness of national and international arrange- often disputed, the importance of the state ments for crisis management and surveil- regulating and supervising the �nancial sector lance, reopening debates and analyses in all is well established in the economic and �nan- these areas. The results of these debates and cial literature. The case for � nancial sector GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 47 regulation has been built around the fol- supervision (Laffont and Tirole 1993; Stigler lowing market failures:4 (a) anticompetitive 1971). According to some authors, that has behavior, (b) market misconduct, (c) infor- contributed to the � nancial crisis (Johnson mation asymmetries, and (d) systemic insta- and Kwak 2010). So the real question is how bility. These failures can impair the capacity best to ensure that regulation and supervision of � nancial markets to deliver ef�cient out- support sound �nancial development. comes and justify regulatory intervention if Shortcomings in private and public institu- the bene�ts outweigh the costs. The �rst two tions have to be addressed in a comprehen- market failures give rise to inef�ciencies that sive way, focusing on the roots of the crisis, need to be resolved through market regula- instead of its consequences. This approach tion, while the last two underpin the case for entails correcting weaknesses prevalent before prudential regulation. Regulation aimed at the crisis in many �nancial markets and insti- curbing anticompetitive behavior is needed tutions, including their supervisory bod- to foster an ef�cient allocation of resources ies (such as deep information asymmetries, and intermediation of funds. Market miscon- distorted incentives, defective governance duct regulation is needed to ensure that par- arrangements, and defective accountability). ticipants act with integrity and that suf�cient Doing so would improve market discipline by information is available to make informed providing information to market participants decisions. Information asymmetries have tra- and supervisory bodies if inef�cient risk tak- ditionally served as the main justi�cation for ing is suf�ciently and timely penalized before prudential regulation, but experiences in the the correction of excesses entails prohibitive financial crisis have raised the importance costs for the whole economic system. But it attributed to prudential regulation in pre- also implies that, given the limits of market venting systemic instability. discipline (due to, for example, coordination States’ regulation and supervision (usually problems, fallacy of composition, and herd done via autonomous or semiautonomous behavior), a complementary and equally nec- agencies) can improve welfare by provid- essary role is reserved for well-designed regu- ing the monitoring functions that dispersed latory and supervisory action. counterparts (in particular, depositors, Because of the crisis, much focus has been shareholders, and bondholders) are unable placed on regulating and monitoring sys- or unwilling to perform (Barth, Caprio, and temic risk. Indeed, using the latest round of Levine 2006). For example, Dewatripont the World Bank’s survey of regulatory and and Tirole (1994) develop a model of banks’ supervisory practices around the world, this capital structure, showing how optimal chapter con�rms that countries have stepped regulation can be achieved using a combina- up efforts on macroprudential policy, as well tion of basic capital adequacy requirements as on resolution regimes and � nancial con- with external intervention when those are sumer protection. All these efforts place more violated, with elements of market discipline demands on supervisors, which introduces being an important complement (though not greater burden in smaller and lower-income a substitute) to this regulation. Dewatripont economies where supervisory capacity is and Tirole, as well as other authors, see the already constrained. key challenge for regulation as providing Breakdowns in incentives are a unify- the right incentives to managers of � nancial ing theme when discussing the roots of the intermediaries. Regulation is seen as a “speed crisis (see Calomiris 2011; Demirgüç-Kunt bump�—a term coined by Joseph Stiglitz— and Servén 2010; Levine 2010, 2011; Rajan mitigating managers’ incentives to gener- 2010). Misaligned incentives in the �nancial ate pro�ts by rapid growth. Politicians and markets and in the regulatory and supervi- regulators are often subject to intense pres- sory framework were among the key factors sure from regulated � rms to modify regula- contributing to the crisis. Incentive confl icts tions, resulting in suboptimal regulation and help explain how securitization went wrong, 48 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 why credit ratings proved inaccurate, and and exit, healthy competition, and disclosure why the crisis cannot be blamed on mark- of quality information, combined with strong to-market accounting or an unexpected loss and timely supervisory action, are essential of liquidity. Contradictory market, bureau- in getting this balance right. But by revealing cratic, and political incentives undermined limits of market and regulatory discipline, �nancial regulation and supervision (Caprio, the crisis has led to a policy debate on the Demirgüç-Kunt, and Kane 2010). Insuf�cient right approach to regulation and supervision. incentives to enforce the existing rules (Barth, This ongoing debate continues to inform reg- Caprio, and Levine 2012a), combined with a ulatory reforms. lack of capacity, resulted in regulations that were not applied and supervisory powers that SOME LESSONS FROM THE were not used in the years leading to the cri- GLOBAL FINANCIAL CRISIS sis. Reducing the likelihood of future crises therefore requires addressing these incentive The global financial crisis—which began issues; along these lines the chapter discusses in 2007 and intensi�ed with the collapse of reform proposals that emphasize greater Lehman Brothers in September 2008—pro- transparency and disclosure, importance of vided a fundamental test of the international incentive compatibility in reforms, and sim- architecture, developed to safeguard the ple regulation. global �nancial system (Rajan 2010). Besides Transparency and disclosure of good macroeconomic factors, the main contribut- information, coupled with the right incen- ing factors identi�ed by scholars and policy tives, help make market participants behave makers include major regulatory and super- in ways consistent with the public interest. visory failures, together with failures in other Complicated regulation is not desirable, since parts of the � nancial system, such as gover- it is harder to implement and supervise, par- nance of private institutions, rating agencies, ticularly in smaller and less developed econo- accounting practices, and transparency. 6 mies with lower supervisory capacity. In most This section concentrates on the shortcom- middle-income and nearly all low-income ings identi�ed by the crisis in micropruden- economies, basic regulations, combined with tial regulation and supervision and market strong supervision and enforcement of trans- discipline. parency, are a better approach. Market disci- The chapter’s focus on shortcomings and pline is not a panacea, but it is an important areas for improvement does not mean that all ingredient in the regulatory and supervisory precrisis regulation failed, or that all super- mix. When regulation is ineffective, market visors performed uniformly badly. Supervi- discipline often breaks down, as illustrated in sion in many jurisdictions has actually per- the recent crisis. formed well. Within advanced economies, This chapter acknowledges the progress Australia, Canada, and Singapore have been made by recent global regulatory reforms, mentioned among examples of countries which include measures to address moral that withstood the global crisis rather well, hazard in too-big-to-fail institutions, abu- in part as a result of their prudent supervi- sive compensation policies, undue activity sory approaches (Palmer and Cerrutti 2009). with over-the-counter derivatives, and biased Also, many emerging markets and develop- credit ratings. It also discusses reform pro- ing economies had limited exposure to the posals that argue for taking these reforms risky behaviors that precipitated the crisis, further, as well as new approaches to regula- and most of these countries averted out- tion and supervision.5 right distress in the � nancial system partly The challenge of � nancial sector regula- because of their conservative prudential and tion is to align private incentives with the supervisory practices. Malaysia and Peru are public interest without taxing or subsidizing just two countries that have been praised for private risk taking. Threats of market entry their prudential policies.7 Nonetheless, some GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 49 emerging market countries suffered direct management purposes. These risks were espe- impacts of the crisis, especially in Europe and cially great in the case of large and intercon- Central Asia, which need to be seen against nected banks.8 Similarly, many of the credit a background of a heavy reliance on par- ratings produced in the run-up to the crisis ent bank funding and a buildup of funding failed to properly reflect systemic risk, rais- imbalances in the run-up to the crisis. ing questions about the role for credit ratings in the regulatory framework. These examples underscore the broader point: supervising the Weaknesses in regulation safety and soundness of individual �nancial and supervision institutions, though very important, does not A major weakness in the precrisis approach necessarily lead to a � nancial system that is to regulation and supervision was that it robust and stable, thus demonstrating the focused on risks to individual institutions and importance of following the example of those did not suf�ciently take into account what a supervisory bodies that have been tradition- confluence of risks implies for the � nancial ally incorporating systemwide considerations system as a whole (systemic risk). Thus the in their supervisory evaluations. crisis raised questions about the effectiveness The second weakness was that the pru- of narrowly focused microprudential policy dential approach suffered from regulatory in preventing systemic risk. Such approaches “silos� along functional and national lines. seek financial stability by focusing on the The approach focused on the risks in individ- safety and soundness of individual �nancial ual institutions and in their legal form, with institutions, with emphasis on institutions separate approaches for regulation and super- accepting retail deposits. Because questions vision of banks, insurance, and securities have been raised on whether this amounts not complemented by strong oversight at the to �nancial stability, many supervisory bod- �nancial group level and systemic level. This ies, in their risk evaluations, take into con- approach allowed transactions to be chan- sideration systemwide developments to assess neled through the entities that were subject to their potential repercussions for individual weaker regulation, and for transactions to be institutions and for the whole system. conducted in the gaps between the regulatory If an institution or market fails, the impact silos to avoid regulation altogether. The rapid on the financial system and economy can growth of the shadow banking system was a exceed the losses sustained by individual case in point. The emergence of the shadow institutions or markets. A microprudential banking system in the United States needs approach that sets regulations and conducts to be seen against the background of differ- supervision to limit the risk in an individual ent regulatory approaches toward deposit- institution does not necessarily limit the risk taking banks and other less-regulated seg- to the financial system. For example, the ments of the financial system. By drawing push for Basel II implementation led in some a “line in the sand� and separating deposit- jurisdictions to increased emphasis on banks’ taking banks from other entities (including internal models and on credit rating agencies investment banks), and by placing risky activ- to evaluate risk. But the implementation of ities in separate legal entities such as special- banks’ internal models focused on the risks in purpose vehicles, policy makers expected that the banks’ own balance sheets (private risks), the prudentially regulated segment—primar- and did not adequately take into account the ily the deposit-taking banking sector—would risks posed by individual banks to the �nan- be isolated from difficulties in the unregu- cial system as a whole (systemic risk, or pub- lated segment of the financial system that lic risk). In addition to this regulatory issue, was populated by well-informed professional many �nancial institutions also suffered from investors. In addition to these functional poor risk management surrounding the mod- silos, there are also national silos: whereas the els that they used for their own internal risk regulated entities have become increasingly 50 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 global, � nancial regulation is national, and adequately capture lending concentration, cross-border regulatory cooperation—despite excessive maturity transformation, and— some progress—still faces serious incentive especially in small, open developing econo- problems and broke down in the face of crisis mies—the indirect credit risk associated with pressures (box 2.1). foreign exchange exposures of unhedged bor- Third, some microprudential regulations rowers. Moreover, the rules encouraged risk were poorly designed, contributing to sys- transfers to entities that were legally separate temic risk. The Basel capital adequacy mea- but not separately capitalized. When market sures considerably misrepresented the sol- sentiment with regard to complex, structured vency of banks. There were various reasons products started to deteriorate, parent banks for this, including the use of risk weights that felt compelled for reputational reasons to underestimated the riskiness of assets such shoulder the losses in those entities.10 Risk as mortgages and sovereign debts, the differ- was also transferred in nontransparent ways ent treatment under the Basel rules of assets owing to the rapidly increasing trade in com- held in banking books and those in trading plex, structured financial products, which books, and the definition of capital.9 The often underwent successive repackaging and regulatory framework has also struggled to sale. Because of these layers of opacity, the BOX 2.1 Distorted Incentives: Subprime Crisis and Cross-Border Supervision Distorted incentives at various levels were a main cies were aware of the growing fragility of the � nan- cause of the � nancial crisis. For example, the poli- cial system associated with their policies during the cies to promote home ownership in the United decade before the crisis, yet they chose (under pres- States created perverse incentives within of�cial and sure from the industry and politicians) not to modify quasi-of� cial agencies, contributing to the buildup those policies. of exposures in subprime mortgages, and to for- Distorted incentives have also played an impor- bearance in regulation and supervisory oversight tant role in regulation and supervision of � nancial (Calomiris 2011; Wallison and Calomiris 2009). institutions across several jurisdictions. Supervisory Regulation also played a role in distorting incen- memorandums of understanding and supervisory tives for rating organizations to conduct appropriate colleges have been used to strengthen cross-border due diligence. Other issues included moral hazard supervision. And some of the colleges have been associated with too-big-to-fail policies (Ötker-Robe useful in good times. But in times of crisis, cross- and others 2011), adverse selection associated with border cooperation almost always breaks down, as the rules for assessing the creditworthiness of bor- during the Fortis failure in 2008. This and many rowers, and the principle or agent problems within other examples confirm that the supervisory task � nancial institutions, related to the nature of own- sharing anchored in the Basel Concordat of 1983 is ership and the structure of executive compensa- not crisis-proof, reflecting misalignments in underly- tion that favored risk taking and higher short-term ing incentives (D’Hulster 2011). Without an agreed returns to the longer-term detriment of shareholders. resolution and burden-sharing mechanism and with Levine (2010) � nds that the design, implementa- deteriorating health of the bank, incentive confl icts tion, and maintenance of � nancial policies in 1996– escalate and supervisory cooperation breaks down. 2006 were primary causes of the � nancial system’s Thus, good practices for cooperation among super- demise. He rejects the view that the collapse was visors are insuf�cient to address the incentive con- only due to the popping of the housing bubble and fl icts. D’Hulster (2011) calls for rigorous analysis the herding behavior of � nanciers selling increas- and review of the supervisory task sharing, so that ingly complex and questionable � nancial products. the right incentives are secured during all stages of Rather, the evidence indicates that regulatory agen- supervision GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 51 extent to which individual financial insti- climate dampened incentives for analysts of tutions were exposed to these toxic assets financial stability to dig deeper and ques- became increasingly nontransparent. So tion the adequacy of the information and though individual banks’ regulatory capital the underlying benign assumptions on which positions appeared sound, some were not, their analysis was based. The crisis also and the capital adequacy of the �nancial sys- revealed severe shortcomings in resolution tem was weakened, increasing systemic risk. frameworks, especially for large financial Fourth, implementation of the rules was institutions active in multiple jurisdictions. constrained by the capacity and incentives of regulators and supervisors. Supervisory Capacity constraints in regulation resources became stressed as � nancial insti- and supervision tutions, instruments, and regulation grew more complex. Information on exposure and In many developing economies and emerg- risk became harder to compile as � nancial ing markets, weak supervisory capacity and groups became more complex and intercon- a lack of regulatory independence are at nected, with operations both locally and least as important as gaps in the regulatory overseas, and spanned many business lines. framework in explaining fragility. Nearly all The regulators also faced conflicts of interest, assessments of developing economies under with some mandated to promote financial the Financial Sector Assessment Program system development as well as supervise it.11 (FSAP) �nd capacity constraints in regulation Some regulators lacked independence, and and supervision. In many of these economies, even supervisors that were legally indepen- licensing and closure decisions are still vested dent found it dif�cult to withstand pressures with ministries of � nance rather than bank from the industry.12 The “revolving door� regulators, which gives rise to a risk of politi- moving staff between supervisory authorities cal interference in these critical decisions, as and the � nancial industry—though hard to well as delays in early intervention in the case avoid completely because having an industry of fragile and weak banks. In many countries background and familiarity with financial in Sub-Saharan Africa, for example, supervi- instruments helps in understanding risks— sory resources are limited, including quali�ed resulted in perceptions of conflicts of interest staff and the availability of analytical tools for some supervisors (Kane 2007). It has also and skills. Supervisory processes focus on been suggested that supervisors exercised compliance with regulatory standards but are regulatory forbearance on the treatment of not set up to identify and manage the chang- subprime mortgages because of political con- ing risks in banking systems. In addition, the siderations. Across borders, misalignments in ability to monitor risk at the institutional and incentives between home and host supervi- systemic level is hampered by insufficient sors impede cross-border sharing of supervi- quality in data and reporting processes. sory information (box 2.1). The West African Economic and Mon- Finally, shortcomings in crisis manage- etary Union Banking Commission, for exam- ment and surveillance compounded several of ple, lacks suf�cient power to enforce correc- the problems identi�ed above. In particular, tive measures in cases of noncompliance with information gaps and asymmetries limited regulations, a situation that has only recently the capacity of �nancial stability assessors to begun to be addressed by political authorities monitor exposures, risk transfers, and threats (Beck and others 2011). Ill-suited regulations, to systemic stability. It was dif�cult to know such as on preapproval of loan applications, how the failure of one institution would are often ignored, undermining supervisory affect others. Systemically important seg- discipline. To ensure certainty and supervi- ments of the � nancial system were not cov- sory discipline, such outdated regulations ered by surveillance and crisis management should be dropped and the focus shifted to arrangements. The political and economic enforcing meaningful ones. Suggestions to 52 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 TABLE 2.1 Examples of Weak Supervisory Capacity Identi�ed in the FSAP Bangladesh FSA 2010 “Although the government has improved the prudential regulatory and supervisory framework for banks, further improvement would be needed to bring the system up to international standards. Loan classi�ca- tion, provisioning, rescheduling, and even capital in banks appears uneven and needs strengthening to be brought to international standards.� Barbados FSA 2009 “To assure continued �nancial stability and competitiveness, important regulatory and remaining super- visory weaknesses in the �nancial system also should be addressed. The prudential oversight of the banking sector could be strengthened by enhancing supervisory capabilities, accelerating the transition to risk-based banking supervision, tightening supervision of large exposures and exposures to related par- ties, improving the criteria for asset classi�cation and provisioning, improving consolidated supervision for banking groups and regional �nancial conglomerates, and establishing more active home-host supervisory cooperation arrangements.� Burundi FSA 2009 “The supervisory approach adopted by the [central bank] BRB is still largely based on monitoring compliance with laws and regulations, despite the fact that the international trend favors the risk-based approach. The level of supervision could also be stepped up by developing closer surveillance methods so as to have greater visibility with respect to the major risk areas and fragilities of each establishment.� Haiti FSSA 2008 “Supervisory procedures appear largely adequate, but the capacity of the supervisory function should be improved. Its operational autonomy can be strengthened in the context of greater central bank indepen- dence. The current staf�ng of the DBS seems insuf�cient, and its budget needs to be increased in a sustained manner, while supervisory staff skills should be upgraded through training focused on banking and risk management.� Lithuania FSA 2008 “Resolving these issues requires an urgent review of supervisory arrangements for Lithuanian �nancial markets. Future supervisory arrangements should be designed with the objectives to (i) strengthen capacity to supervise the interactions of banks with their related entities; . . .� Mozambique FSA “(ii) increase capacity of BM’s supervisory staff, especially in the areas of 2009 on-site inspections and risk management� Papua New Guinea “Enhance the capacity of the supervisory staff through training, so that the BPNG can move to full FSSA 2011 risk-based supervision. The BPNG should assert itself more rigorously to ensure full compliance with the supervisory regime. It is good to seek consensus but less desirable to leave necessary prudential statements in draft for over �ve years (as has happened with the revised large exposures prudential standard). With new tools and the complete suite of regulations recommended above, the BPNG will be in a position to ensure good risk management practices are followed and move to risk-based supervision. It will need further assistance to build capacity to achieve this, so that its committed and professional staff can work in partnership with �nancial businesses to focus on identifying and managing the key risks in the sector.� Source: World Bank FSAP website (http://worldbank.org/fsap). improve supervisory capacity are among the economy poses signi�cant risks if steps are most common recommendations in FSAPs. taken too fast or not sequenced appropriately, For example, the 2011 FSAP assessment on given “considerable capacity constraints for El Salvador noted that, despite an ambitious quali�ed personnel in � nancial institutions project to move toward risk-based supervi- and in financial sector supervision.� (For sion, “it is essential to further upgrade super- additional examples, see table 2.1.) visory capacity, both in quantitative and in These deficiencies weigh increasingly in qualitative terms� and that “the existing a globalized world that is moving toward regulatory framework has signi�cant gaps.� more complex regulations. The survey results In another recent example, the 2011 FSAP show that a move toward Basel II and Basel assessment on Rwanda warns that the ambi- III and the increased complexity of postcrisis tious agenda to improve access to finance regulations are adding pressures on resource and provide more long-term �nancing to the requirements. In line with these observations, GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 53 a recent joint report by FSB, International but also some emerging markets) relied on Monetary Fund (IMF), and World Bank market discipline to safeguard financial (2011), endorsed by G-20 leaders in the soundness and stability. Market discipline Cannes Summit, calls for countries that have requires that markets objectively assess the less internationally integrated � nancial sys- risks and value of financial instruments tems or substantial constraints in supervisory and financial institutions, and price them capacity to focus on reforms to ensure compli- accordingly. ance with the more basic principles of sound The crisis has made it clear that such regulation before considering a move to the objective market valuation does not always Basel II and Basel III standards. The same occur. For example, in the �rst half of 2007, report also calls for further development of the stock market valuation of Irish banks supervisory capacity in developing economies was at or close to their long-term maximum. through targeted and well-coordinated tech- Also, spreads between Greek debt and Ger- nical assistance and other capacity-building man bund were very small for years, as were activities. These efforts need to be part of a the spreads of other euro area countries, broader, sustainable strategy to overcome not providing much indication of what was capacity constraints in regulation and super- to take place. Similarly, credit default swap vision in developing economies. spreads for southern European countries Strengthening of supervisory capacity were negligible for many years compared and improvements in regulations are areas to their peers in Northern Europe, which where donors can provide help. One of the allowed some countries to go on a lending tools in this regard is the Financial Sector binge for many years. It was only through Reform and Strengthening (FIRST) Initia- the economic slowdown during the crisis and tive, a multidonor grant facility managed by escalation of events in Greece that market the World Bank. Strengthening supervisory perceptions started to change substantially capacity and improving regulations account and credit default swap spreads on govern- for more than a half of FIRST’s recent proj- ment paper shot up (and became more closely ects. Individual donors have also provided correlated with bank risk measures). support directly to various projects aimed at These observations are reminders of the strengthening capacity in regulatory agencies. tendency of economic agents and the � nan- For example, the State Secretariat for Eco- cial system to be overly tolerant of risk in nomic Affairs (SECO) in Switzerland has an credit cycle upswings and excessively risk extensive program of banking sector training averse in downswings. Put differently, the in Vietnam, which includes practical train- failure of market discipline needs to be seen ing for the regulatory body, complemented against the collective tendency of � nancial by a train-the-trainer project for Vietnam’s markets to underestimate risks in boom times two largest universities, management train- and overestimate it in times of bust. ing for bank managers, and technical assis- Still, the failures of market discipline in tance in modernizing the central bank and its the run-up to and during the crisis do not strategy to develop the banking sector. (For mean that financial markets did not pro- information on SECO’s programs in the area vide useful signals. As shown for example of capacity building, see the relevant back- by Haldane (2011), equity markets were dif- ground materials at http://www.worldbank ferentiating between banks in trouble and .org/�nancialdevelopment.) those that were not several years before the financial crisis, when intervention could have vastly reduced the subsequent costs. Weaknesses in market discipline Papers that examined previous crises find and the role of incentives similar relationships. Markets can provide Before the crisis, � nancial systems in many useful signals, but the real question is, when jurisdictions (especially advanced economies, do they provide such signals? And how can 54 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 these signals be caught by market agents and ample evidence that large � nancial institu- supervisors? tions enjoyed an implicit market subsidy The idea of market discipline rests on prior to the crisis, consistent with the moral the notion that, given the right incentives hazard associated with too-big-to-fail poli- and information, rational market par- cies (Goldstein and Véron 2011; Ötker-Robe ticipants would penalize institutions that and others 2011). They could take on more take on excessive risk. The Basel II capital risk and expand their balance sheets rapidly framework sought to expand the role of to boost short-term pro�ts without having market discipline in the regulatory frame- to increase their capital. Indeed, one long- work. Rating agencies were given a role in known problem that came to the fore dur- the evaluation of the risks in the portfolio ing the crisis was the too-big-to-fail problem under so-called Pillar I (Minimum Capi- (Rajan 2010), in which institutions that are tal Requirements), and an explicit role for too large or too interconnected to be allowed market discipline was introduced under to fail are given more favorable treatment so-called Pillar III (Market Discipline). during crises. This condition severely dis- Reliance on markets to safeguard � nancial torts their risk-taking incentives during nor- stability was also evident well beyond the mal times by undermining market discipline Basel rules. It was reflected, for example, in to be exercised by their unsecured debtors the limited attention by of�cials to the risks and transaction counterparts. The implicit posed by unregulated entities in the shadow assumption is that given the prohibitive banking system or to the lack of informa- consequences of failure of these large and tion on risk transfers.13 The assumption was highly interconnected � nancial institutions, that the regulated � nancial institutions have policy makers will do whatever it takes to incentives to be prudent in managing expo- prevent these institutions from collapsing. sures to their counterparties. This expectation translates into a fund- But market discipline in the run-up to the ing advantage and skews incentives toward crisis did not work well—mainly because leveraging and risk taking.14 The problems incentives of market players were distorted associated with the too-big-to-fail condition and they did not have access to the needed are often exacerbated by shortcomings in the information. Many institutions and instru- resolution framework for failing financial ments were allowed to grow highly com- institutions, especially when they operate plex and nontransparent. Information on across borders. In most cases, these weak- interconnections and exposures of finan- nesses originate in insolvency frameworks cial institutions was lacking. The increasing that do not distinguish between financial use of over-the-counter � nancial derivatives companies—especially banks—and nonfi- enabled � nancial institutions to transfer or nancial corporations. Part of the answer is to to take on risk in nontransparent ways, rap- strengthen bank resolution frameworks and idly and without the necessary capital for to put greater emphasis on resolvability in ultimate risk-taking institutions to be able to the context of ongoing supervision, also by withstand losses when they became appar- demanding that banks establish their resolu- ent. In many cases, assessment of the entities tion plans (also called living wills).15 and instruments was outsourced to special- Executive compensation is one aspect of ized institutions, such as rating agencies and these inadequate governance structures that auditing � rms, while the incentives of these attracted close attention during the crisis. agencies to conduct due diligence were often Before the crisis, compensation was gener- compromised by conflicts of interest. As a ally a no-go area for supervisors. Supervisors result, effective market discipline could not rarely had adequate powers to address issues function. related to risk and compensation structures. The lack of effective market discipline The collapse of banks with executives who also resulted from moral hazard. There is were allegedly paid for performance clearly GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 55 raised many questions about the link between study by Ellul and Yerramilli (2010) �nd that executive pay and risk taking. Philippon and commercial banks with a strong commitment Reshef (2009) show that, whereas in 1980 to risk management—as proxied by the ratio bankers made no more than their counter- of the compensation of the chief risk of�cer parts in other parts of the economy, by 2000 relative to the compensation received by the wages for employees in the � nancial sector chief executive officer—fared much better were 40 percent higher than for those with during the subprime crisis than those with the same formal quali�cations in other sec- weaker commitments to risk management. tors. The last time such a discrepancy was Although risk managers, acting in the inter- observed was just prior to the Great Depres- est of their stockholders, are the � rst line of sion—an irony that has not been lost on crit- defense against imprudent investing, pruden- ics of bank compensation, who range from tial regulation and supervision are the second regulators to the Occupy Wall Street protest- line of defense.17 ers. But the level of compensation alone may not be the real problem. Leading economists, REGULATION AND SUPERVISION such as Alan Blinder and Raghuram Rajan, IN CRISIS VERSUS NONCRISIS have emphasized that a much more important COUNTRIES (and dif�cult) question to answer is how the structure of performance pay may encourage How do regulatory and supervisory practices excessive risk taking at all levels of the insti- in countries at the epicenter of the global tution, from traders and underwriters right financial crisis differ from the rest of the up to the �rm’s chief executive of�cer and the world? What can one learn from those dif- board of directors. ferences? And how have the actual national But how exactly the structure of executive regulatory and supervisory practices changed pay affects risk taking is still a topic of heated in recent years as a result of Basel II and debate. Some have argued—even before the other initiatives and in response to the global crisis—that executive compensation at banks �nancial crisis and its aftermath? To answer must have several features to discourage these questions, this section provides a status short-term and excessive risk taking: pay- update and analysis of the regulatory and ing bankers with equity or stock options, for supervisory practices around the world. The instance, should ensure that if the �rm’s mar- section relies on the recently updated data ket value gets wiped out, the same fate awaits from the World Bank’s Bank Regulation and the paycheck of its senior management. But Supervision Survey (see box 2.2 for an intro- matters may be more complex. Incentive duction to the survey; all the country-level schemes may unduly emphasize immediate data are publicly available at http://www revenue generation over a prudent long-term .worldbank.org/�nancialdevelopment). assessment of credit risk (as was likely the case To examine the regulatory and super- in mortgage lending); and bonuses awarded visory differences between crisis and noncri- to managers today may entail risks for the sis countries, this section and the paper by institution that do not become apparent until Čihák, Demirgüç-Kunt, Martínez Pería, and much later. Both aspects of bank compensa- Mohseni (2012) compare country of�cials’ tion have become the focus of increased regu- inputs from the World Bank’s Bank Regula- lation intended to discourage bank executives tion and Supervision Survey and juxtapose from excessive risk taking. But policy mak- them with the countries’ experience during ers’ understanding of how incentives at banks the crisis. Specifically, to distinguish crisis translated into actual risk-taking behavior is and noncrisis countries, this chapter uses an still limited, and regulators struggle to come existing and often-used database of bank- up with rules that can rein in reckless risk ing sector crises, last updated in Laeven and taking without extinguishing banks’ abil- Valencia (2012). Laeven and Valencia use a ity to reward actual performance.16 A recent set of well-de� ned criteria to assess the 143 56 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 2.2 What Is in the World Bank’s Bank Regulation and Supervision Survey? An important input into this chapter was the update supervisors and researchers, and detailed guidelines of the World Bank’s Bank Regulation and Supervi- were drafted by a senior banking regulator to make sion Survey. The survey is a unique dataset of bank the questions more speci�c and clearer. regulation and supervision around the world. In the Data for 143 jurisdictions (see map B2.2.1) for early 2000s, the World Bank created a database of 2010–11 allow comparisons across countries and bank regulation and supervision around the globe with the previous three rounds. The survey consists (Barth, Caprio, and Levine 2001). The second, of information from over 730 questions and sub- updated iteration of the database was issued in 2003, questions in 14 sections. About half of the questions and the third version was issued in 2007. The survey are the same as in the previous three survey rounds has been widely used in research and policy work. (for reasons of comparability), and about half are The current round of the survey provides compre- new (mostly on macroprudential issues, consumer hensive information on the state of regulation and protection, and Basel II implementation). supervision around the world as of 2011. It is the The survey contains questions in the following � rst comprehensive look at regulation since 2007. 14 sections: 1. Entry into banking; 2. Ownership; Some of the questions have been kept unchanged 3. Capital; 4. Activities; 5. External auditing from the 2007 survey, for reasons of comparability. requirements; 6. Bank governance; 7. Liquidity and Other questions have been reformulated to result in diversi�cation requirements; 8. Depositor (savings) more precise answers. Several questions were added, protection schemes; 9. Asset classification, provi- in particular on consumer protection and macro- sioning, and write-offs; 10. Accounting/information prudential regulation. The survey involved a major dis closure; 11. Discipline/problem institution exit; effort to ensure the consistency of responses across 12. Supervision; 13. Banking sector characteristics; countries. Its design involved expertise of both 14. Consumer protection. MAP B2.2.1 Coverage of the 2011 Bank Regulation and Supervision Survey covered by survey not covered by survey Source: World Bank. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 57 TABLE 2.2 Differences between Crisis and Noncrisis Countries (percent, unless indicated otherwise) Noncrisis Crisis Difference Question (Yes/No) % Yes % Yes P-valuea Is tier 2 allowed in regulatory capital? 85.3 100.0 0.07 Is tier 3 allowed in regulatory capital? 27.5 80.0 0.00 Was advanced internal ratings-based approach offered to banks in calculating capital 44.7 94.7 0.00 requirements for credit risk? Are there minimum levels of speci�c provisions for loans and advances that are set 78.8 30.0 0.00 by the regulator? Can the supervisory agency require commitment/action from controlling 83.2 65.0 0.06 shareholder(s) to support the bank with new equity? Are asset/risk diversi�cation requirements employed to oversee more closely and/or 39.1 13.3 0.07 limit the activities of large/interconnected institutions? Is there a regulatory requirement for general provisions on loans and advances? 69.9 25.0 0.00 Do you have an asset classi�cation system under which banks have to report the 89.3 65.0 0.00 quality of their loans and advances using a common regulatory scale? Does accrued, though unpaid, interest/principal enter the bank’s income statement 23.9 50.0 0.02 while the loan is classi�ed as nonperforming? Is there a regulatory limit on related party exposures? 97.3 84.2 0.01 Are external auditors subject to independent oversight by banking supervisory 25.9 9.5 0.10 agency? In cases where the supervisor identi�es that the bank has received an inadequate 93.6 75.0 0.01 audit, does the supervisor have the powers to take actions against the bank? Question (Quantitative) Median Median P-valueb Risk-based capital ratio of banking system (end of 2008) 14.9 12.8 < .01 Risk-based capital ratio of banking system (end of 2009) 16.5 14.6 < .01 Risk-based capital ratio of banking system (end of 2010) 16.5 15.9 < .05 Tier 1 capital ratio of banking system (end of 2008) 12.9 9.8 < .01 Tier 1 capital ratio of banking system (end of 2009) 14.0 11.6 < .01 Tier 1 capital ratio of banking system (end of 2010) 14.6 12.0 < .01 Source: World Bank’s Bank Regulation and Supervision Survey (database), 2011 data. See also www.worldbank.org/�nancialdevelopment. Note: The following countries included in the Bank Regulation and Supervision Survey had a systemic banking crisis between 2007 and 2011 according to Laeven and Valencia (2012): Austria, Belgium, Denmark, Germany, Greece, Iceland, Ireland, Kazakhstan, Latvia, Luxembourg, Netherlands, Nigeria, Spain, Ukraine, United Kingdom, and United States. The following countries had borderline systemic crises: France, Hungary, Italy, Portugal, the Russian Federation, Slovenia, and Switzerland. a. For questions with yes or/no responses, Student’s t-test was used to test for the equality of the means (percentage of “yes� responses) between crisis and noncrisis. b. For quantitative questions, Stata’s cendif utility was used to test for the equality of the medians between crisis and noncrisis. See Newson (2002) for more on this utility. countries covered by the World Bank’s Bank • The crisis countries allowed for less strin- Regulation and Supervision Survey. Of those, gent de� nitions of capital and had lower they identify 16 countries—mostly advanced actual tier 1 capital. Whereas 80 percent economies, but also some EMDEs—that of crisis countries allowed tier 3 in regula- experienced a systemic banking crisis tory capital and 100 percent allowed tier between 2007 and 2011 and 7 countries that 2, only some 28 percent among noncrisis experienced a borderline systemic crisis in the countries allowed tier 3 and 85 percent same period (see table 2.2 for a list). All the allowed tier 2. other countries in the database are treated as • The median level of tier 1 capital to assets noncrisis countries. was 13 percent for noncrisis countries in When one uses the data from the survey to 2008, compared to 10 percent among cri- compare regulation and supervision in crisis sis countries. countries to the rest of the world (table 2.2), • The share of crisis countries that allow the following differences stand out:18 banks to calculate their capital require- 58 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 ment for credit risk based on the banks’ efforts to coordinate and monitor progress in internal ratings models is 95 percent, strengthening �nancial regulation.20 about twice as a large as in the rest of the The expansion of the Financial Stability world (45 percent). Forum and its reestablishment as the FSB, • Regulators and supervisors in crisis coun- with a broader mandate in the area of �nan- tries were less able to require bank share- cial stability and the coordination of the holders to support the bank with new global financial sector reform, is the most equity. Although the regulator had this important development in the international power in 83 percent of noncrisis countries, architecture of � nancial regulation in recent this was true in only 65 percent of crisis years. countries. As part of the regulatory reform package • Although almost 70 percent of noncrisis coordinated by the FSB, the Basel Commit- countries had a regulatory requirement for tee on Banking Supervision (BCBS) intro- general provisions on loans and advances, duced a new global regulatory framework only 25 percent of crisis countries had such for bank capital adequacy and liquidity. provisions in place. Close to 90 percent This new framework, called Basel III, aims of noncrisis countries had an asset clas- to strengthen bank capital requirements and si�cation system under which banks had introduces new regulatory requirements on to report the quality of their loans using bank liquidity, while limiting bank leverage. a common regulatory scale, while 65 per- This proposal (summarized in table 2.3) con- cent of crisis countries had such systems tains many useful steps that help to address in place. Half of crisis countries allowed the problems highlighted by the crisis. For accrued though unpaid interest and prin- example, the elimination of tier 3 and most cipal to enter the bank income statement of tier 2 capital and other measures should when loans are nonperforming, but only help in raising the quality of banks’ capital, 24 percent of noncrisis countries allowed while the signi�cantly more stringent regula- this. tory treatment of securitizations would result • Crisis countries have relatively less strict in more capital to be held against the credit limited party exposure limits and audit risk of these positions. procedures. In addition to the steps taken by the Basel • Finally, crisis countries also had less scope Committee, the FSB has come up with a for market discipline, in terms both of pro- broad range of proposals that, in some cases, viding incentives to monitor and of ensur- are meant to address certain failures in mar- ing quality of information to enable accu- ket discipline that became strikingly apparent rate monitoring; for example, they had during the crisis. In particular, the FSB has more generous deposit insurance coverage speci�cally addressed issues relating to com- and lower quality of �nancial information pensation practices, credit rating agencies, made publicly available. � lling of information gaps, and methods to deal with the too-big-to-fail issue. As regards compensation practices, com- pensation at significant financial institu- GLOBAL REGULATORY tions was recognized by FSB as one factor REFORMS19 among many that contributed to the � nan- In response to the deficiencies in financial cial crisis that began in 2007. A part of the regulation revealed by the global financial official response was the issuance of FSB’s crisis, leaders of the world’s major econo- Principles for Sound Compensation Practices mies designated the G-20 to be the premier (FSB 2009b) and the related Implementation forum for international economic coopera- Standards (FSB 2009c). The stated aim of tion. They also established the FSB to include these principles and standards is to enhance major emerging economies and welcomed its the stability and robustness of the � nancial GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 59 TABLE 2.3 Summary of the Basel III Framework Proposed changes Speci�c steps Raising quality, consistency, and The predominant form of tier 1 capital must be common shares and retained transparency of the capital base earnings. Tier 2 capital instruments will be harmonized. Tier 3 capital will be eliminated. Strengthening risk coverage of the Promote more integrated management of market and counterparty credit risk. capital framework Add the credit valuation adjustment risk due to deterioration in counterparty’s credit rating. Strengthen the capital requirements for counterparty credit exposures arising from banks’ derivatives, repo and securities �nancing transactions. Raise the capital buffers backing these exposures. Reduce procyclicality. Provide additional incentives to move over-the-counter derivative contracts to central counterparties. Provide incentives to strengthen the risk management of counterparty credit exposures. Raise counterparty credit risk management standards by including wrong-way risk. Introducing a leverage ratio as a The committee therefore is introducing a leverage ratio requirement that is intended supplementary measure to the Basel II to put a floor under the buildup of leverage in the banking sector. risk-based framework Introduce additional safeguards against model risk and measurement error by supplementing the risk-based measure with a simpler measure that is based on gross exposures. Reducing procyclicality and promoting Dampen any excess cyclicality of the minimum capital requirement. countercyclical buffers Promote more forward-looking provisions. Conserve capital to build buffers at individual banks and the banking sector that can be used in stress. Protecting the banking sector from Requirement to use long-term data horizons to estimate probabilities of default. periods of excess credit growth. Downturn loss-given-default estimates, recommended in Basel II, to become mandatory. Improved calibration of the risk functions, which convert loss estimates into regulatory capital requirements. Banks must conduct stress tests that include widening credit spreads in recessionary scenarios. Promoting stronger provisioning practices Advocate a change in the accounting standards toward an expected loss approach. (forward looking provisioning) Introducing a global minimum liquidity A 30-day liquidity coverage ratio requirement underpinned by a longer-term structural standard for internationally active banks liquidity ratio called the net stable funding ratio. Source: Based on Basel Committee on Banking Supervision 2011a. system. They are not to be used as a pretext stability–threatening herding effects that cur- to prevent or impede market entry or market rently arise from credit rating thresholds being access. hardwired into laws, regulations, and market T he reform of credit rating agen- practices, the FSB has drawn up principles cies is another important part of the FSB to reduce reliance on credit ratings in stan- agenda. In an effort to reduce the � nancial dards, laws, and regulations (FSB 2010b). The 60 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 2.3 Reforming Credit Rating Agencies Credit rating agencies have not met the expectations Though references to risk ratings in regulations placed on them by investors and policy makers. For are undesirable, the alternatives have drawbacks. example, empirical evidence suggests that ratings In particular, bank models of risk assessment have have often been lagging indicators that show at best proved even less reliable than credit ratings, includ- only information already known by the market (see, ing in the largest banks where risk management for example, Afonso, Furceri, and Gomes 2011; was widely believed to be most advanced (see, for Arezki, Candelon, and Sy 2011). instance, UBS 2008). Replacing references to rat- Much of the postcrisis debate on credit rating ings with references to market-based risk indicators agencies has revolved around confl icting interests could sharply increase procyclicality because such because of the commingling of rating and advisory indicators are typically much more volatile than services. The issue is that many larger credit rating credit ratings. As a result, it is to be expected that agencies offer “credit rating advisory services� that ratings will be complemented with other measures essentially advise an issuer on how to structure its of risk. But eliminating references to credit ratings in bond offerings and “special purpose entities� so as regulations is impractical and undesirable given the to achieve a given credit rating for a certain debt lack of proper alternatives. Moreover, contemplating tranche. This creates potential confl icts of interest, such steps given the current period of market stress of course, because credit rating agencies may feel could increase short-term volatility. obligated to provide issuers with those ratings if issu- Véron and Wolff (2012) argue that the role of ers followed their advice on structuring the offering. credit ratings in regulation should be reduced—but Some credit rating agencies avoid this conflict by that eliminating it entirely would have important refusing to rate debt offerings for which its advisory downsides, at least in the short term. Transferring services were sought. This was an important reason the responsibility for ratings to public authorities is why many of the risky, complex structured � nancial unlikely to be a good alternative because of inherent products had very favorable ratings. confl icts of interest. Goodhart (2008) and Caprio, Credit rating agencies derive some of their impor- Demirgüç-Kunt, and Kane (2010) suggest that credit tance from the fact that the regulatory system relies rating agencies need to bond the quality of their on their assessments. This reliance is observed in work by subjecting it to effective independent review bank regulation, which in some circumstances sets and setting aside some of their fees in a fund from banks’ capital requirements in relation to asset risks which third-party special masters of expedited civil as assessed by the rating agencies. Similar regula- judgments could indemnify investors for provable tions exist for insurance and other � nancial market harm. participants. Following failures of ratings in the U.S. Reform of credit rating agencies is not just subprime mortgage-based securities market, work an issue for advanced economies. Many EMDEs has been undertaken to reduce regulatory reliance have adopted or have been adopting regulatory on credit ratings. However, this is proving dif�cult at frameworks that are similar to those of the ad- times, not least because it complicates the adoption of vanced economies, including the reliance on credit global supervisory standards that do refer to ratings. ratings. For illustration, data from the 2011 Bank At the global level, a review of this issue by the Finan- Regulation and Supervision Survey show that 17 cial Stability Board has concluded that it may take a percent of EMDE regulators require their commer- number of years for market participants to develop cial banks to have external credit ratings; the com- enhanced risk management capability to enable parable number for advanced economy regulators is reduced reliance on credit rating agencies (FSB 2010b). 8 percent. principles aim to trigger a signi�cant change risk assessment practices instead. They set out in existing practices that would end mechanis- broad objectives for standard setters and reg- tic reliance by market participants on credit ulators to follow by de�ning the more speci�c ratings and establish stronger internal credit actions that will be needed to implement the GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 61 changes over time. For a discussion on credit gation capabilities, risk governance, and rating agencies, see box 2.3. internal controls Much work at the FSB level has been devoted to �lling information gaps. In 2009, The FSB, in coordination with the Basel the G-20 �nance ministers and central bank Committee, has come up with a list of 29 governors endorsed recommendations to banks considered global SIFIs. These banks address information gaps identified by the will need to meet the resolution planning FSB Secretariat and IMF staff. The FSB, in requirements by the end of 2012. National cooperation with the IMF and others, has authorities may decide to extend these launched a major initiative to fill existing requirements to other institutions in their information gaps (FSB and IMF 2010). jurisdictions. The group of global SIFIs will To deal with the too-big-to-fail issue, the be updated annually and published by the FSB has developed, in response to requests FSB each November. The methodology and from G-20 leaders, a set of policies to address data used by the FSB will be publicly avail- the systemic and moral hazard risks associ- able so that markets and institutions can rep- ated with systemically important financial licate the authorities’ determination.23 institutions (SIFIs). 21 The policies—to be The FSB acknowledged that consis- implemented from 2012, with full implemen- tent implementation will be critical to the tation from 2019—basically consist of the effectiveness of these measures. Legislative following:22 changes will be required in many jurisdic- tions to implement the FSB Key Attributes • A new international standard, titled “FSB of Effective Resolution Regimes and to Key Attributes of Effective Resolution strengthen supervisory mandates and capa- Regimes,� setting out the responsibilities, bilities. Other requirements will demand a instruments, and powers that national high degree of cooperation among authori- resolution regimes should have to enable ties, and �rms will have to review and change authorities to resolve failing financial their structures and operations. � rms in an orderly manner and without exposing taxpayers to the risk of loss NATIONAL REGULATION AND • Requirements for resolvability assess- SUPERVISION IN RESPONSE TO ments and for recovery and resolution THE CRISIS24 planning for global SIFIs, and for the development of institution-speci�c cross- What is the effect of the global regulatory border cooperation agreements so that reforms so far at the national level? The home and host authorities of the global World Bank survey of bank regulation and SIFIs are better prepared for dealing with supervision is useful in answering this ques- crises and have clarity on how to cooper- tion. The results from the survey underscore ate in a crisis the evolutionary nature of the regulatory and • Requirements for banks determined to be supervisory changes at the national level. To globally systemically important to have illustrate this point, for the qualitative ques- additional loss absorption capacity tai- tions in the survey, 85 percent of yes or no lored to the impact of their failure, rising responses were unchanged between 2007 and from 1.0 percent to 2.5 percent of risk- 2011. Similarly, most of the quantitative indi- weighted assets, to be met with common cators showed relatively little overall move- equity ment throughout the crisis. • More intensive and effective supervision This relatively slow evolution notwith- of SIFIs, including through stronger super- standing, the World Bank survey shows visory mandates, resources, and powers, notable changes in individual countries in and higher supervisory expectations for several areas. For example, in an attempt to risk management functions, data aggre- respond to the crisis, countries introduced a 62 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 2.1 Introduction of Bank Governance Frameworks The survey also indicates that many juris- (In response to the global �nancial crisis, dictions resorted to increasing the amount have you introduced new regulations in the following areas?) covered by deposit protection systems as a means to avert the systemic consequences of a widespread mistrust in banking institutions Compensation for executives (�gure 2.3). The crisis experience and the survey also Independence of the Board provide a unique opportunity to reexamine the broader framework for regulation and Existence of a direct reporting supervision. One of the most visible devel- line from the chief risk officer to opments in �nancial sector regulation in the the Board or Board Committee past 20 years has been a shift from the tradi- Existence of a Board or tional sector-by-sector approach to supervi- a risk committee sion toward a greater cross-sector integration of � nancial supervision Čihák and Podpiera Other 2008). This shift, which was to a large extent 0 10 20 30 40 50 60 70 in response to the growing integration of the banking, securities, and insurance markets, Number of responses has an important impact on the practice of supervision and regulation around the globe. Source: World Bank’s Bank Regulation and Supervision Survey (database), 2011. Note: The �gure shows the number of positive responses in each area. A country could respond Box 2.4 discusses the crisis experience with positively in several areas. the regulatory frameworks. The World Bank’s Bank Regulation and Supervision Survey con� rms that macropru- FIGURE 2.2 New Insolvency Frameworks dential policies received renewed impetus (Have you introduced signi�cant changes to the bank resolution after the crisis, with many new macropru- framework in your country as a result of the global �nancial crisis?) dential bodies involved (such as the Finan- cial Stability Oversight Council in the United States and the European Systemic Risk Board Introduced a separate bank in the European Union). This involvement insolvency framework led to rapid growth in new � nancial stabil- Implemented coordination ity reports, with India, Italy, and the United arrangements among States being some of the recent entrants, and domestic authorities it also encouraged a trend toward increased publication of �nancial sector stress tests (�g- Revisions under ure 2.4). consideration Most developing economy supervisory authorities still use the Basel I capital regime, No revisions though the majority plan on implementing the Basel II capital requirements soon (fig- ure 2.5a). In the survey, some 75 percent of 0 10 20 30 40 50 60 70 80 90 responding jurisdictions, including many Number of responses developing economies, indicated their inten- tion to implement Basel II. Basel II allows for Source: World Bank’s Bank Regulation and Supervision Survey (database), 2011. several approaches, some relatively simple and similar to Basel I, so in principle, for plethora of new requirements on bank gov- developing economies, especially those facing ernance frameworks (figure 2.1), and they important supervisory capacity constraints, have sought or are seeking to strengthen a simplified version of the standardized bank insolvency frameworks (figure 2.2). approach is probably the most appropriate GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 63 option. Yet many are aiming to adopt more FIGURE 2.3 Introduction of Deposit Protection Schemes complex approaches (�gure 2.5b), sometimes (Have you introduced changes to your deposit protection system without justi�cation in terms of the complex- as a result of the global �nancial crisis?) ity of the institutions that are to be super- vised or the types of transactions in which they are involved. Indeed, experience from Increase in amount covered World Bank country work indicates that in some small or lower-income countries, the full range of options proposed by the BCBS Expansion of coverage (types of exposures, nature is not properly thought through, resulting in of depositor, and so on) the adoption of overly complex regulations for the level of economic development and Government guarantee of complexity of the � nancial system. The sur- deposits and bank debts vey indicates that introducing Basel II already had substantial impacts in many countries Only minor changes (�gure 2.6), with the implementation being or no change more challenging for the developing econo- mies than for developed economies. 0 5 10 15 20 25 30 35 40 Developing-economy regulators offer a Number of responses variety of reasons why they want to adopt Basel II, although adoption is not manda- Source: World Bank’s Bank Regulation and Supervision Survey (database), 2011. tory outside the member states of the BCBS. Speci�cally, some are concerned that Basel narrow de�nitions of capital and without risk I is beginning to be perceived as an inferior weighting) to their minimum requirements. standard by international investors and that This trend is likely to continue as countries developing-economy financial institutions move toward Basel III. and markets may be penalized by inter- As for the impact of Basel III, recent cal- national market participants or that their culations (such as Majnoni 2012) suggest domestic banks may eventually be denied that Basel III implementation may, in con- access to foreign markets if they do not com- trast to Basel II, be relatively easier for devel- ply with the latest Basel standards. Accord- oping economies than for developed ones, ing to Financial Stability Institute (2004), the given that the former have built relatively main driver among nonmember countries of higher capital buffers. Indeed, minimum the BCBS to move toward Basel II is the fact required as well as actual risk-based capital that foreign-controlled banks or local subsid- ratios of banking systems tend to be rela- iaries of foreign banks operating under Basel tively higher in developing economies (�gure II expect regulators in low-income countries 2.8). However, higher capital (and liquidity) to adopt the framework as well. Whether or buffers may also be warranted, consider- not these concerns are justified, they have ing that emerging-market and developing- accelerated the diffusion of the Basel frame- economy banks operate in a more volatile works across the developing world, as docu- economic environment. Also, a common mented by the World Bank survey. observation from assessments under the The survey results suggest a somewhat World Bank/IMF’s Financial Sector Assess- increased emphasis on higher-quality capital ment Program and other diagnostic work is in regulatory capital relative to earlier sur- that high reported capital buffers overstate veys. For example, respondents in the more the true resilience of �nancial institutions in recent survey were less likely to include sub- light of de�ciencies in accounting and regu- ordinated debt in regulatory capital (�gure latory frameworks, especially as regards loan 2.7). Also, since the 2007 survey, seven coun- classi�cation, provisioning, and consolidated tries have added basic leverage ratios (with supervision. To examine the likely impact 64 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 2.4 Institutional Structures for Regulation and Supervision Regarding the broader architecture for regulation the twin peak model and against the sectoral model and supervision, three broad models are being used ˇ (Cihák and Podpiera 2008). Indeed, during the around the world: a three-pillar or “sectoral� model global � nancial crisis, some of the twin peak juris- (banking, insurance, and securities); a two-pillar or dictions (particularly Australia and Canada) have “twin peak� model (prudential and business con- been relatively unaffected, while the United States, a duct); and an integrated model (all types of super- jurisdiction with a fractionalized sectoral approach vision under one roof). One of the arguably most to supervision, has been at the crisis epicenter. How- remarkable developments of the past 10 years, con- ever, the crisis experience is far from black and � rmed by the World Bank’s Bank Regulation and white, with the Netherlands, one of the examples of Supervision Survey, has been a trend from the three- the twin peaks model, being involved in the Fortis pillar model toward either the two-pillar model or failure, one of the major European bank failures. It the integrated model (with the twin peak model is still early to make a � rm overall conclusion, and gaining traction in the early 2000s). isolating the effects of supervisory architecture from In a recent study, Melecky and Podpiera (2012) other effects is notoriously hard. examined the drivers of supervisory structures for There is one area where the postcrisis policy con- prudential and business conduct supervision over sensus is rather clear, though, and that relates to the the past decade in 98 countries, � nding among other role of the central bank in the supervisory frame- things that countries advancing to a higher stage of work. Recent policy papers on the subject (such as economic development tend to integrate their super- Nier and others 2011) emphasize the importance of visory structures, small open economies tend to opt central banks playing an important role in macro- for more integrated supervisory structures, � nancial prudential policy. Indeed, the World Bank’s bank deepening makes countries integrate supervision pro- regulation survey underscores the growing role of gressively more, and the lobbying power of the con- central banks in the supervisory framework and the centrated and highly pro�table banking sector acts as growing emphasis on macroprudential policy. The a negative force against business conduct integration. emphasis here is on macroprudential, as views differ (The related data on the structure of supervision are on the appropriate involvement of central banks in available on the website accompanying this report, microprudential supervision. In a recent study on the http://www.worldbank.org/� nancialdevelopment.) subject, Masciandaro, Pansini, and Quintyn (2011) How do these various institutional structures used empirical evidence from the crisis to make a compare in terms of crisis frequency and the limit- case for keeping macro- and microprudential super- ing of the crisis impact? Cross-country regressions vision institutionally separate to allow for more using data for a wide set of developing and devel- checks and balances and thus reduce the probability oped economies provide some evidence in favor of of supervisory failure. of Basel III on developing economies, World Finally, the survey suggests there is further Bank staff have undertaken in-depth analy- scope for improving disclosures and incen- sis of individual bank data. The results (box tives for stakeholders to monitor financial 2.5) suggest that the impact of the new capi- institutions, hence the need to address mar- tal regulations may be broadly manageable, ket discipline. The � ndings in this area are whereas the liquidity regulations may be somewhat mixed. Deposit insurance coverage more challenging, given the dif�cult exter- has increased during the crisis and, coupled nal funding environment, as well as the rela- with too-big-to-fail policies, is further erod- tively undeveloped local � nancial markets. ing incentives to monitor. Although the sur- However, there are important differences vey suggests that some elements of disclosure across regions, as well as within each region and quality of information have improved, and across �nancial institutions. it is not clear whether market discipline has been strengthened overall. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 65 HOW TO STRENGTHEN THE FIGURE 2.4 Financial Stability Reporting and Stress Test CRISIS RESPONSE Publication, 1995–2011 Are the global and national regulatory and supervisory responses suf�cient to address Percent the issues highlighted by the crisis? Is any- 90 thing missing in the crisis response so far? 80 As illustrated in this chapter, the � nancial 70 crisis has triggered much discussion and 60 many regulatory reform initiatives on the 50 global level as well as on the national level. 40 Economists have been following this reform process, and have voiced concerns that the 30 reforms are only going halfway (Beck 2011; 20 Shadow Regulatory Financial Committee 10 2011; Squam Lake Group 2010; London 0 School of Economics 2010). Economists, 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 regulators, and policy makers agree that the challenge is to design regulations to Countries that publish financial stability reports minimize the occurrence and cost of future Countries that publish stress test results crises; however, there is less agreement on the proposed approaches to regulation and ˇ Source: Cihák, Muñoz, Teh Sharifuddin, and Tintchev 2012. supervision. Table 2.4 provides a summary of selected proposals. reflect issues of regulatory complexity as well One common theme emerging from these as the capacity of the regulatory approach to studies involves concerns about the effective- address systemic risk that can lead to �nancial ness and ef�ciency of the regulatory approach crises. The trend toward regulating more and adopted by the of�cial sector. The concerns the growing complexity of regulation distorts FIGURE 2.5 Push to Implement New Basel Rules a. Year of Basel II adoption b. Selected approach to calculate minimum capital requirements Countries adopting Basel II (%) 100 90 Developing 80 countries 70 60 50 40 Developed 30 countries 20 10 0 2000 2005 2010 2015 2020 Standardized Foundation IRB Advanced IRB approach approach approach Median Deviation 2007 survey 2011 survey Source: World Bank’s Bank Regulation and Supervision Survey (database). Note: IRB = international ratings-based. 66 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 2.6 Impact of the Move to Basel II FIGURE 2.7 Quality of Capital (What was the impact of moving to Basel II on (Is subordinated debt included in regulatory capital?) the overall regulatory capital level in the banking system?) Respondents (%) Percent 100 40 90 35 80 70 30 60 25 50 20 40 30 15 20 10 10 5 0 2007 2011 0 Increased Increased Neutral/ Decreased Decreased substantially slightly little change slightly substantially Source: World Bank ’s Bank Regulation and Supervision Survey (database), 2011. All Developed Developing regulation that may limit innovation and hin- Source: World Bank’s Bank Regulation and Supervision Survey (database), 2011. der the ability of the �nancial system to per- Note: Percentage of countries responding in each category (of all countries that implemented Basel II). form its role in supporting growth and devel- opment. Overall, these reforms emphasize (a) FIGURE 2.8 Capital Adequacy Ratios: Minimum and Actual the importance of greater transparency and disclosure, (b) closer attention to incentives, so that regulations are “incentive-robust,� Actual risk-based capital ratio of the banking system as of end-2010, % and (c) simplicity, that is, keeping regulatory 40 38 rules as simple as possible to make it more 36 dif�cult for market participants to circum- 34 vent rules and easier for supervisors to moni- 32 tor and enforce them. 30 28 26 24 Asymmetric information, transparency, 22 and disclosure 20 18 Further enhancing the disclosure of infor- 16 mation should be a key component of regu- 14 latory reform. Asymmetric information—a 12 situation in which one party to the �nancial 10 8 transaction, usually the debtor, has access to 6 information material to the valuation of the 6 7 8 9 10 11 12 13 14 15 16 transaction that is not available to the other Minimum required risk-based regulatory capital ratio as of the end of 2010 (%) party, usually the creditor (Bebczuk 2003; Stiglitz and Rothschild 1976)—is a central Source: World Bank’s Bank Regulation and Supervision Survey (database), 2011. problem in �nancial systems because it lim- its the capacity of the investors, lenders, and incentives by facilitating regulatory arbitrage analysts to monitor effectively and to price and undermining the ability of supervisors correctly the risks in financial institutions to monitor and enforce these regulations. and instruments. The problems of asymmet- The concerns also reflect the risk of excessive ric information have increased as � nancial GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 67 BOX 2.5 Impact of the Basel III Implementation in Developing Economies This box examines implications of the Basel III mon equity component of capital. The de� nition of regulatory measures for developing economies. The capital will contain only a limited amount of certain analysis, which closely follows Ötker-Robe, Paz- intangibles and quali�ed assets (for example, banks arbasioglu, and others (2010), focuses on the impact can count up to 10 percent of DTAs resulting from of Basel III capital requirements for banks to have timing differences, MSRs, and significant invest- higher and better quality capital with greater loss ments in unconsolidated subsidiaries, capped at 15 absorption characteristics, taking into account the percent for the sum of DTAs, MSRs, and signi�cant Basel Committee on Banking Supervision (BCBS) investments in unconsolidated subsidiaries). Analy- rules on capital deductions and the market risk sis follows the BCBS indication that market risk framework agreed to in July 2010. It also covers the capital requirements will increase by an estimated impact of the new liquidity requirements in the form average of three to four times for large, internation- of the so-called net stable funding ratio (NSFR). ally active banks. Because of data constraints, this analysis does not Overall, the analysis suggests that the share of include the short-term liquidity coverage ratio and assets with less loss-absorbing characteristics to be the leverage ratio. The sample covers 127 banks in deducted from core tier 1 capital is relatively small 42 countries over six regions. on average for EMDE banks, except in the Latin The calculations are based, in large part, on com- America and Caribbean region (�gure B2.5.1). The pany reports and data from the Bankscope database. proportion of core tier 1 capital to be deducted varies An array of assumptions common to all banks are greatly across banks, reflecting their business char- made, given the lack of access to more granular acteristics and differences in tax systems; hence, the country-speci�c data on the various components of needed increase in capital to meet the new require- banks’ capital bases on a consistent basis. The � nd- ments can be large for some banks. Other intan- ings should hence be interpreted with caution. Since gible assets form the core of items to be deducted regional averages are affected by the sample, they (more than 30 percent), followed by net DTAs due may not fully represent the actual vulnerability of a to loss carry forwards (23 percent) and invest- given region to the new requirements. ments in unconsolidated subsidiaries (20 percent). According to the new capital standards, banks Net DTAs seem particularly important for Latin will be required to deduct most of their assets with America and the Caribbean (LAC), East Asia and less loss-absorbing characteristics—such as minority Paci�c (EAP), and Southern Africa, and investment interests, goodwill, net deferred tax assets (DTAs), in unconsolidated subsidiaries is relatively high in investment in unconsolidated subsidiaries, and Europe and Central Asia (ECA) and EAP. If applied mortgage servicing rights (MSRs)—from the com- immediately, the proposed deductions and market FIGURE B2.5.1 EMDEs: The Impact of Basel III Capital and Liquidity Requirements a. % of core tier 1 to be deducted b. Impact of Basel regulation on core capital ratio by end-2012 Percent Percent 30 28.0% 10 9.11% 25 8.22% 8 7.58% 20 6 15 4 10.1% 9.4% 10 8.0% 6.2% 2 5 4.0% 3.3% 0.92% 0.64% 0.61% 0 0 ur r 1 s W n ur r 1 –1 d , 2 re on tR ei 11 ne be ca L Af and l A nd ac sia sia Af aran , c Tie , c Tie tio co AL nt A nt 2 2 cti rke eas 20 ai rib eri ra a 01 dP tA hA an a sia c a re re s, et ra III ifi ric ric du rth st nt pe h tio e tio e Ca m ma Incr Sa ra Cor ra cor r an Eas No le Ea l ut De he in A Ce Euro se ing e So b- rn tiv Ba I l II Su d t at ea mula idd se an L M Ba Cu (box continued next page) 68 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 2.5 Impact of the Basel III Implementation in Developing Economies (continued) c. Current and adjusted core tier 1 ratios d. Net stable funding ratio (available stable funding/required stable funding) Percent 12 Percent 11.3% 10.6% 10.6% 10.9% 10.7% 160 10 10.0% 9.5% 9.8% 9.5% 9.0% 9.1% 8.3% 7.9% 140 8.2% 8.2% 8 7.6% 7.6% 7.2% 7.4% 7.0% 6.3% 120 6 100 4 80 2 60 0 40 L be ca l A nd Af and ac sia sia Af aran AL rib eri ra a dP tA hA an sia a c a 20 ifi ric ric nt pe rth st h Ca m Sa an Eas No le Ea ut he in A Ce Euro So b- 0 Su d t Lat idd M ac sia Af and Af aran sia be ca l A nd rib eri ra a dP tA hA c a a an sia an ifi ric ric rth st nt pe h Ca m Sa an Eas No Ea ut he in A Ce Euro Core Tier 1 ratio Basel III core ratio, 2010 b- So e Su l d t at idd an L M New Basel III core ratio, 2012 e. NSFR vs. share of wholesale funding f. NSFR vs. loan to deposit ratio Net stable funding ratio Net stable funding ratio (%) 180 180 160 160 140 140 120 120 100 100 80 80 60 60 40 40 20 20 0 0 0 10 20 30 40 50 60 70 80 90 0 50 100 150 200 250 300 Share of wholesale funding in total funding Loan to deposit ratio East Asia and Pacific Middle East and North Africa East Asia and Pacific Middle East and North Africa Europe and Central Asia South Asia Europe and Central Asia South Asia Latin America and the Caribbean Sub-Saharan Africa Latin America and the Caribbean Sub-Saharan Africa risk adjustments would lower the core tier 1 ratio by dependence on parent funding and underdeveloped about 1–3 percentage points on average. The overall local capital markets) or loan-to-deposit ratios are impact on the core tier 1 ratio is the largest for LAC, high (ECA and some LAC countries). However, Middle East and North Africa (MENA), and ECA other factors may also affect the level of NSFR (for regions, but most banks in the latter two regions still example, low levels of government securities in asset pass the required 7 percent level comfortably after portfolios may result in low NSFRs). Moreover, fur- the adjustments. ther challenges may be ahead in meeting the NSFR There is also wide variation in banks’ ability to requirement: for example, upcoming rollover needs meet the required 100 percent level for the net stable of European banks and sovereigns may raise the cost funding ratio (NSFR). The calculations suggest that of term funding that may spill over to EMDEs and the NSFR may have varying degrees of impact on result in competition for deposits. There are also EMDEs across regions, with wide variations within challenges associated with holding high levels of (liq- a given region. The ECA and LAC regions seem to uid) government securities; this would help in meet- be the most vulnerable to the NSFR, where depen- ing the NSFR target but expose the bank to higher dence on wholesale funding (ECA, given the high sovereign risk. Source: World Bank staff calculations based on Bankscope data for 127 banks in 42 countries. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 69 TABLE 2.4 Summary of Selected Proposals for Regulatory Reform Proposal by Speci�c Steps Admati and Hellwig 2012; Replace risk-weighted capital ratios by (signi�cantly higher) leverage ratios, Hellwig 2010 combined with strong disclosures about risk exposures. Barth, Caprio and Levine Rather than focusing on the regulations themselves, focuses on how to better 2012a oversee the regulators. Establishment of a “sentinel� agency, watching over the regulators on behalf of the taxpayers. Bartlett 2012 Redesign bank disclosures to facilitate credit modeling by market participants. (Illustrates that basic credit risk modeling, combined with appropriate bank disclosures, could have enhanced investors’ ability to detect the portfolio risk leading to recent banking crises, without revealing sensitive position-level data.) Brunnemeier and others Develop more “prompt corrective action�-types of rules to facilitate “leaning 2009 against the wind.� Calomiris 2011 Proposes “incentive-robust� reform proposals to address mortgage risk subsidization, regulators’ inability to measure risks ex ante and losses ex post, the too-big-to-fail problem, liquidity risk, macroprudential regulations that vary over the cycle, prudential regulations to encourage the greater use of clearinghouses in clearing over-the-counter transactions, and design of appropriate guidelines to constrain government assistance to banks during crises. Caprio, Demirgüç-Kunt, and Make oversight more adaptive to changes (innovations) and hold supervisors Kane 2010 accountable for their adaptiveness. Regulators should disclose information on the value and measurement of potential claims that institutions make on the government’s safety net. Establishing the right incentive structure for supervisors requires a chain of reforms. ˇ ihák, Demirgüç-Kunt, and C Reorient the approach to �nancial regulation to have at its core the objective of Johnston 2012a addressing incentives on an ongoing basis. As part of this, consider conducting regular “incentive audits.� Claessens and others 2010 Recognize mitigation of systemic risks as an explicit objective of all agencies involved in supervision, to enhance accountability: Clear mandates and tools commensurate with these mandates to preserve �nancial stability; suf�cient resources; clear allocation of responsibilities among agencies; Clear communication among agencies. de la Torre and Ize 2011 Establish strong and independent supervisory agencies, populated by highly skilled civil servants. Demirgüç-Kunt 2011; Scale back explicit deposit insurance from large banks as an additional measure Rajan 2010 to claw back implicit guarantees and remove the too-big-to-fail subsidies. Enriques and Hertig 2010 Strengthen internal and external governance of supervisors: (a) strong CEOs with boards’ and commissions’ powers limited to basic policy-making decisions and monitoring, (b) increased line responsibilities for staff, (c) requirement subjecting supervisors to stronger disclosure requirements. FSA (The Turner Address the need for more intrusive supervision, more outcomes-oriented Review) 2009 supervision, and more risk-based supervision, more systemic supervision, and international coordination of supervision. Masciandaro, Pansini, Make a clearer organizational distinction between macro- and microprudential Quintyn 2011 supervision to allow for more checks and balances to improve supervisory governance. (table continues next page) 70 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 TABLE 2.4 Summary of Selected Proposals for Regulatory Reform (continued) Proposal by Speci�c Steps Palmer and Cerutti 2009 Summon the “will to act� by (a) leaning more against the wind; (b) strengthening the context of supervision (independence, leadership, accountability); (c) strengthening supervisory processes by making them more intensive, result-oriented, risk-based, and proactive; (d) strengthening macroprudential surveillance and mitigating procyclicality; and (e) improving cross-border supervisory cooperation. Viñals and others 2010 Implement more intrusive supervision—“skeptical but proactive supervision� that is comprehensive, adaptive, and conclusive. Achieve changes through (a) enabling legislation and budgetary resources; (b) clear strategy; (c) robust internal organization; and (d) effective coordination with other agencies. Create revisions through (a) clear mandate, (b) independence and accountability, (c) skilled staff, (d) healthy relationship with industry, and (e) partnership with board. Weder di Mauro 2009 Establish more independence and accountability for supervisors to address time-inconsistency issues; offer higher compensation levels for supervisors; and set up supervision at supranational levels (Europe) to eliminate local industry capture. Wellink 2011 Address the need for “intrusive supervision.� instruments, structures and interconnections, supervision. These initiatives, which focus regulatory and accounting rules, and institu- primarily on ensuring better information for tions’ risk control and assessment techniques supervisory purposes, go some way toward have become more complex. addressing the weaknesses highlighted by the Asymmetries of information and principal- crisis. agent issues abound between buyers and sell- The focus on collecting better data for ers of �nancial services and products, as evi- supervision should not detract from the need denced by the malpractices in the U.S. for much better public disclosure of informa- mortgage sector in the run-up to the crisis. tion. The value of transparency and disclo- Professional bankers possess expert knowl- sure has been emphasized by many observers edge, and obtaining such knowledge is time- as well as by recent research (Bartlett 2012; consuming and costly. This puts the client at Demirgüç-Kunt, Detragiache, and Tressel a disadvantage, notably when monitoring 2008). Greater disclosure would allow credi- compliance with contractual arrangements. tors, investors, and analysts to assess directly In principle, disgruntled consumers can seek the solvency of the �nancial institutions. The legal recourse, but the legal process is time- �nancial crisis illustrated that �nancial insti- consuming, costly, and uncertain. These con- tutions are rather opaque organizations for ditions highlight the general case for detailed investors in capital markets. Although bank disclosure requirements and conduct of busi- regulatory policy has long sought to promote ness regulation. Moreover, information asym- market discipline of banks through enhanced metries are an important rationale for pru- public disclosure, bank regulatory disclosures dential regulation and supervision of banks are notoriously lacking in granular, position- accepting deposits from retail clients, as these level information concerning their credit nonprofessional consumers are ill-equipped investments, largely because of conflicting to evaluate the safety and soundness of banks. concerns about protecting the con�dentiality As part of the crisis response, the FSB and of a bank’s proprietary investment strategies others have launched useful initiatives on and customer information. data and information gaps. At the national When particular market sectors experi- level, many regulators have started collect- ence distress, investors are thus forced to ing additional data to allow for strengthened speculate as to which institutions might be GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 71 exposed, potentially causing signi�cant dis- Further improvements in transparency ruptions in credit markets and contributing and disclosures are seriously needed, both on to systemic risk. Bartlett (2012) argues that the systemwide level and on the level of indi- redesigning bank disclosures to facilitate vidual � nancial institutions. As regards the credit modeling by market participants has disclosures on systemic risks, many countries the potential to meaningfully increase mar- have been publishing so-called � nancial sta- ket discipline while minimizing the disclo- bility reports. Recent research on the subject sure of sensitive bank data. He illustrates suggests that simply publishing a financial how even basic credit risk modeling, when stability report seems to have no impact on combined with appropriate bank disclosures, �nancial stability (Čihák, Muñoz, Teh Shari- could have signi�cantly enhanced investors’ fuddin, and Tintchev 2012). The effective- ability to detect the portfolio risk leading to ness of such reports in signaling and address- two recent severe banking crises. Moreover, ing systemic risk has, however, been affected because such an approach leverages the same by a number of factors. In the absence of a aggregate metrics banks themselves use to unifying analytical framework for assessing monitor their risk exposure, the proposed systemic risk, most �nancial stability reports disclosure regime would impose a limited were rather descriptive and refrained from disclosure burden on banks while avoiding explicit statements about the level of systemic the need to reveal sensitive position-level risk present in the financial system. Data data. gaps, particularly in the nonbank sector, led It would be naive, of course, to think that many reports to focus on the banking sec- all creditors, investors, and analysts have the tor, impeding a true systemwide perspective. resources and capacity to understand, assess, Also, the articulation of financial stability and identify these increasingly more complex analysis into remedial policies, aimed at curb- structures, institutions, and instruments. ing the buildup of systemic risk, was prob- Indeed, the collective tendency of � nancial lematic. There is thus substantial scope for �rms, non�nancial corporations, and house- improving such reports (as well as the asso- holds to overexpose themselves to risk in the ciated information on systemic-level risks) in upswing of a credit cycle, and to become terms of clarity, consistency over time, and overly risk-averse in a downswing, has been coverage. The ongoing work on good prac- well documented. These tendencies raise tices in macroprudential surveillance (such as some questions about the capacity of � nan- Nier and others 2011) could usefully address cial markets and investors to instill discipline transparency and disclosures of systemic risk. on the behavior of �nancial entities, and they At the level of individual �nancial institu- underscore the importance of having both tions, further reforms are needed to ensure a strong supervision and market discipline. higher quality of disclosures. Much reliance One of the important advantages of com- has been placed on the external auditors, a plementing strong supervision with market sector that came to be dominated by the “Big discipline is that, with suf�cient disclosures Four� (KPMG, PwC, Ernst & Young, and and proper incentives, investors and analysts Deloitte). In the run-up to the �nancial crisis, would be more likely to develop their own many �nancial institutions were given a clean assessments of capital adequacy and liquidity, bill of health by the external auditors, only and there could be scope for competition and to be bailed out a few months later as the evolution in the design of the most appropri- � nancial crisis unfolded. In the wake of the ate measures. This approach would limit the crisis, the European Union has proposed a likelihood of “groupthink� and focusing too draft law to tighten supervision of the exter- much on a single and possibly flawed proxy nal auditors. At the global level, the FSB has or rating system. Ultimately, the approach requested action from several global bodies would help in limiting the buildup of risk that to ensure greater international consistency in occurred prior to the �nancial crisis. audit practices, and to provide more speci�c 72 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 guidance for external audits. The focus of source of financial instability in finan- these reforms is on ensuring higher-quality cial systems. Modern financial systems information for supervisors, but it is impor- with limited liability encourage risk-tak- tant to use the momentum also for improving ing incentives in financial institutions, 25 the quality of information that investors get and these incentives can be exacerbated from the audits. by badly designed regulations and safety Finally, a further push is needed to nets. The literature has identi�ed a number strengthen global accounting standards. of market failures that are relevant to �nan- Many observers have recently awakened cial stability and that result from incentive to the importance of accounting because problems. 26 of the controversy over the mark-to-market Of course, most regulations affect incen- principle (or more appropriately, “fair value tives in one way or another, and the impor- accounting�) and its impact during the �nan- tance of designing regulations in a way that cial crisis. In late 2007 and early 2008, sev- is incentive-compatible is being increasingly eral prominent � nanciers and analysts pro- recognized in international forums. The tested that the rapid decreases in the market recent regulatory reforms at the global level, prices of U.S. mortgage-based securities and as well as at the country level, have included other assets were meaningless and caused by measures on systemically important �nancial liquidity shortfalls. According to them, mark- institutions, compensation policies, a reduced ing these assets to reduced amounts in bal- role of credit ratings, initiatives on data, and ance sheets would precipitate an unnecessary information gaps, all of which go some way crisis. With hindsight, this analysis was not toward addressing the weaknesses high- completely correct. It is widely accepted that lighted by the crisis. �nancial markets can overshoot in their cor- Reconciling private incentives with pub- rections, which in a mark-to-market environ- lic interest by regulation is key, but far from ment could create contagion effects; however, trivial. Addressing one incentive issue by a for the most part, the reason for the reduced new or amended regulation often only leads market prices was a permanent loss of value to creating incentive breakdowns elsewhere. rather than a temporary effect resulting from And some of the existing incentive issues lack of liquidity in the markets. In this case, (for example, lack of incentives of supervi- the transparency that fair value accounting sors from different jurisdictions to share provided played a key role in pushing these relevant information in situations of stress) institutions’ management through the recog- have not really been fully addressed, in spite nition of losses. Box 2.6 provides a viewpoint of the efforts made at the international level on this topic. under the coordination of the FSB. As dem- Greater transparency and better infor- onstrated, for example, in Barth, Caprio, and mation would not have the desired positive Levine (2012a), regulators have often failed effect on sustainable �nancial development if to implement the regulations and exercise the market participants did not have incentives powers that they already had. They point out to monitor performance. The next section that among other factors, psychological bias therefore turns to the issue of incentives. in favor of the industry—similar to that pre- vailing in sports, where referees regularly call games in favor of home teams—also operates Incentive issues in �nance. In the authors’ view, therefore, the The identi�cation and correction of incen- key issue to address is not necessarily more tive problems that create systemic risk regulations (although some additional regula- should be at the center of any framework tions may be appropriate), but it is how to get that seeks to maintain financial stabil- regulators to enforce the rules. So, how are ity. Many economists believe incentive governments to ensure that addressing the problems are perhaps the most important incentive breakdowns is indeed central to the GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 73 BOX 2.6 Accounting Standards (Viewpoint by Nicolas Véron) Sometimes it takes a narrow lens to distinguish the community, and setting incentives for individual true features of big objects. The future of � nancial jurisdictions to adopt and enforce standards to globalization, whatever one’s perspective on its foster genuine cross-border comparability of � nan- dangers or merits, is one of the biggest questions cial statements. At this point, it is still too early to of the moment. By contrast, accounting has often judge whether the attempt to make IFRS the domi- been perceived as boring. But the policy debate on nant global accounting language can succeed on an accounting, and especially on International Finan- enduring basis. But there are already important les- cial Reporting Standards (IFRS), entails large sons that have wide signi� cance beyond the com- stakes and important lessons for global financial munity of accounting professionals. integration. First, global financial rules are not a utopian As is so often the case, the policy debate is vision but a reality. The initial success of IFRS has obscured by the weight of special interests. Most been remarkable. Their adoption has been smooth notably, � nancial industry executives and their lob- and has generally improved financial reporting byists were able to convince many policy makers and quality, starting in the European Union in 2005 but non–accounting experts that fair value accounting increasingly now in other countries as well. Given had been a major aggravating factor in the initial the right conditions, � nancial regulatory harmoni- phase of the crisis, in late 2007 and early 2008, zation can work across continents. even as later developments clearly demonstrated that Second, the crisis has increased the need for pub- there had been no excessive undershooting of mar- lic oversight of � nancial rules, but it is not yet clear ket prices during that period. A similar sequence had how this can be done effectively and consistently. A happened a few years earlier in the United States, monitoring board of public entities was created in when corporate advocates managed to delay the 2009 to oversee the IFRS Foundation, but its con- accounting recognition of the cost of stock options struction is awkward and raises concerns about its for nearly a decade. In accounting, as in most other legitimacy and future effectiveness. For the foresee- areas of � nancial regulation, the issues are technical able future we will have to rely on trial-and-error and jargon-ridden, and the potential � nancial conse- experimentation for international � nancial regula- quences are large, so that public debates and policy tory bodies, which in most cases cannot take exist- decisions are easily captured. Therefore, governance ing national arrangements as a direct model. arrangements are crucial. Third, those global bodies that exist have yet to The governance of accounting standard setting adapt to the ongoing rebalancing of the � nancial has widely varied over time and across countries. world. The IFRS Foundation is registered in the The general trend of the past few decades has been United States; its staff is in London; its monitor- toward standard setters that are more independent ing board gives permanent seats only to the United vis-à-vis governments and special industry interests. States, European Union, and Japan; and it still But the challenges related to the IFRS are unprece- caters largely to audiences in the developed world, dented because these standards are set at the global even as large emerging economies represent a rap- level. There is no global government to oversee the idly increasing share of global finance. We don’t IFRS Foundation, the organization that sets the stan- know whether or how China, India, and others dards, or to enforce consistent IFRS implementation will take responsibility at the global level for trans- across countries. Nor is there any coherent global parency and integrity in � nancial reporting. But if representation of investors, whose information needs efforts to empower them in formal global institu- the standards are primarily meant to serve. tions are not accelerated, it is hard to see how such Answers may lie in the IFRS Foundation mak- institutions can ful� ll their potential. ing more efforts to organize the global investor 74 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 regulatory framework? One suggestion, pro- approaches to assessing �nancial stability, in posed by Barth, Caprio, and Levine (2012a), particular, stress testing and use of supervi- instead of focusing on the regulations them- sory codes and standards. The incentive audit selves, focuses on how to better oversee the would focus on the key elements that moti- regulators. Speci�cally, they propose a senti- vate and guide �nancial decision making. nel agency, which would watch the regulators A focus on correcting incentive issues and on behalf of taxpayers. The agency would putting them front and center does not mean have no regulatory powers, but it would have discarding the existing regulatory and super- the ability to obtain all the information avail- visory frameworks. In other words, what is able to regulatory agencies, along with the needed is an evolutionary change, not a revo- duty to report on the key systemic risks and lution. The effort would involve a reorienta- on what the regulators are doing to address tion toward a regulatory and supervisory those risks. framework that is more focused on the suf- Going in a similar direction, Masciand- �ciency of information disclosures, factors aro, Pansini, and Quintyn (2011) emphasize that influence the incentives to monitor activ- the distinction between macro- and micro- ities within �nancial �rms, corporate gover- prudential supervision. Using empirical evi- nance and compensation practices, conflicts dence from the crisis, they make a case for of interest, and explicit and implicit guar- keeping macroprudential supervision insti- antees. A greater focus on incentives, which tutionally separate from microprudential could be achieved by extending or improving supervision. In other words, it seems better existing efforts, would support the regula- for macroprudential supervisors not to have tory approaches by identifying the underlying direct microprudential supervisory powers, distortions that can give rise to systemic risk, as long as they have the ability to obtain all including in the design and application of the the relevant information to assess the key regulations themselves. The effort would also systemic risks and the steps to address those help to prioritize the regulatory response on risks. Such an arrangement allows for more a systemwide basis and increase the attention checks and balances and thus reduces the on bolstering effective market discipline in probability of supervisory failure. support of regulation. Many countries have recently been put- ting in place committees or new agencies to Keeping regulatory rules as simple carry out macroprudential regulation and as possible, and promoting strong supervision. In some cases, these agencies or enforcement bodies maintain a clear separation between macro- and microprudential supervision. In It is important to keep the rules as simple as other countries, however, macroprudential needed for their effective monitoring, and to committees or agencies are also involved in promote strong enforcement. 27 This point microprudential supervision. This is different relates closely to the previous two points on from the sentinel agency, which cannot have asymmetric information and incentive prob- any direct regulatory or supervisory powers, lems: complexity can make it difficult for to limit the risk of conflicts of interest. supervisors, investors, and others to distill How are governments and agencies to the relevant information about soundness increase the focus on incentives in prac- of � nancial institutions and about systemic ˇ tice? Cihák, Demirgüç-Kunt, and Johnston risk. Also, combined with bad incentives for (2012a, 2012b) propose that the identi�ca- � nancial institutions’ managers, complexity tion of incentive problems be based on a can make it easier for �nancial institutions to specific analysis of incentives, an “incen- bypass regulatory rules and use obfuscation tive audit� (box 2.7). Such audits, building to their advantage. on a methodology proposed in an earlier However, finance is not a simple busi- paper by Johnston, Chai, and Schumacher ness. Imposing simplicity through draconian (2000), could be used to complement other measures would be either too costly or not GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 75 ˇ BOX 2.7 Incentive Audits (Viewpoint by Martin Cihák, Asli Demirgüç-Kunt, and R. Barry Johnston) Introducing incentive audits could strengthen � nan- to address incentive issues by adapting regula- cial sector policy. The basic idea of such audits is to tion, supervision, and other measures. In Iceland, more regularly and systemically evaluate structural the analysis was done as a postmortem, bene�ting factors that affect incentives for risk taking in the from hindsight. But it is feasible to do such analy- � nancial sector. The key issues that audits would sis ex ante. Indeed, much of the information used need to cover include contract design, banking pow- in the commission’s report was available (read- ers, banking relationships, structure of ownership ily, or with moderate data-gathering effort) even and liabilities, industrial organizations, existence of before the crisis. Also, the commission had relatively guarantees, and the adequacy of safety nets. modest resources (three members and small sup- How would incentive audits look in practice? To port staff), illustrating that incentive audits need get a clearer idea, one can take a 2010 report by a not be very costly or overly complicated to per- parliamentary commission examining the roots of form. As the commission’s report points out, “it the 2007 � nancial crisis in Iceland. The report notes should have been clear to the supervisory authori- the overly rapid growth of the three major Icelandic ties that such incentives existed and that there was banks as a major contributor of the crisis, and docu- reason for concern,� but supervisors “did not keep ments the underlying “strong incentives for growth,� up with the rapid changes in the banks’ practices,� which included the banks’ incentive schemes as well and instead of examining the underlying reasons for as the high leverage of the major owners (Special the changes, they took comfort in the banks’ capital Investigation Commission 2010). The report maps ratios exceeding a statutory minimum and appear- out the network of confl icting interests of key bank ˇ ing robust in narrowly de� ned stress tests (Cihák owners, who were also the largest debtors of these and Ong 2010). banks. Existing approaches do not entirely overlook This illustration captures the basic notion of issues related to incentives. Many reports on � nan- an incentive audit. There are other examples (such cial stability focus very narrowly on a quantitative as Calomiris 2011, who uses an incentive-based ˇ description and analysis of trends (Cihák, Muñoz, approach to propose a reform of the U.S. regulatory Teh Sharifuddin, and Tintchev 2012), but some do ˇ framework). Cihák, Demirgüç-Kunt, and Johnston mention the misalignments between private sector (2012b) provide a more detailed description of the incentives and public interests. So this area could audit, going from a top-level examination of the key be usefully extended and made a more permanent elements of the � nancial environment in an econ- feature of the reporting. Also, in the context of the omy—market structure and � nancial instruments, World Bank/IMF Financial Sector Assessment Pro- government safety net, legal framework, and quality gram, when assessments collect information on own- of enforcement—to a more detailed and prioritized ership of � nancial institutions, they look into issues assessment of incentives, mindful of the likely effect such as safety nets. The idea of incentive audits, on the behavior of the main agents in the system. therefore, is not to build a new assessment from The checklist of key features would be accompanied scratch, but to raise the pro� le of incentive-related by guidance with evaluation methodology for con- issues and bring more structure to the assessments. sistent application across countries. Incentive audits should be seen not as a replacement To be effective, incentive audits would have to of other parts of the overall assessment of vulner- be performed regularly, and their outcome used abilities, but as a complement. feasible. For example, requiring all banks have been circulated—for example, those to move from limited to unlimited liability that suggest directly constraining � nancial could bring banks’ size down substantially, institutions’ size or substantially narrowing but it would not be feasible in contemporary the range of permissible activities—would financial systems. 28 Other proposals that potentially have serious side effects and 76 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 weaknesses, such that their implementation critical of the risk-weighted asset concept should be subject to careful analysis before as a proxy for actual exposures, and sug- their actual adoption. One possible way to gested replacing risk-weighted capital ratios keep institutions from becoming systemically with (significantly higher) leverage ratios, important, as noted, for example, in Rajan combined with strong disclosures about risk (2010), may not be through crude prohibi- exposures. Their argument is based on the tions on size or activity, but through collect- observation that the system of risk-calibrated ing and monitoring of information about capital requirements, in particular under the exposures among institutions and risk con- model-based approach, played a key role in centrations in the system. allowing banks to be undercapitalized prior An important aspect of the debate on com- to the crisis, with strong systemic effects for plexity relates to issues of capital adequacy deleveraging multipliers and for the function- and leverage. Empirical evidence suggests ing of interbank markets. The issue is not that when faced with uncertainty, markets trivial, of course, partly because of the short- tend to pay more attention to more basic indi- and medium-term adverse effects such high cators that are more dif�cult to bypass. For ratios could have on �nancial intermediation example, Demirgüç-Kunt, Detragiache, and and �nancial development. Merrouche (2012) used a multicountry panel Another alternative approach to current of banks to study whether better-capitalized risk-based capital regulation would be to banks experienced higher stock returns dur- have a simple leverage ratio (which is simple ing the financial crisis. They found that a enough to monitor and enforce) adjusted stronger capital position was associated with upward by the loan spreads banks charge better stock market performance, and that their customers. As discussed in Calomiris the relationship is stronger when capital is (2011), using loan spreads to measure loan measured by the leverage ratio rather than the default risk is desirable because these spreads risk-adjusted capital ratio, and that higher- are accurate forecasters of the probability quality forms of capital, such as tier 1 capi- that a loan will become nonperforming (Ash- tal and tangible common equity, were more craft and Morgan 2003). This would be an relevant. These empirical �ndings, of course, example not only of a simple regulation but do not imply that using leverage as the only also of an incentive-robust one, since banks tool is a complete solution; nonetheless, the clearly would not have an incentive to lower authors use these results to make a case for their interest rates just to reduce their capi- relatively greater reliance on simpler capi- tal budgeting against a loan, because doing tal ratios, such as the leverage ratio, that are so would reduce their income and defeat the more dif�cult to circumvent. purpose of circumventing the regulation. 29 Of course, a sole reliance on the lever- An added advantage of this approach would age ratio, which is not risk-sensitive, could be that monitoring interest rates is fairly become problematic, as it could give banks uncomplicated even in the least developed of an incentive to shift to riskier activities to emerging markets. boost returns. Therefore, it is important to Other approaches focus on complement- take into account the relative riskiness of ing basic capital ratios (such as a common the various assets; however, views differ as equity requirement) by so-called contingent to who should do it and how. Basel III has capital (CoCo) requirements. These authors recognized the usefulness of basic leverage provide evidence that seems to imply that a ratios and narrower (high-quality) measures CoCo requirement, complementing common of capital and has combined them with the equity, would be an effective prudential tool. risk-based ratios. CoCos can help the prompt recapitalization Several authors have proposed to move of banks after significant losses of equity even further. For example, Hellwig (2010) but before the bank has run out of options and Admati and Hellwig (2012) have been to access the equity market. That dynamic GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 77 BOX 2.8 Regulatory Discipline and Market Discipline: Opposites or Complements? In a major precrisis study of banking regulation T he global � nancial crisis has highlighted fail- around the world, Barth, Caprio, and Levine (2006) ures both in regulatory and supervisory discipline have used the data from the earlier versions of the as well as in market discipline. It has thrown a World Bank’s Bank Regulation and Supervision Sur- particularly unfl attering light on market discipline, vey to examine the various regulatory approaches and highlighting its intrinsic limitations, especially when compare them to the outcomes that countries care key assumptions are not met. At the same time, the about. They concluded that the standard features of crisis has also highlighted the serious regulatory and banking supervision and regulation do not reduce— supervisory failures that contributed to the crisis. and may even increase—the chance that countries In a recent paper, Barth, Caprio, and Levine experience banking crises. Nor do these rules and (2012b) follow up on this analysis, using the crisis regulations lead to more developed banking sectors or observations as well as data from the 2011 Bank more ef�cient banks. These �ndings are, according to Regulation and Supervision Survey database. The the authors, consistent with private interest views and results of their analysis support a complementary the fact that “few countries have highly developed role for regulation and market discipline, as high- democratic institutions� (13). In contrast, policies lighted recently by Haldane (2011). Market disci- that enable private markets to better monitor banks pline and state-imposed regulatory and supervisory and that encourage private actors to “discipline� discipline are complementary, since each has its own banks are associated with desirable outcomes. But limitations, and market failures often have their critically, they found no link between market moni- roots in regulatory failures. toring and the likelihood of a bank crisis. incentive feature of a properly designed The debate on these proposals is still ongo- CoCo requirement would encourage effective ing, so this is unlikely to be the last word. risk governance by banks, provide a more But most proposals argue that the regula- effective solution to the too-big-to-fail prob- tory framework should include well-de� ned lem, reduce forbearance risk (supervisory capital and liquidity measures that are moni- reluctance to recognize losses), and address tored and enforced by a strong supervisory uncertainty about the appropriate amount of body, which should be held accountable for capital banks need to hold (and the changes its activities. This need for strong enforce- in that amount over time). Calomiris and ment of regulatory rules should also include Herring (2011) examine this proposal in transparency and disclosure requirements detail, concluding that if a proper CoCo (as pointed out, for example, by the Shadow requirement had been in place in 2007, the Financial Regulatory Committee 2011). In disruptive failures of large � nancial institu- other words, proposals suggest market dis- tions, and the systemic meltdown after Sep- cipline should become an important comple- tember 2008, could have been avoided. They ment to the supervisory discipline provided note that, to be effective, (a) a large amount by an independent but accountable supervi- of CoCos relative to common equity should sory body (box 2.8). be required, (b) CoCo conversion should be The issue of regulatory complexity is, based on a market value trigger, (c) all CoCos of course, much broader than the capital should convert if conversion is triggered, and requirements and the issue of risk weighting. (d) the conversion ratio should be dilutive of The U.S. Shadow Regulatory Committee has preexisting equity holders. However, how long advocated simpler and more transpar- these untested instruments would actually ent regulations (Shadow Financial Regula- perform in case of need remains to be seen. tory Committee 2011). However, the broader 78 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 trend, con� rmed also by the Bank Regula- that did not have a banking crisis in 2007 tion and Supervision Survey, is toward more through 2009—had more stringent defini- complex regulations, not only in developed tions of capital, higher capital levels, and less economies but also in developing economies. complex regulatory frameworks. They had Various observers have pointed out the bur- stricter audit procedures, limits on related den of this new complexity for the industry;30 party exposures, and asset classi�cation stan- but in many countries, especially the smaller dards; and their supervisors were more likely and lower-income ones, this also adds sub- to require shareholders to support distressed stantially to the already existing capacity banks with new equity. Noncrisis countries constraints. This complexity makes it more were also characterized by better quality of dif�cult for supervisors to ensure that regula- � nancial information and greater incentives tions are actually and effectively implemented to use that information—among other rea- and, importantly, for taxpayers to see what sons, because they have relatively less gener- is being done to keep the system safe and to ous deposit insurance coverage. hold supervisors accountable. The global financial crisis has also trig- gered a healthy policy debate on approaches to regulation and supervision. This ongoing CONCLUSION debate among regulators, policy makers, and Although the overall role of the state in academics has led to multiple reform propos- � nance is an open question, there is clearly als, highlighting the diversity of views. This an important role for the state in � nancial is likely to inform the regulatory reform pro- sector regulation and supervision. This is the cess and improve future outcomes. one area where the role of the state is not in This chapter reviewed the progress with dispute; the real debate is over how to ensure regulatory reforms at the global level as well that the role is performed well. as in individual countries, and identi�ed the Good regulation needs to better align pri- advances made so far in many areas. It also vate incentives with public interest, without recognized a number of reform proposals taxing or subsidizing private risk taking. that suggest improvements on the current Supervision is meant to ensure implementa- approaches to regulation and supervision. tion of rules and regulations and to address These proposals aim to limit regulatory arbi- limitations of market discipline. trage opportunities and better employ regu- The global financial crisis underscored latory resources and capacity. Among the limitations in both regulatory and market common themes of these proposals are calls discipline. It emphasized the importance of for greater regulatory simplicity and trans- combining strong, timely, and anticipatory parency as a way to enhance accountability, supervisory enforcement with a better use as well as for more proactive identifying and of market discipline. It also highlighted the addressing of incentive problems. importance of the basics, that is, solid and transparent legal and institutional frame- NOTES works to promote �nancial stability. In many developing economies, the conclusion is that 1. For example, Greenspan (2005, para. 17–19) building up supervisory capacity needs to be remarked that “regulatory reform, coupled with innovative technologies, has stimulated a top priority. the development of �nancial products, such Lessons can be learned by analyzing regu- as asset-backed securities, collateral loan lation and supervision in economies that were obligations, and credit default swaps that at the epicenter of the global � nancial crisis facilitate the dispersion of risk. . . . These and those that were not. The World Bank’s increasingly complex �nancial instruments new global survey, presented in this chap- have contributed to the development of a far ter, suggests that noncrisis countries—those more flexible, ef�cient, and hence resilient GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R 79 �nancial system than the one that existed just poses. For example, U.S. supervisors did raise a quarter-century ago.� alarms about the risks of subprime lending, 2. See FSB Charter, article 1. but a signi�cant tightening of the pruden- 3. This report uses a broad concept of “the tial practices did not occur before the crisis, state� that includes not only government but reflecting pressures from the industry and also autonomous or semiautonomous agen- lawmakers (Levine 2010). Reviews of com- cies such as a central bank or �nancial super- pliance with the Basel Core Principles �nd vision agency. that some of the weakest areas relate to the 4. For example, Carmichel and Pomerleano operational independence of the regulators, (2002) and de la Torre, Ize, and Schmukler and that—despite some progress in recent (2011). years—this is still an issue in many develop- 5. In addition to regulation, another state inter- ing economies (Čihák and Tieman 2011). vention that can also have an impact on �nan- 13. For more on shadow banking and FSB-related cial sector risk taking is �nancial taxation. It response see FSB (2012). has received some policy attention during the 14. Demirgüç-Kunt and Huizinga (2011), using crisis (see, for example, IMF 2010c). None- a wide international sample of banks, present theless, the regulatory approach is likely to evidence that casts doubts on the need for sys- remain central for practical policy in the fore- temically large banks even from the narrower seeable future, and is therefore the focus of this perspective of bank shareholders. It suggests chapter (for a discussion of the pros and cons that bank growth has not been in the interest of regulation and taxation, see Keen 2011). of bank shareholders in small economies, and 6. Masciandaro, Pansini, and Quintyn (2011) it is not clear whether those in larger econo- provide an overview of the literature on the mies have bene�ted. Inadequate corporate causes on the crisis, focusing on supervisory governance structures in �nancial institutions and regulatory failures. have enabled managers to pursue high-growth 7. See the respective IMF country reports (IMF strategies at the expense of shareholders, pro- 2010a; IMF 2012). viding support for greater government regula- 8. Data on larger and more interconnected tion (as also argued, for example, in Barth, �nancial institutions show that they have Caprio, and Levine 2012a). been taking on more risk and have been more 15. The work of the FSB in this area is meant to likely to experience �nancial stress than oth- make credible to all counterparts of a �nan- ers (Ötker-Robe and others 2011). cial institution the possibility of its failure, so 9. Many banks, especially in advanced econo- they exercise due monitoring on management mies, held a relatively small part of capital and control of shareholders. as equity, with the remainder being in capital 16. Another related, although less explored, facet with weak loss-absorbing characteristics that of market discipline is the forced departure had little value during the crisis. Given the of managers from underperforming �nancial large differences and lack of transparency in institutions. Schaeck and others (2011) �nd the de�nition of capital, it was hard to assess that when banks take on too much risk and and compare the adequacy of capital across get into trouble, their managers do get forced institutions. out. But it is often too late for the banks, 10. Banks could reduce their risk capital require- which tend to remain in trouble for years ments through shifts in assets to legally remote after the turnover. entities excluded from asset de�nitions, 17. For more on this subject, see also a recent through credit default swaps that reduced the debate between Rene Stulz of Ohio State regulatory risks in their portfolio, or through University and Lucian Bebchuk of Harvard credit enhancements that improved the rat- Law School and others at the “All About ings of assets and thus the need to hold regu- Finance� blog on the World Bank’s website. latory capital. (http://blogs.worldbank.org/allabout�nance/ 11. For example, the mandates of the United the-aaf-virtual-debates). Kingdom’s Financial Services Authority and 18. Gaps and weaknesses in regulation and Switzerland’s Federal Banking Commission. supervision were not the only factors con- 12. Even supervisors that are independent in tributing to the crisis, as discussed earlier in principle can be overruled for political pur- this chapter, though they are expected to have 80 T H E S TAT E A S R E G U L AT O R A N D S U P E R V I S O R GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 played an important role. Also, although the pros and cons of returning to unlimited systemic crisis prevention is not the only liability for banks. Although such a move objective of regulation and supervision (for would almost certainly bring banks’ size example, some regulations focus on customer down “with a bang� (as noted by Charles protection, anti–money laundering, and so Goodhart in his March 1, 2012, debate on), crisis prevention is usually seen as a with Robert Pringle at centralbanking.com), key objective, so juxtaposing regulation and it would not be feasible in a contemporary supervision in crisis and noncrisis countries �nancial system. does offer interesting insights. Nevertheless, 26. For a broad overview of the underlying fac- users should note that these �ndings are cor- tors of �nancial crises, see de la Torre and Ize relations, and do not imply causality. (2011). In addition to incentive breakdowns 19. This section does not aspire to be an all- and asymmetric information, they also men- inclusive compendium of all reforms. More tion issues of collective cognition (“nobody work is ongoing or contemplated, for really understands what is going on�) and instance on accounting standards, shadow costly enforcement (“crises are a natural part banking, �nancial supervision, and mar- of the �nancial landscape�). ket infrastructures. For more on these, see 27. The long-running debate on rules versus http://www.financialstabilityboard.org/ principles in supervision (Mersch 2007) has publications/r_111104.pdf. been intensi�ed by the crisis. In the context of 20. See the Leaders’ statement after the Pittsburgh this debate, for example, the incentive audits Summit in September 2009 (G-20 Leaders, proposal calls for a greater emphasis on well- 2009). de�ned principles. Speci�cally, the principles 21. See FSB (2011b). SIFIs are �nancial institu- need to address the various misalignments in tions whose distress or disorderly failure, incentives, both among market participants because of their size, complexity, and inter- and among regulators. But emphasizing well- connectedness, would cause signi�cant dis- de�ned principles does not mean that one can ruption to the wider �nancial system and do away with rules. Similarly, the proposal to economic activity. focus on incentives does not mean abolishing 22. In addition to the four steps listed here, stron- microprudential supervision. Effective rules ger international standards for core �nancial and ef�cient principles are both essential to market infrastructures are to be �nalized in promote �nancial integration and reinforce early 2012, aiming to reduce contagion risks �nancial stability. It would be naive to think when failures occur. that principles applied on a stand-alone basis 23. On May 31, 2012, the International Asso- can eliminate the need for rules. ciation of Insurance Supervisors released for 28. The intrinsic desirability of such a measure public consultation its assessment method- for �nancial stability purposes is far from ology for identifying globally systemically undisputed, since the solvency analysis that important insurers (see http://www.iaisweb. creditors are expected to make would shift its org/view/element_href.cfm?src=1/15384. focus from the bank per se to the personal pdf). wealth of the bank proprietors, which would 24. This report focuses on global trends. Recent probably prove to be a quite challenging task. regional reports of the World Bank provide 29. Interest rates on deposits could also be used related updates on regulatory and supervisory because these too are associated with bank trends in individual regions, such as Latin risk (Acharya and Mora 2012). However, America and the Caribbean (de la Torre, deposit rates tend to be sensitive to bank risk Ize, and Schmukler 2011) and Middle East only very close in time to bank insolvency and North Africa (Rocha, Arvai, and Farazi because of explicit and implicit deposit insur- 2011). ance. 25. Banks have not always operated with limited 30. For the United States, see for example, the liability, pre-19th-century England being a Economist 2012; for the European Union, case in point. Several authors have discussed see, for example, Wall Street Journal 2011. 3 The Role of the State in Promoting Bank Competition • Competition in the banking sector promotes ef�ciency and �nancial inclusion, with- out necessarily undermining �nancial stability. • Even if the recent crisis is perceived as an episode where competition exacerbated private risk taking and helped destabilize the system, the correct public policy is not to restrict competition. What is needed is a regulatory framework that ensures that private incentives are aligned with public interest. • The state can play a role in enhancing banking competition by designing policies that guarantee market contestability through healthy entry of well-capitalized institutions and timely exit of insolvent ones and by creating a market-friendly informational and institutional framework. • Governments should be mindful of the consequences of their intervention during crises and limit negative consequences on bank competition and risk taking. he recent crisis reignited the interest of to future instability as a result of moral haz- T policy makers and academics in assess- ing bank competition and rethinking the role of the state in shaping competition policies ard problems associated with too-big-to-fail institutions. Box 3.1 presents a recent debate on the relationship between competition and (that is, policies and laws that affect the extent �nancial stability. to which banks compete).1,2 Some believe that Another reason why competition mat- increases in competition and �nancial inno- ters is related to the changing mandate of vation in markets such as subprime lending central banks and bank regulatory agen- contributed to the recent � nancial turmoil. cies. Although traditionally the primary goal Others worry that the crisis and government of bank regulators has been to ensure bank support of the largest banks increased bank- stability, this is changing. According to the ing concentration, reducing competition and World Bank’s Bank Regulation and Supervi- access to �nance, and potentially contributing sion Survey, updated in 2011, 71 percent of GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 81 82 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 3.1 Two Views on the Link between Competition and Stability In a recent debate held by The Economist magazine, were weakened by banks’ off balance sheet vehicles, two banking professors expressed contrasting views which were used to hold securitized assets.� about the role of bank competition in promoting On the other hand, Thorsten Beck, Professor of stability. Economics and Chairman of the European Banking According to Franklin Allen, Nippon Life Pro- Center at Tilburg University, argues that “compe- fessor of Finance and Economics, Wharton School, tition in banking is not dangerous per se; it is the University of Pennsylvania, “more competition does regulatory framework in which banks operate and make banking more dangerous.� But he also cautions which sets their risk-taking incentives that drives that “competition is only one of the factors contribut- stability or fragility of banking. Competition can ing to instability.� He goes on to say that “the experi- be a powerful source of useful innovation and ef�- ence of a number of countries in the past and dur- ciency, ultimately bene�tting enterprises and house- ing the recent crisis provides some insights into the holds; competition can also foster stability through relevant issues. Historically, the comparison that has improved lending technologies; competition, how- often been made is between the stability of the Cana- ever, can also endanger stability if mixed with the dian banking system compared to the United States’ wrong kind of regulation.� experience. In the late 19th and early 20th century, “Risks and dangers in banking arise primarily the United States had many banking crises, while from a regulatory framework that is not adapted Canada did not. The standard explanation for this is to the market structure. Large � nancial institutions that Canada had a few large banks, while the United turn too-big-to-fail because the regulator does not States had many small banks. In the recent crisis, the have any means to properly discipline and resolve banking system in Canada and also that in Australia them. Similarly, competition results in herding and were very resilient. Six banks dominate the Canadian increased fragility risk in the absence of macro-pru- � nancial system, while there are four major banks dential tools to counter asset price and credit booms together with a few small domestic banks in Austra- and take into account co-variation between banks’ lia. However, the United Kingdom, whose banking risk pro� les. The experience from the last crisis has system has a broadly similar structure to Australia’s, led to reform attempts exactly in these two areas: with four major banks and a few other small domes- resolution, especially of systemically important tic and foreign banks, had a very different experience. � nancial institutions, and macro-prudential regula- The lesson of this comparison is that competition is tion. It is thus not market structure or competition only one of the many factors that are important. In per se, that drives fragility, but a regulatory frame- addition to the competitive nature of the industry, work that sets the wrong incentives.� funding structure and the institutional and regulatory “The challenge is to maintain competition in the environments are important. These factors are well market to the bene�t of the real economy, while at illustrated by the recent experience of Canada, Aus- the same time creating a regulatory framework that tralia, and the United Kingdom. Canadian and Aus- minimizes the negative implications that competi- tralian banks mainly relied on depositary funding. tion can have for stability. Such a framework would This funding source proved stable through the crisis. include additional capital charges for size, com- In contrast, British banks increasingly used wholesale plexity and systemic importance of banks, macro- funding from financial markets. Canada and Aus- prudential regulations that take into account the tralia also have much more conservative regulatory interaction between financial institutions, and— environments than the UK. For example, in Canada, most critically—a resolution framework that allows capital regulation is stricter than the Basel agreements resolving even the largest � nancial institutions, thus require. Banks’ foreign and wholesale activities are reducing the perverse incentives stemming from a limited. The mortgage market is also conservative in too-big-to-fail status.� terms of the products offered, with less than 3 percent This discussion suggests that both sides share being subprime and less than 30 percent being securi- more in common than they disagree with, but see tized. In the UK a ‘light touch’ regulatory framework Economist 2012 http://www.economist.com/debate/ was implemented. An illustration is that capital ratios days/view/706 for more. Source: The Economist 2012 (reprinted with permission). GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 83 bank regulators report that their mandate and bank pricing behavior. It then reviews also includes promoting � nancial inclusion the evidence on the implications of banking and economic development. Also, 65 percent competition for bank efficiency, access to mention issues of market conduct, and nearly �nance, and �nancial stability. After that, the 25 percent mention competition policy. chapter analyzes the policy drivers of compe- Hence, either directly or indirectly—because tition and highlights the role of the state as competition influences market conduct and a regulator and enabler of a market-friendly access to � nance—competition is an impor- informational and institutional environment. tant issue for regulators.3 It also examines the impact on competition This chapter presents measures of bank of government actions during crises. The competition and describes basic trends across chapter concludes by summarizing the policy economies and over time. By illustrating implications. various approaches to measuring competi- tion and discussing factors that drive it, the BANK COMPETITION: chapter seeks to provide guidance to policy MEASUREMENT AND makers. STYLIZED FACTS The chapter conveys four main messages: There are three main approaches to assess- • Bank competition improves efficiency ing bank competition: measures of bank across banks and enhances access to �nan- concentration under the “structure-conduct- cial services, while not necessarily eroding performance� paradigm, regulatory indica- the stability of the �nancial system. tors that measure the contestability of the • Policies to address the causes of the recent banking sector, and direct measures of bank crisis should not restrict competition. The pricing behavior or market power based on correct public policy should establish a the “new empirical industrial organization� regulatory framework that supervises and literature. ensures the alignment of private incentives An alternative approach used by some with public interest. studies to analyze bank competition is based • The state should promote competition on interest spread decomposition (box 3.2). both as a regulator and as an enabler of a But spreads are outcome measures of effi- market-friendly informational and institu- ciency, and in addition to the competition tional environment. Policies that improve environment, cross-country differences in market contestability—through healthy spreads can reflect macroeconomic perfor- entry of well-capitalized institutions and mance, the extent of taxation on financial timely exit of insolvent ones, opportune intermediation, the quality of the contractual flow of adequate credit information, and and judicial environment, and bank-speci�c contract enforceability—will enhance factors such as scale and risk preferences. So competition among banks. this chapter instead presents direct measures • State interventions during crises may cre- such as the Panzar-Rosse H-statistic, the ate barriers to exit that permit insolvent Lerner index, and the so-called Boone indica- and inef�cient banks to survive and gener- tor. Box 3.3 summarizes these measures.4 ate unhealthy competition. Governments Competition may vary within economies should take steps to eliminate distortions and across products (for example, by type of in risk taking and limit their negative con- loan, such as corporate or consumer). Ideally, sequences on bank competition. competition should be measured by business line for different markets (box 3.4). But such The chapter first discusses alternative disaggregated data are often not available, measures of competition and presents trends and most measures cannot be computed sep- across economies and over time, using mea- arately for these submarkets. Accordingly, in sures of market concentration, contestability, what follows, country and regional measures 84 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 3.2 Decomposing Bank Spreads to Make Inferences about Bank Competition Bank interest spreads are frequently used as an indi- Demirgüç-Kunt and Huizinga (1999) and Beck cator of the ef�ciency of the banking system (Beck and Fuchs (2004) follow the identities above to con- and Fuchs 2004; Demirgüç-Kunt and Huizinga duct an accounting decomposition and an economic 1999; Demirgüç-Kunt, Laeven, and Levine 2004). analysis of the determinants of bank net interest An accounting decomposition of bank spreads or margins using data for 80 countries between 1988– of interest margins (the value of a bank’s net inter- 95, in the first case, and focusing on 38 banks in est income divided by assets) can be derived from a Kenya for the year 2002, in the second case. straightforward accounting identity: To the extent that high spreads are explained by high pro�t margins, these studies infer that lack of Before-tax pro�ts to assets (BTP/TA) = After-tax competition could be a factor. In the economic anal- pro�ts to assets (ATP/TA) + taxes to assets (TA/A) ysis of spreads, Demirgüç-Kunt and Huizinga (1999) From a bank’s income statement, before-tax pro�ts regress spreads and pro�ts on measures of concen- must satisfy the accounting identity: tration (as an indicator of competition) and conclude that, aside from other factors, lack of bank compe- BTP/TA = NI/TA + NII/TA – OV/TA – LLP/TA tition drives bank spreads and pro�ts across coun- where NI is net interest income, NII refers to non- tries. Similarly, Beck and Fuchs (2004) conclude that interest income, OV stands for overhead costs, and the high pro�t margins that explain part of the high LLP refers to loan loss provisioning. The identities spreads in Kenya are due to lack of competition in above allow for a decomposition of net interest mar- the banking sector. gins (NI/TA) into its components: NI/TA = ATP/TA + TA/A − NII/TA + OV/TA + LLP/TA are used to illustrate different approaches to region have the largest CR5 concentration assessing bank competition. ratios (�gure 3.2). The structure-conduct-performance para- Concentration measures are not good pre- digm assumes that there is a stable, causal dictors of competition.5 The predictive accu- relationship between the structure of the racy of concentration measures on banking banking industry, � rm conduct, and perfor- competition is challenged by the concept of mance. It suggests that fewer and larger �rms market contestability. The behavior of banks are more likely to engage in anticompetitive in contestable markets is determined by behavior. In this framework, competition is threat of entry and exit. Banks are pressured negatively related to measures of concentra- to behave competitively in an industry with tion, such as the share of assets held by the low entry restrictions on new banks and easy top three or �ve largest banks and the Her- exit conditions for unpro�table institutions— �ndahl index. even if the market is concentrated. Figure 3.1 depicts the asset share of the Figure 3.3 depicts two (admittedly imper- five largest banks (CR5) in developed and fect) proxies of regulatory indicators that developing economies, showing that banking capture entry conditions into the banking systems are more concentrated in developing industry: an index of barriers to entry and than developed economies. Across regions, the share of banking licenses denied. These banking systems in Sub-Saharan Africa and two indicators are from the World Bank’s the Gulf Cooperation Council (GCC) coun- Bank Regulation and Supervision Survey, tries of the Middle East and North Africa and they capture entry restrictions into the GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 85 BOX 3.3 Measuring Banking Sector Concentration and Competition Banking concentration can be approximated by the Higher values of the H-statistic are associated concentration ratio—the share of assets held by the k with more competitive banking systems. Under a largest banks (typically three or �ve) in a given econ- monopoly, an increase in input prices results in a rise omy—or the Her� ndahl-Hirschman index (HHI), in marginal costs, a fall in output, and a decline in the sum of the squared market share of each bank in revenues (because the demand curve is downward the system. The HHI accounts for the market share sloping), leading to an H-statistic less than or equal of all banks in the system and assigns a larger weight to 0. Under perfect competition, an increase in input to the biggest banks. Instead, concentration ratios prices raises both marginal costs and total revenues completely ignore the smaller banks in the system. by the same amount (since the demand curve is per- The concentration ratio varies between nearly 0 and fectly elastic); hence, the H-statistic will equal 1. 100. The HHI has values up to 10,000. If there is A frequently used measure of markups in bank- only a single bank that has 100 percent of the mar- ing is the Lerner index, defined as the difference ket share, the HHI would be 10,000. If there were a between output prices and marginal costs (relative large number of market participants with each bank to prices). Prices are calculated as total bank revenue having a market share of almost 0 percent, the HHI over assets, whereas marginal costs are obtained would be close to zero. from an estimated translog cost function with The Panzar and Rosse (1982, 1987) H-statistic respect to output. Higher values of the Lerner index captures the elasticity of bank interest revenues to signal less bank competition. input prices. The H-statistic is calculated in two steps: The Boone indicator measures the effect of ef� - ciency on performance in terms of pro�ts. It is calcu- 1. Running a regression of the log of gross total rev- lated as the elasticity of pro�ts to marginal costs. To enues (or the log of interest revenues) on log mea- calculate this elasticity, the log of a measure of prof- sures of banks’ input prices. its (such as return on assets) is regressed against a 2. Adding the estimated coef� cients for each input log measure of marginal costs. The elasticity is cap- price. Input prices include the price of depos- tured by the coef�cient on log marginal costs, which its (commonly measured as the ratio of interest are typically calculated from the � rst derivative of a expenses to total deposits), the price of personnel translog cost function. The main idea of the Boone (as captured by the ratio of personnel expenses to indicator is that more-ef�cient banks achieve higher assets), and the price of equipment and � xed capi- pro�ts. The more negative the Boone indicator is, tal (approximated by the ratio of other operating the higher the level of competition is in the market, and administrative expenses to total assets). because the effect of reallocation is stronger. banking industry. The first indicator, an The competitive environment of the overall index of barriers to entry, summa- banking system can also be affected by rizes the information needed to obtain a the strategic reactions of banks. The new banking license. Higher index values indicate empirical industrial organization literature more stringent requirements for bank entry. provides three indicators of banks’ pricing The second indicator of contestability is the behavior.6 share of applications for bank licenses that First, the H-statistic measures the elastic- were denied. Regulations concerning entry ity of banks’ revenues relative to input prices to the banking sector are, on average, more (Panzar and Rosse 1982, 1987). Under per- stringent in developing economies than in fect competition, an increase in input prices developed ones. Between 2001 and 2010, the raises both marginal costs and total revenues share of denied banking licenses declined for by the same amount, and hence the H-statis- both groups of countries. tic equals 1. Under a monopoly, an increase 86 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 3.4 Analyzing Bank Competition Using Disaggregated Business Line Data: Evidence from Brazil Urdapilleta and Stephanou (2009) use disaggregated existence of more substitute providers (like capital business data for banks in Brazil to analyze the driv- markets or overseas banks) in the corporate sector ers of bank revenues, costs, and risks in the retail keeps loan rates and fees lower. Similarly, the study and corporate segments. The study allocates rev- cites easier access to credit information for large cor- enues, costs, and all other line items in the � nancial porations as another reason why competition in this statements of the banking system into different busi- segment is higher. ness lines. This approach results in � nancial state- Among the policies that can foster competition ments and ratios by business line. Other public data in the retail segment, the study mentions promoting sources were used and assumptions made to estimate the portability of bank accounts, permitting positive notional � nancial statements for each business line. credit information sharing, and expanding payment Interviews with senior management served as a con- system interconnection. All these allow customers to sistency check on the overall methodology. switch banks more easily and, therefore, force banks A key finding of the analysis is that the retail to compete more actively. banking segment has signi�cantly higher returns (39 The study illustrates how differences across mar- percent) than the corporate segment (16 percent), ket segments, which tend to be averaged out in an despite being riskier and costlier. In particular, aggregate analysis, need to be taken into account higher lending rates and fees more than compensate when designing public policy in banking. The study for additional expenses. also highlights that a great deal of in-depth knowl- The study argues that one of the reasons for edge of the banking sector is required to be able to lower profitability in the corporate sector is the use the practitioner approach to obtaining pro�tabil- higher degree of competition among providers in the ity measures by business line and to be able to assess segment. In particular, the study mentions how the bank competition across market segments. FIGURE 3.1 Five Bank Concentration Ratio (CR5): in input prices results in a rise in marginal Developed and Developing Economies costs, a fall in output, and a decline in rev- enues, leading to an H-statistic less than or Median asset share of 5 largest banks, % equal to 0. Panzar and Rosse (1987) show 90 that when H is between 0 and 1, the system 80 operates under monopolistic competition. 70 In general, the H-statistic is interpreted as a measure of the degree of competition in the 60 banking market.7 50 Second, the Lerner index captures the dif- 40 ference between output prices and marginal 30 costs of production—that is, the markup of 20 output prices over marginal costs (Lerner 10 1934).8 0 Finally, the Boone indicator is based on 1996–2010 2000–07 2000–10 the association between firm performance Developed economies Developing economies and ef�ciency (Boone 2001; Boone, Grif�th, and Harrison 2005; Hay and Liu 1997). See Source: Calculations based on Bankscope (database). box 3.3 for further details on the calculation GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 87 FIGURE 3.2 Five Bank Concentration Ratio (CR5): Developing Regions, Median Values, 1996–2010 Median concentration 100 90 80 70 60 50 40 30 20 10 0 East Asia Europe and Latin America GCC Middle Non-GCC Middle South Asia Sub-Saharan and Pacific Central Asia and the East and East and Africa Caribbean North Africa North Africa Source: Calculations based on Bankscope (database). of these pricing indicators of banking indicator, a measure of the effect of effi- competition. ciency on performance in terms of profits. Figure 3.4 depicts the H-statistic for An increase in the Lerner index or the Boone developed and developing economies. Bank indicator indicates a deterioration of the com- pricing behavior was more sensitive to petitive conduct of � nancial intermediaries. changes in the price of inputs among devel- Banking competition in developed economies oped compared with developing economies deteriorated initially (1996–2003), increased in 1996 –2007, indicating that banking in the run-up to the global financial crisis systems in developed economies behave (2004–08), and worsened afterward (2009– more competitively. But bank competition 10). The initial deterioration could be associ- declined in 2008–10 for developed econo- ated with the drop in competition observed mies, while it improved for developing econ- in the euro area after the adoption of the omies. It can be argued that the declining European Monetary Union (Sun 2011) and in trend in developed economies may be attrib- line with � ndings of less competitive behav- uted to the implications on industry struc- ior of banks in large and integrated �nancial ture and competitive conduct of the recent markets (Bikker and Spierdijk 2008).9 systemic banking crisis and its associated It is important to note that the simple large-scale policy responses. observation that competition increased Figure 3.5 examines the competitive before the crisis does not necessarily suggest behavior of banking systems across develop- that greater competition in itself spurred the ing regions. Latin America has the systems crisis. Recent studies suggest the problem was with the highest sensitivity of output to input that the increase in competition occurred in prices, whereas those in the Middle East and an environment where regulation and super- North Africa appear to be the least competi- vision were too lax and incentives for ade- tive (see box 3.5 for further details). quate risk management were missing (Barth, Figure 3.6 shows the evolution of the Caprio, and Levine 2012; Caprio, Demirgüç- Lerner index, a measure of market power Kunt, and Kane 2010). that compares output pricing and marginal On the other hand, the financial cri- costs (that is, markup), as well as the Boone sis—and the subsequent policy responses by 88 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 3.3 Regulatory Indicators of Market Contestability a. Entry barriers: b. Share of denied licenses: Developed versus developing economies Developed versus developing economies Average of region Average of region, % 8 25 7 20 6 5 15 4 3 10 2 5 1 0 0 Developed economies Developing economies Developed economies Developing economies 2001 2010 2001 2010 c. Entry barriers across developing regions d. Share of denied licenses across developing regions Average of region Average of region, % 8 80 7 70 6 60 5 50 4 40 3 30 2 20 1 10 0 0 ia ia ic sia GC bbe ica GC Afr ast ric t ca ic sia GC bbe ica GC Afr ast ric t ca Af as Af as As As cif cif fri fri lA ri er lA ri er an C M an an C M ica So a an C M an an C M ica So a h E th E h E th E Pa Pa nA nA th th Ca m No ort dle or ddle Ca m No ort dle or ddle ra tra u u he in A he in A d d t ra ra d d an en an en dN i dN i dN i dN i ha ha d t Lat d t Lat dC dC ia ia Sa Sa As As an an b- b- st st n- Su n- Su pe pe Ea Ea an an ro ro Eu Eu 2001 2010 2001 2010 Sources: Calculations based on Bank Regulation and Supervision Survey (database), World Bank, 2007 data; Barth, Caprio, and Levine 2001, 2004, 2006; ˇ ihák and others 2012. C Note: The index of entry into banking requirements captures whether various types of legal submissions are required to obtain a banking license. Higher scores indicate greater restrictions on entry into banking. On the other hand, the share of denied licenses is the ratio of denied to total license requests. governments—may have affected the com- and Spain). The Lerner index and the Boone petitive conduct of � nancial intermediaries indicator for developing economies evolve in in developed economies.10 Sun (2011) � nds a similar fashion, with a smoother trend in that bank competition in developed econo- the Boone indicator than the Lerner index. mies deteriorated during this period, espe- Deterioration of bank competition may cially in countries that had large credit and have taken place in spite of �nancial reforms housing booms (such as the United States across developing economies—especially in GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 89 FIGURE 3.4 Bank Competition: Developed vs. with variability across geographical regions Developing Economies (�gure 3.7). Though GCC countries in the Middle East and North Africa display the H-statistic (OLS regional estimates) least competitive banking systems, Latin 0.65 American banking systems have the most 0.60 competitive systems in developing regions. 0.55 0.50 0.45 THE IMPACT OF COMPETITION 0.40 ON THE BANKING SYSTEM 0.35 0.30 Competition affects the banking industry 0.25 along three dimensions: ef�ciency, access to 0.20 �nance, and stability. 0.15 0.10 0.05 Competition and banking ef�ciency 0.00 1996–2007 2000–07 2000–10 There are two views on the direction of cau- Developed economies Developing economies sality between competition and efficiency. The “quiet life� hypothesis argues that Source: World Bank staff, based on Bankscope (database). monopoly power allows banks to relax their Note: OLS = ordinary least squares. efforts and increases their costs, predicting a positive link from competition to ef�ciency (Hicks 1935). Alternatively, the “efficient countries with weak institutions (low bureau- structure� hypothesis predicts a negative cratic quality and low transparency) and low relationship between competition and ef�- levels of economic development (Delis 2012). ciency, where causality runs from ef�ciency The Lerner index and Boone indicator in to competition (Demsetz 1973). According developing country regions mimic the aver- to this view, better managed, more ef�cient age for developing economies—although � rms can secure the largest market shares, FIGURE 3.5 Bank Competition across Developing Regions, 1996–2007 H-statistic (OLS regional estimates) 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 East Asia Europe and Latin America GCC Middle Non-GCC Middle South Asia Sub-Saharan and Pacific Central Asia and the East and East and Africa Caribbean North Africa North Africa Source: Calculations based on Bankscope (database). Note: H-statistic �gures are calculated following the methodology described in Demirgüç-Kunt and Martínez Pería 2010. 90 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 3.5 Banking Competition in the Middle East and North Africa Banking sectors in the Middle East and North Third, the paper compares the extent of banking Africa region (MENA) are among the deepest in sector competition in the region to that observed in the developing world (see table B3.5), but are they other regions of the developing world. Finally, the competitive? paper analyzes the factors that explain differences Anzoategui, Martínez Pería, and Rocha (2010) in competition between MENA and other regions. analyze bank competition in the region in four dif- The estimations of the H-statistic and the Lerner ferent ways. First, the study analyzes two distinct index show that banking sectors in MENA operate measures of competition, the H-statistic and the under monopolistic competition. Comparisons over Lerner index, over a longer period of time, 1994– time indicate that competition within MENA, both 2008. Second, the paper examines the behavior of among Gulf Cooperation Council (GCC) countries competition in the region and tests for differences and non-GCC economies, has not improved and, in across two subperiods: 1994–2001 and 2002–08. many cases, worsened. TABLE B3.5.1 Competition in MENA and across Regions H-statistics Lerner index Regions (1994–2008) (2002–08) (1994–2008) (2002–08) Middle East and North Africa 0.520 0.482 0.320 0.373 GCC countries 0.497 0.470 0.360 0.435 Non-GCC countries 0.528 0.508 0.241 0.258 p-value GCC = non-GCC 0.640 0.640 0.050 0.010 East Asia and Paci�c 0.614 0.584 0.230 0.265 p-value East Asia and Paci�c = GCC 0.070 0.120 0 0 p-value East Asia and Paci�c = non-GCC 0.020 0.140 0.810 0.890 Eastern Europe 0.685 0.694 0.182 0.196 p-value Eastern Europe = GCC 0 0 0 0 p-value Eastern Europe = non-GCC 0 0 0.240 0.240 Latin America and the Caribbean 0.743 0.765 0.215 0.234 p-value Latin America and the Caribbean = GCC 0 0 0 0 p-value Latin America and the Caribbean = non-GCC 0 0 0.580 0.630 Former Soviet Union 0.659 0.669 0.271 0.266 p-value Former Soviet Union = GCC 0.010 0 0 0 p-value Former Soviet Union = non-GCC 0 0 0.520 0.860 South Asia 0.710 0.677 0.244 0.272 p-value South Asia = GCC 0 0.010 0.020 0 p-value South Asia = non-GCC 0 0 0.970 0.800 Sub-Saharan Africa 0.521 0.518 0.223 0.169 p-value Sub-Saharan Africa = GCC 0.700 0.510 0.040 0.020 p-value Sub-Saharan Africa = non-GCC 0.830 0.850 0.810 0.450 Note: GCC = Gulf Cooperation Council, MENA = Middle East and North Africa. Relative to other regions, MENA is lagging suggests that a worse credit information environ- behind in terms of bank competition. The evaluation ment and stricter regulations and practices govern- of the factors explaining differences in banking sec- ing bank entry are at least partly to blame. tor competition between MENA and other regions GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 91 leading to more concentration and less FIGURE 3.6 Bank Competition: Developed vs. Developing competition. Economies Although studies that examine the link between concentration and efficiency find a. Lerner index, 1996–2010 mixed results,11 the overwhelming major- Median across banks ity of recent empirical studies conclude that 0.30 competition brings about improvements in ef�ciency in both developed and developing 0.25 economies. Using data for more than 14,000 0.20 banks operating in Europe and the United States, Schaeck and Čihák (2008) find a 0.15 positive effect of competition on pro�t and cost ef�ciency. Similarly, using a technique 0.10 to obtain joint estimates of efficiency and 0.05 market power among banks in the European Monetary Union, Delis and Tsionas (2009) � nd a negative relationship, which is in line 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 with the quiet life hypothesis. Comparable � ndings are obtained when Developed economies Developing economies the sample of economies is extended to b. Boone indicator, 1997–2010 include developing economies. Using data on Median across banks net interest margins and overhead costs for 0.000 over 1,400 banks in 72 developed and devel- oping economies, Demirgüç-Kunt, Laeven, –0.005 and Levine (2004) � nd that tighter regula- –0.010 tions on bank entry and bank activities lead to higher costs of � nancial intermediation. –0.015 Lin, Ma, and Song (2010) find a similar result for 2,500 banks operating in 74 econ- –0.020 omies. Finally, focusing on 60 developing –0.025 economies, Turk-Ariss (2010) �nds a signi�- cant negative association between bank mar- –0.030 ket power (as measured by the Lerner index) 97 98 99 00 01 02 03 04 05 06 07 08 09 10 19 19 19 20 20 20 20 20 20 20 20 20 20 20 and cost ef�ciency.12 Overall, the literature examining the link Developed economies Developing economies between direct measures of competition and ef�ciency suggests that more bank competi- Source: Calculations based on Bankscope (database). Note: Lerner index estimations follow the methodology described in Demirgüç-Kunt and Martínez tion increases bank ef�ciency in both devel- Pería (2010). The regional estimates for the Lerner index are based on the median of bank esti- oped and developing economies. mates within the region. Boone indicator estimations follow the methodology used by Schaeck and Cˇ ihák (2010a) with a modi�cation to use marginal costs instead of average costs. Data are pooled by region in order to estimate the regional Boone indicator. Boone indicator data are not shown for Sub-Saharan Africa because of a lack of adequate data Competition and access to �nance Theory makes ambiguous predictions regard- ing the effect of competition on access to asymmetries and agency costs, competition finance. The conventional market power can reduce access by making it more dif- hypothesis argues that competition in the ficult for banks to internalize the returns banking market reduces the cost of � nance from investing in lending, in particular, with and increases the availability of credit. On opaque clients.13 the other hand, the information hypothesis Most of the empirical studies on this posits that in the presence of information question used concentration as a measure 92 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 3.7 Bank Competition across Developing Regions banking sectors.15 Using data on growth in value added from 1980–90 for 16 countries, and measuring competition at the country- a. Lerner index, 1996–2010 level (using the Panzar and Rosse H-statistic), Median across banks 0.55 Claessens and Laeven (2005) � nd that com- 0.50 petition is positively associated with indus- trial growth. They suggest that competi- 0.45 tive banking sectors are better at providing 0.40 �nancing to �nancially dependent �rms. 0.35 Exploiting a rich dataset on small and 0.30 medium-sized enterprises in Spain, Carbó- 0.25 Valverde, Rodríguez-Fernández, and Udell 0.20 (2009) also find evidence that competition 0.15 promotes access to �nance, using the Lerner 0.10 index.16 In sum, similar to the findings on 0.05 the link between competition and ef�ciency, 0.00 the evidence that measures bank competition 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 directly suggests that competition is ben- 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 e�cial for the banking sector. In particular, b. Boone indicator, 1997–2010 bank competition enhances access to credit. Median across banks 0.55 Competition and banking stability –0.01 Competing theories explain the link between –0.02 competition and stability.17 The traditional view predicts that competitive banking sys- –0.03 tems are less stable because competition reduces bank profits and erodes the char- –0.04 ter value of banks, consequently increasing –0.05 incentives for excessive risk taking (Chan, Greenbaum, and Thakor 1986; Keeley 1990; –0.06 Marcus 1984). Furthermore, in more com- –0.07 petitive environments, banks earn lower informational rents from their relationship 97 98 99 00 01 02 03 04 05 06 07 08 09 10 19 19 19 20 20 20 20 20 20 20 20 20 20 20 with borrowers, reducing their incentives to East Asia and Pacific Non-GCC Middle East properly screen borrowers, again increasing Europe and Central Asia and North Africa the risk of fragility (Allen and Gale 2000, Latin America and the Caribbean South Asia 2004; Boot and Greenbaum 1993). Competi- GCC Middle East and North Africa Sub-Saharan Africa tion can also destabilize the banking sector through its impact on the interbank market Source: Calculations based on Bankscope (database). and the payments system. Note: Lerner index estimations follow the methodology described in Demirgüç-Kunt and Martínez Pería (2010). The regional estimates for the Lerner index are based on the median of bank esti- For example, if all banks are price-takers mates within the region. Boone indicator estimations follow the methodology used by Schaeck in a competitive market, banks have no incen- and Cˇ ihák (2010a) with a modi�cation to use marginal costs instead of average costs. Data are pooled by region in order to estimate the regional Boone indicator. Boone indicator data are not tives to provide liquidity to a troubled bank, shown for Sub-Saharan Africa because of a lack of adequate data. leading to bank failure, and creating negative repercussions for the entire sector (Allen and Gale 2000). A somewhat different argument of competition, obtaining mixed results.14 in support of the competition-fragility view But studies that focus on direct measures is that more concentrated banking systems of competition and contestability show that have larger banks, which in turn allow them access to �nance is easier in more competitive to diversify their portfolios better. A final GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 93 argument refers to the number of banks to be 2007a) � nd that more competitive banking supervised. Given that a more concentrated systems (de� ned as those with fewer regula- banking system typically implies a smaller tory restrictions on bank entry and activities) number of banks, this might reduce the are less likely to suffer systemic banking dis- supervisory burden and enhance the overall tress. This � nding is con� rmed by Schaeck, stability of the banking system. Čihák, and Wolfe (2009), who �nd a negative The competition-stability view argues relationship between bank competition and that market power in banking boosts prof- systemic bank fragility using the H-statistic its and stability, yet ignores the potential to measure competition. Schaeck and Čihák impact of market power on borrower behav- (2010b) identify bank capitalization as one ior (Boyd and de Nicoló 2005). Because of the channels through which competition banks in less competitive sectors can charge fosters stability. Using data for more than higher interest rates, this may induce � rms 2,600 European banks, they show that banks to assume greater risk—resulting in a have higher capital ratios in more competitive higher probability that loans become non- environments. This is consistent with Berger, performing. Similarly, higher interest rates Klapper, and Turk-Ariss (2009), who find might attract riskier borrowers through the that banks in more competitive banking sys- adverse selection effect. Thus, in contrast tems take greater lending risks, but compen- to the charter-value hypothesis, the com- sate with a higher capital-asset ratio, result- petition-stability view predicts that bank ing in an overall lower level of bank risk, as actions will result in more risk taking and measured by the z-score. greater fragility in more concentrated and Measures of bank risk, such as the less competitive banking systems. Advo- z-score, ignore systemic stability, but regu- cates of the competition-stability view also lators are concerned with systemic stabil- disagree with the notion that concentrated ity much more than the absolute level of banking systems are easier to monitor than risk of individual banks. In a recent paper, less concentrated banking systems with Anginer, Demirgüç-Kunt, and Zhu (2012) many banks, since larger banks can be more introduce a new measure of systemic risk complex and, hence, harder to supervise. taking by banks. Using Merton’s 1973 con- The early empirical literature on the link tingent claim pricing framework, they cal- between competition and stability is mixed. culate the default probability for each bank Some country studies have shown that in the system. They measure systemic risk increasing competition leads to greater indi- as the codependence in default probability vidual bank risk taking.18 In the context of across banks. After controlling for various the U.S. subprime crisis, Dell’Ariccia, Igan, bank- and country-level variables, Anginer, and Laeven (2012) document that the rapid Demirgüç-Kunt, and Zhu (2012) � nd a posi- growth of credit in U.S. mortgage markets tive relationship between competition and in the run-up to the crisis was accompanied systemic stability. They also show that lack by a reduction in lending standards (lower of competition (as measured by the Lerner loan application denial rates), which they index) has a more adverse effect on systemic argue was in part explained by the entry of stability in countries with low levels of for- new and large lending institutions. How- eign ownership, weak investor protection, ever, some previous studies failed to �nd that generous safety nets, and weak regulation larger banks are less likely to fail as would be and supervision. predicted by the competition-fragility view The advantages of competition in an ef�- (Boyd and Graham 1991, 1996; Boyd and cient and inclusive � nancial system are sig- Runkle 1993; De Nicoló 2000). nificant. Recent studies provide evidence On the other hand, studies using cross- questioning the conventional view that com- country, time-series data sets offer evidence petition is bad for stability. Importantly, supporting the competition-stability view. policy bodies such as the OECD Competition Beck, Demirgüç-Kunt, and Levine (2006, Committee have suggested that to promote 94 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 banking stability, policy makers should practices, the state can shape the information design and apply better regulations and environment and influence the extent of bank supervisory practices rather than limit bank competition. competition (OECD 2010). The institutional environment can also have an impact on bank competition. For example, to the extent that corruption is ram- DRIVERS OF BANK COMPETITION pant in the economy, there will be less scope The main drivers of competition are entry for a level playing �eld in the �nancial sector, and exit policies, underlying information and and competition will suffer as a result. Simi- institutional environment, and competitive larly, to the degree that creditor rights are not pressures in the � nancial sector.19 The state protected, there will be less incentive for new can directly influence all three. The state banks to enter the banking sector. The state can also affect bank competition by owning directly influences the institutional environ- banks. Box 3.6 analyzes the determinants of ment by the laws that it promotes and the banking competition across economies. extent to which it upholds compliance. Entry and exit policies in banking are All else being equal, the entry of foreign important for competition because they keep banks and the presence of nonbank interme- incumbents on their toes. The threat of entry diaries are likely to affect bank competition. and exit to the industry forces banks to worry Foreign banks often bring new technologies about providing good, affordable products and new products to banking sectors, creat- and limits their ability to exercise market ing an incentive for local banks to compete. power. Entry policies include regulations on Similarly, the presence of a liquid stock mar- licensing, as well as the practice by regulators ket or other �nancial intermediaries that can of approving new licenses. Exit policies refer provide � nancing to � rms is likely to foster to regulations as well as the measures taken competition in the banking sector, because by regulators to close insolvent banks. The banks will have to compete to provide �nan- state can directly affect bank competition by cial services to � rms. Once again, the state promoting policies and practices that facili- has a role to play here by introducing regula- tate bank entry and exit. A delicate balance tions and practices that foster the entry and needs to be struck where regulators foster operation of nonbank competitors. contestability (to streamline requirements for Finally, government ownership of banks bank licensing, speed up the licensing pro- can also affect bank competition. On the cess, and implement ef�cient bank resolution) one hand, government banks can spur com- without jeopardizing bank stability (that is, petition if (because they typically do not maintaining a licensing process that keeps maximize pro�ts) they push other banks to out un�t bankers). lower prices. On the other hand, if govern- Access to credit history information about ment banks dominate the system and other potential borrowers also facilitates compe- banks are crowded out, competition falls. tition in the banking sector. Dell’Ariccia, Box 3.6 shows that banking systems are Friedman, and Marquez (1999) show that to more competitive in countries with lower the extent that access to credit information entry barriers, greater foreign bank par- is restricted, incumbent banks are better able ticipation, and more developed capital mar- to exercise market power and limit competi- kets (which are also associated with greater tion. At the same time, greater disclosure of development of nonbank � nancial interme- information regarding the terms of banking diaries). 20 Greater information disclosure, products will generate greater awareness by as captured by depth of credit information, bank clients and promote bank competi- also promotes competition. Box 3.7, on the tion. By promoting the establishment and other hand, highlights the importance of operation of credit bureaus and by having in consumer protection measures to enhance place consumer protection regulations and banking competition. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 95 BOX 3.6 An Econometric Analysis of Drivers of Bank Competition The Lerner index is used as the summary measure institutional framework is measured by the index of of competition across 83 countries and is regressed control of corruption, which measures the degree to against variables capturing the following: which public power is exercised for private bene�t— such as state capture by elites and private interests 1. Entry and exit policies in the banking sector. (Kaufmann, Kraay, and Mastruzzi 2009). 2. The information and institutional environments. Competitive pressures from within and outside 3. Competitive pressures from within and outside the the banking sector are captured by indicators of banking sector. Bank-level Bankscope data for the market structure, presence of foreign banks, liquid- period 2000–10 are used to compute the Lerner ity of the stock market, and importance of nonbank index for all countries with at least �ve banks. intermediaries such as insurance companies and pension funds. Market structure is measured by the The � rst group of barriers to entry comprise mea- share of assets held by the top �ve banks in the sys- sures such as the number of documents and proce- tem. To the extent that there is some validity to the dures required to obtain a banking license, the per- structure-conduct-performance paradigm, this vari- centage of denied applications for banking licenses, able is expected to be positive and signi�cant. Cross- and the minimum entry capital required for banks. border banking as captured by the share of banking The second group includes a synthetic indicator of assets held by foreign banks may refl ect a greater activity restrictions that captures a bank’s ability to degree of market contestability. To the extent that engage in activities other than banking (say, securi- this is the case, promoting foreign bank participa- ties, insurance, or real estate). Higher values of this tion may increase competition in the industry. On index represent greater restrictions. Both indexes are the other hand, if foreign bank entry is associated constructed with data from the World Bank 2006 with mergers and acquisitions, it might not enhance Bank Regulation and Supervision Survey. competition. Finally, a more competitive banking The availability of credit history information is system may arise from greater interindustry compe- captured by the depth of credit information index tition. In short, the development of nonbank � nan- gathered from Doing Business (see http://www. cial intermediaries may affect the market power doingbusiness.org). This index takes values from 0 of the banking sector. The relative importance of to 6, and higher values indicate greater availability such intermediaries is approximated by the value of of information. The extent to which the government shares traded to gross domestic product (GDP), the requires that � nancial institutions disclose informa- value of life insurance premiums to GDP, and the tion about � nancial contracts to potential users of share of pension fund assets to GDP. this service is measured by an index on the strict- The table below suggests that the banking sec- ness of � nancial contract disclosure requirements, tor is more competitive (the Lerner index is lower) which was constructed based on questions from in countries with greater contestability (lower entry the World Bank Regulation and Supervision Sur- barriers), greater information disclosure, better vey. This contract disclosure index is also a measure institutions, more foreign bank participation, and of consumer protection. The quality of the overall more liquid stock markets. (Continued on next page) STATE INTERVENTIONS holidays, provide blanket guarantees, inject DURING CRISES AND BANKING capital, increase deposit insurance coverage COMPETITION limits, and extend liquidity support to banks During crises, governments, central banks, on an unprecedented scale (Laeven and Valen- and other authorities in charge of the super- cia 2008, 2010). During tranquil periods, the vision and regulation of financial institu- competitive effects of rescue operations on tions introduce deposit freezes, declare bank individual banks (such as capital injections, 96 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 3.6 An Econometric Analysis of Drivers of Bank Competition (continued) TABLE B3.6.1 Cross-Country Determinants of Banking Competitiona Variable [1] [2] [3] [4] [5] [6] State as regulator: Market contestability Entry barriers 0.0401* 0.0376 0.013 0.0279 0.0242 0.0537** [0.024] [0.023] [0.028] [0.021] [0.019] [0.021] Share of bank licenses denied –0.0025 0.0229 –0.0616 –0.0132 –0.0589 [0.038] [0.037] [0.049] [0.042] [0.043] Restrictions on bank activities –0.0074 –0.0049 –0.01 –0.0087 –0.009 [0.005] [0.005] [0.008] [0.006] [0.006] Minimum entry capital required (ln) 0.0025** 0.0022* 0.0011 0.0006 0.0005 0.0031** [0.001] [0.001] [0.001] [0.001] [0.001] [0.001] State as enabler of market-friendly environments Depth of credit information –0.0110* –0.0130** [0.006] [0.006] Strictness of �nancial contract –0.0074 disclosure requirements (0-4) [0.010] Control of corruption –0.0291** –0.0172 –0.0299* –0.0318*** –0.0235* –0.0193 [0.011] [0.013] [0.017] [0.011] [0.012] [0.015] State as bank owner Government bank participation –0.0417 –0.0793* –0.0854 –0.044 –0.0394 –0.0734 [0.046] [0.044] [0.064] [0.047] [0.047] [0.048] Competitive pressures within the banking sector and from other parts of the �nancial sector Foreign bank participation –0.0594** –0.0677 –0.0614 –0.0859*** –0.0715** –0.0822** [0.028] [0.034] [0.050] [0.029] [0.030] [0.034] Concentration (CR5) 0.1119** 0.0816 0.0191 [0.055] [0.063] [0.065] Stock market value traded –0.0371* –0.0616** (ratio to GDP) [0.020] [0.023] Life insurance premium 0.3105 (ratio to GDP) [0.445] Pension fund assets 0.0448 (ratio to GDP) [0.031] Constant 0.0178 0.0195 0.2706 0.0503 0.1403 –0.0858 [0.191] [0.189] [0.221] [0.176] [0.152] [0.158] Observations 83 71 38 72 64 42 R-squared 0.189 0.231 0.369 0.273 0.355 0.399 Note: GDP = gross domestic product. a. A country's Lerner index is the median estimate across banks over the period 2000–10. Robust standard errors are in brackets. ***p < 0.01 **p < 0.05 *p < 0.1 emergency liquidity facilities, and assisted affect large numbers of institutions, with mergers) tend to be relevant only for a limited potential implications for industry structure number of distressed institutions and their and competitive conduct in �nancial systems competitors (Gropp and others 2011; Hak- over longer periods of time. enes and Schnabel 2010). However, episodes An emerging body of research exam- of systemic banking crises frequently result ines the effects of bank bailouts and other in large-scale, repeated policy responses that policy responses on risk taking at the bank GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 97 BOX 3.7 Consumer Protection and Competition in South Africa Among emerging markets, South Africa has a well- ments and barriers to entry in the payments systems, developed � nancial sector and one of the largest cap- but the inquiry put special emphasis on measures of ital markets. In 2010, the outreach of the banking consumer protection. sector, as measured by its credit to the private sector, Advances in consumer protection are justified totaled 145 percent of GDP, while stock market cap- on the grounds that they can promote competition italization to GDP amounted to 278 percent of GDP. and depth. Providing better information to custom- In recent years, major South African banks have also ers can lead to rising price and product competi- expanded throughout the region—notably, Absa and tion among banking intermediaries. In this context, Standard Bank (Beck and others 2011). However, South Africa established a series of mechanisms the banking sector is heavily concentrated, with the and institutions that promoted a more sound infor- �ve largest banks accounting for over 90 percent of mation environment. The National Credit Act 34 total assets and deposits in the system in 2010. Esti- (enacted in 2005), for instance, provides a general mates of market power in the South African banking framework to promote responsible lending practices, industry show that there is evidence of monopolis- protect South African consumers from unfair credit tic competition (Greenberg and Simbanegavi 2009; and credit marketing practices and, more generally, Mlambo and Ncube 2011), which is consistent with establish norms and standards on consumer credit. the fact that large banks tend to avoid competition It also created the National Credit Regulator to among themselves, as reported by the Competition ensure the law’s compliance, investigate complaints, Commission on Banking (OECD 2008). promote � nancial literacy, and provide a knowledge The lack of competition in the South African platform on credit practices. On the other hand, banking sector has been documented in several cases of noncompliance and appeals to decisions reports prepared for the National Treasury and the of the regulators were allowed to be presented and South African Reserve Bank (Falkena and others solved by the National Consumer Tribunal. 2004; Competition Commission of South Africa In spite of the advances on abusive lending prac- 2006). As manifested by high prices and poor qual- tices, the mechanisms established by the National ity of financial services, low rates of innovation, Credit Act to restructure consumer debt have been and � nancial exclusion, the lack of competition was slow. For instance, Beck and others (2011) point attributed partly to high concentration and pro�t- out that as few as 5 percent of the 150,000 applica- ability in retail banking and payments (OECD tions made by overly indebted consumers have been 2010). This led to the Banking Enquiry launched by � nalized by courts. Finally, better use of consumer the Competition Commission in August 2006. The protection networks requires � nancially aware con- result of this inquiry led to several recommendations sumers. Efforts to raise the effectiveness of �nancial to address problems of restrictive interbank arrange- literacy programs are required. Sources: Beck and others 2011; Greenberg and Simbanegavi 2009; Mlambo and Ncube 2011; OECD 2010. level (Berger and others 2010; Duchin and entrenchment of the supported institutions; Sosyura 2011; Farhi and Tirole 2012; Gropp, assisted mergers of large �nancial institutions Hakenes, and Schnabel 2011; Hakenes and increase concentration, presumably reducing Schnabel 2010; Hoshi and Kashyap 2010; competition in retail markets and reinforcing Richardson and Troost 2009). However, how the perception that these banks are too big these actions affect competition has received to fail (Beck and others 2010; Hakenes and less attention. This issue is of vital impor- Schnabel 2010). tance because of the unintended (and pos- More research is clearly needed in this sibly detrimental) effects for consumer wel- area, though a number of studies suggest fare. For example, guarantees can result in that state interventions that favor some bank 98 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 services (such as guarantees, liquidity sup- number of banks in the system can easily lead port, recapitalizations, and nationalizations) to lower levels of bank competition (OECD force competitors to behave more aggres- 2009). In that case, governments should pro- sively, leading to decreased margins and mote the entry of deserving institutions to increased competition (Gropp, Hakenes, and mitigate the negative impact of mergers. Schnabel 2011; Hovakimian and Kane 2000; Overall, it is clear that state interven- Kane 1989). Using data for 138 countries that tions during crises can have an impact on witnessed a variety of policy responses dur- bank competition and potentially on future ing 46 banking crises, Calderón and Schaeck banking stability. Governments should avoid (2012) �nd that Lerner indexes and net inter- insolvency resolution policies that not only est margins drop as a result of state interven- distort risk-taking incentives and jeopardize tions such as guarantees, liquidity support, future stability, but also have implications for recapitalizations, and nationalizations. the level of competition in the banking sector. This apparent increase in competition should be interpreted with caution—espe- IMPLICATIONS FOR THE DESIGN cially if the type of state intervention under OF COMPETITION POLICIES consideration delays the exit of inefficient and insolvent banks. Kane (2000) argues Bank competition increases efficiency and that some state interventions can constitute �nancial inclusion. Recent evidence suggests a barrier to exit by allowing these banks that bank competition can even enhance sys- to survive beyond their “natural death.� temic � nancial stability. Hence, bank com- State support to these zombie banks—a petition should not be restricted with the term coined by Kane (1989)—would allow hope of promoting stability. Instead, the state these institutions to bid up deposit rates and should design and enforce regulations that accept low interest rates on high-risk loans create the right incentives to safeguard stabil- and investments, thus reducing pro�t mar- ity, while at the same time promote competi- gins in the industry (Kane 2000; Kane and tion and ef�ciency. Rice 2001). In sum, states can create zombie The state can shape bank competition banks by distorting risk-taking incentives of through its actions as a regulator and an the system and generating unhealthy com- enabler of a market-friendly and information- petition. Calderón and Schaeck (2012) show rich environment. In particular, banking that the increase in competition that might sectors are more competitive in economies result from state interventions during crises where the state designs, implements, and is also accompanied by other negative conse- enforces regulatory frameworks that ensure quences. Despite the evidence that the cost of greater contestability. More speci�cally, poli- borrowing is reduced and credit is restored as cies designed to ease the entry of deserving a result of these state interventions, access to institutions (those that can pass �t and proper credit by opaque borrowers such as small and tests) and promote timely exit may prevent medium enterprises is reduced. incumbent banks from exercising market To avoid the distorting effects of state power and lead to a more competitive envi- interventions, sunset clauses and exit plans ronment. Related to the less stringent barriers are important (Beck, Coyle, and others 2010). to entry, foreign bank penetration may also By providing credible signals that interven- be conducive to greater competition. tions are temporary, governments can reduce The state can also promote competition the negative repercussion for competition. in the banking sector by ensuring banks’ Similarly, addressing governance deficien- and consumers’ access to information as cies in the supported institutions can also be well as by building up a sound institutional important in reducing incentives for excessive framework that levels the playing �eld. Free risk taking. Finally, measures such as gov- flow of credit information among banks and ernment-sponsored mergers that reduce the transparency of financial products offered GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N 99 by banks to potential consumers should be and full integration—as opposed to differ- ensured by the state. A sound institutional ent types of partial integration supervisory framework that ensures the enforcement of models. contracts, property rights, and the rule of law Competition can bring important bene�ts that limits the exercise of public power for to the �nancial sector and should not be sac- private bene�t is conducive to greater com- ri�ced for the sake of stability. Instead gov- petition in the banking sector. Finally, poli- ernments should implement the measures dis- cies to promote deeper and more diversi�ed cussed in this chapter to monitor competition financial markets—especially the develop- and strengthen the information and institu- ment of nonbanking � nancial institutions— tional frameworks to promote competition, also appear to increase the level of competi- while ensuring that regulations and super- tion in the banking sector. visory practices are in place that safeguard Government interventions during crises banking stability. that prevent the exit of insolvent institutions and increase market power through mergers can also affect competition and bank risk NOTES taking. In trying to mitigate the impact of 1. The chapter focuses on competition in the �nancial crises, governments should be aware banking sector rather than the broader �nan- of the potentially negative consequences of cial sector. But it does touch on the impact of their actions on bank competition and future nonbank intermediaries on bank competition. bank stability. See Motta (2004) for a broader discussion on Finally, implementation of competition the theory and practice of competition policy. policies depends on the institutional arrange- 2. For example, in February 2010, the OECD ment in place. The increased integration Competition Committee held a discussion on across different parts of the �nancial industry competition, concentration, and stability in has led to a shift in �nancial supervision from the banking sector. a sector-based approach to more integrated 3. Several developed and developing economies approaches, including (a) a fully integrated have competition agencies that have mandates supervisory model with one agency (a Finan- or that can influence market outcomes in banking. Competition agencies enforce anti- cial Supervision Authority) carrying out all trust laws (for example, assess the competitive supervisory roles (such as microprudential, harm of mergers, deter anticompetitive behav- business conduct, and competition policy), or ior, and minimize distortions from state aid) (b) a so-called functional or objective-based and promote measures to enable �rm entry approach in which sectorally integrated agen- and rivalry (that is, competition advocacy). cies undertake different supervisory roles.21 4. In general terms, the view of competition An example of the latter is the twin peaks presented here and discussed in the litera- model. One agency is responsible for pruden- ture is based on the notion that banks pri- tial supervision in the � nancial system, and marily compete in deposit and loan markets. another one oversees market conduct, con- However, in practice, especially in developed sumer protection, and corporate governance economies, banks offer a variety of services in all sectors.22 The effectiveness of the twin (such as market making, asset management, peaks approach, as argued by its advocates, is and underwriting) where market power may arise. Payment systems are another area in guaranteed by having a clear focus and divi- which there might be signi�cant deviations sion of roles, to minimize turf battles between from marginal cost pricing. agencies, as well as strong collaboration, to 5. See Cetorelli (1999), Claessens and Laeven work together in overlapping areas (Kremers (2004), and Demirgüç-Kunt, Laeven, and and Schoenmaker 2008, 2010). Evidence so Levine (2004), among others. Nevertheless, far suggests that regulatory quality is stron- concentration measures are presented �rst ger in systems with objectives-based super- because they are the most widely used and eas- vision—in favor of the twin peaks model iest to compute of measures of competition. 100 T H E R O L E O F T H E S TAT E I N P R O M O T I N G B A N K C O M P E T I T I O N GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 6. There is a growing literature on measuring tration facilitates access credit, whereas Beck, and explaining bank competition using direct Demirgüç-Kunt, and Maksimovic (2004) measures of competition: Anzoategui, Mar- and Chong, Lu, and Ongena (2012) �nd the tínez Pería and Melecký (2010); Anzoategui, opposite result using data for 74 countries Martínez Pería, and Rocha (2010); Beck, de and for Chinese small and medium enter- Jonghe, and Schepens (2011); Berger, Klap- prises, respectively. per, and Turk-Ariss (2009); and Delis (2012); 15. At the same time, using bank-level data, Beck, Demirgüç-Kunt and Martínez Pería (2010); Demirgüç-Kunt, and Martínez Pería (2008) Schaeck and Čihák (2008, 2010a, 2010b); �nd that barriers to banking (minimum bal- Schaeck, Čihák, and Wolfe (2009); Turk- ances to open deposit accounts and to obtain Ariss (2010), among others. loans, as well as documentation require- 7. Note that the H-statistic can only be used to ments to access �nancial services) are higher test the hypothesis of perfect competition if in countries with greater entry restrictions for the market is in long-run equilibrium (returns banks. on bank assets are not related to input prices). 16. At the same time, the authors �nd that their However, tests using the H-statistic for the results for the Lerner index are not consistent null of monopoly are still valid, since the with results using concentration as a measure long-run pro�t condition does not apply in of competition. They conclude that “research- the case of a monopoly. ers and policymakers need to be very careful 8. Measuring marginal costs is dif�cult and in drawing strong conclusions about market requires certain assumptions about the cost power and credit availability based on analy- function of banks. Typically, studies that cal- ses that rely exclusively on concentration as a culate the Lerner index assume a translog cost measure of market power.� function. 17. See Beck (2008) for a thorough review of the 9. Bikker and Spierdijk (2008) argue that banks theoretical and empirical literature on bank in large and integrated �nancial markets (such competition and stability. as in Europe after establishing the European 18. See, for example, Keeley (1990) and Dick Monetary Union) are pushed by rising capi- (2006), in the case of the United States; tal market competition and tend to shift from Jimenez, Lopez, and Saurina (2007) in the traditional intermediation to more sophisti- case of Spain. cated and complex products associated with 19. See Anzoategui, Martínez Pería, and Rocha less price competition. 2010; Claessens and Laeven 2004; Delis 2012; 10. In recent work, Calderon and Schaeck (2012) Demirgüç-Kunt and Martínez Pería 2010. use data for 138 countries, of which 43 expe- 20. Rocha, Arvai, and Farazi (2011) �nd that the rienced banking crises. Their analyses show lack of competition in the Middle East and that government interventions (such as North Africa banking sectors is the outcome blanket guarantees, liquidity support by the of barriers to entry and lack of competition central bank, recapitalizations, and nation- from nonbanking �nancial intermediaries, alization of banks) during crises signi�cantly among other factors. increase competition in banking systems, and 21. Čihák and Podpiera (2008) discuss the differ- the distortionary effects cannot be reversed ent types of supervisory arrangements. easily. 22. Countries with a twin peaks supervisory 11. See Berger (1995), Goldberg and Rai (1996), structure are Australia (Australia Prudential and Berger and Hannan (1998), among Regulation Authority and the Australian others. Securities and Investments Commission) and 12. One exception is Casu and Girardone (2009), the Netherlands—with the Netherlands Bank who �nd a positive causation between market (DNB) in charge of prudential supervision power and ef�ciency for 2,701 banks operat- and the Authority for the Financial Mar- ing in France, Germany, Italy, Spain, and the kets supervising market conduct (Čihák and United Kingdom from 2000 to 2005. Podpiera 2008; Kremers and Schoenmaker 13. See Petersen and Rajan (1995) and Marquez 2008). (2002). 14. Using U.S. data, Petersen and Rajan (1995) and Zarutskie (2006) �nd that bank concen- 4 Direct State Interventions • Lending by state-owned banks is less procyclical than lending by private banks, and some state banks played a countercyclical role during the global �nancial crisis. How- ever, this lending did not always target the most constrained borrowers and contin- ued even after economic recovery ensued, questioning the effectiveness of the policy. Furthermore, the evidence based on previous episodes of downturns and recoveries is mixed. • Moreover, research �nds that efforts to stabilize aggregate credit by state-owned banks come at a cost, particularly through the deterioration of the quality of intermediation and resource misallocation. This effect undermines the bene�ts of using state banks as a countercyclical tool. • The empirical evidence largely suggests that government bank ownership is associated with lower levels of �nancial development and slower economic growth. Policy makers need to avoid the inef�ciencies associated with government bank ownership by pay- ing special attention to the governance of these institutions and ensuring, among other things, that adequate risk management processes are in place, which is particularly chal- lenging in weak institutional environments. • Another popular form of intervention during the recent crisis was through credit guar- antee programs. Rigorous evaluations of these programs are rare, but existing studies suggest that the bene�ts are modest and costs are often signi�cant. Success hinges on overcoming the challenges of getting the design right, particularly in underdeveloped institutional and legal settings. tate-owned banks were typically cre- underserved segments of the economy—nota- S ated to ful�ll long-term development roles by �lling market gaps in long- term credit, infrastructure, and agriculture bly, small and medium enterprises (SMEs).1 In practice, however, widespread evidence shows that state banks have generally been very inef- �nance, and to promote access to �nance for �cient in allocating credit, more often than GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 101 102 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 not serving political interests. Nevertheless, discusses public credit guarantees that were the recent global �nancial crisis underscored also widely used to offset the credit crunch the countercyclical role of state-owned banks and mentions other policies adopted by in offsetting the contraction of credit from central banks to deal with the recent crisis. private banks, leading to arguments that this The main messages from this chapter are as function is an important one that can poten- follows. tially better justify their existence. First, lending by state-owned banks is During the recent global � nancial crisis, less procyclical. During the recent global countries pursued a variety of strategies to � nancial crisis, state-owned banks in some restart their �nancial and real sectors. As the countries expanded their lending portfolios balance sheets of private banks deteriorated and were credited with assisting in the eco- and they curtailed their lending activities, nomic recovery. However, it is unclear that many countries used state-owned banks to the recent crisis illustrates that state-owned step up their � nancing to the private sector banks can effectively play a countercyclical (for example, Brazil, China, and Germany).2 role. Lending growth continued even after Some economies relied heavily on the use economic recovery was under way, and loans of credit guarantee programs. For exam- were not directed to the most constrained ple, Canada, Chile, Finland, Germany, the borrowers. Furthermore, the evidence from Republic of Korea, Malaysia, and the Neth- previous crises on this issue is also mixed. erlands extended new and special schemes or Any bene�ts of credit stabilization that may refueled existing ones to alleviate the impact arise from countercyclical lending by state- of the credit crunch on SMEs. Finally, other owned banks come at a cost. The state faces countries, such as the United States, United a direct cost to the extent that outlays to the Kingdom, and those of the euro area, adopted � nancial sector (such as state treasury pur- a number of unconventional monetary and chases of securities, recapitalization, and so �scal measures to prop up credit markets.3 forth) and public debt burdens are piling up.5 The crisis and the actions adopted by Perhaps more important, � nancial interme- different countries reignited the debate on diation by state-owned banks is subject to whether there is a need for direct government the risk of political capture and often leads intervention in the �nancial sector. Support- to a deterioration of the quality of loans and ers of state-owned banks argue that these misallocation of resources in the medium to institutions provide the state an additional long term. The bulk of the empirical evidence tool for crisis management and, relative to suggests that state ownership of banks tends central banks, they may be more capable of to be associated with lower � nancial devel- undertaking the role of safe haven for retail opment, greater financial instability, and and interbank deposits, creating a �re break slower rates of economic growth. Hence, the to mitigate contagion, and stabilizing aggre- trade-off between the potential bene�ts of a gate credit. countercyclical role of state banks and the On the other hand, those who oppose long-term adverse impact on credit allocation government bank ownership point out that requires careful consideration. agency problems and politically motivated Second, focusing on the governance of lending render state-owned banks breed- state-owned banks ideally may help policy ing grounds for corruption. Furthermore, makers address the inef�ciencies associated rewarding political cronyism may build up with these institutions. Policy makers need to large �scal liabilities and threaten public sec- design a clear mandate, allow the institutions tor solvency and �nancial fragility, as well as to work as a complement to (rather than a misallocate resources and retard development. substitute for) the private banks, and adopt This chapter reevaluates the merits of risk management practices that allow them government bank ownership in the wake of to guarantee a � nancially sustainable busi- the global � nancial crisis.4 The chapter also ness. However, these governance reforms are GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 103 particularly challenging in weak institutional FIGURE 4.1 Trends in Government Ownership environments. of Banks Third, credit guarantee schemes were also a popular intervention tool used during the Asset share of government-owned banks in the recent crisis. However, because these pro- financial system (%) grams are small scale, they are not likely to 25 have a large macroeconomic impact. Fur- thermore, research suggests that the bene�ts 20 of these programs in terms of � nancial and economic “additionality� are rather mod- 15 est, and they may bring along signi�cant �s- cal and economic costs.6 The effectiveness of 10 these programs relies on their design. Best practices include (a) leaving credit assess- 5 ments and decision making to the private sector, (b) capping coverage ratios and delay- 0 Developed economies Developing economies ing the payout of the guarantee until recov- 2001–07 2008–10 ery actions are taken by the lender so as to minimize moral hazard problems, (c) pricing Sources: Calculations based on Farazi, Feyen, and Rocha 2011; Bank guarantees so they take into account the need Regulation and Supervision Survey (database), World Bank 2003, 2007, 2011; Barth, Caprio, and Levine 2001, 2004, 2008). for � nancial sustainability and risk minimi- zation, and (d) encouraging the use of risk management tools. be attributed to the recent bailouts, mergers, The rest of the chapter is organized as fol- recapitalizations, and nationalization of dis- lows. It �rst presents data on bank ownership tressed � nancial institutions that were more structure worldwide, highlighting the regions common among developed economies than where state-owned bank presence is signi�- developing economies. cant. Next, it analyzes the countercyclical Though government bank ownership is nature of government bank lending during more prevalent in the developing world, it has the recent crisis and in previous episodes. It declined considerably over time (�gure 4.2). also discusses the evidence on the longer-term Since the 1970s, the share of state-owned implications of government bank ownership banks relative to the total assets of the bank- and then turns to credit guarantee schemes. ing system declined sharply in all emerging regions, from an average of 67 percent in 1970 to 22 percent in 2009. The retrench- OWNERSHIP STRUCTURES ment of government participation has been AROUND THE WORLD dramatic in the Eastern European and Cen- State-owned banks account for less than 10 tral Asian region. The massive privatization percent of banking system assets in devel- program in transition economies launched oped economies and double that share in in the early 1990s reduced the government developing economies (�gure 4.1). In a com- stake in the banking system from almost parison of the periods 2001–07 and 2008– full ownership (88 percent in 1985) to inter- 10, the recent global crisis brought about a mediate levels of government participation small surge in government bank ownership (20 percent in 2009).7 Finally, the World in developed countries (6.7 percent to 8 Bank report Financing Africa illustrates the percent, respectively), while the opposite is dramatic transformation of the ownership true in developing economies (government structure in African banks over the last 50 bank ownership declined from 20.5 to 17.3 years (Beck and others 2011). From a system percent). The uptick in the share of state- mostly dominated by state-owned banks, the owned banks in developed countries may region now has the second highest share of 104 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 4.2 Government Ownership across Developing Regions, 1970–2009 Asset share of government-owned banks in the financial system (%) 100 80 60 40 20 0 East Asia Europe and Latin America Middle East and South Asia Sub-Saharan and Pacific Central Asia and the Caribbean North Africa Africa 1970 1985 1995 2002 2009 Sources: Calculations based on Farazi, Feyen, and Rocha 2011; Bank Regulation and Supervision Survey (database), World Bank 2003, 2007, 2011; Barth, Caprio, and Levine 2001, 2004, 2008. foreign-owned banks—trailing only the tran- have an asset market share between 20 and sition economies of Eastern Europe and Cen- 50 percent. While the average market share tral Asia.8 of government-owned banks is negligible The retrenchment of state-owned banks in in Eastern European nations, it is still large the �nancial sector can be attributed to their among countries in Central Asia.10 Almost poor � nancial performance as well as their half of the countries that responded to the lat- less than stellar contribution to �nancial and est round of the World Bank’s Bank Regula- economic development—especially in coun- tion and Supervision Survey report a market tries where they dominated the banking sys- share for state-owned banks below 5 percent. tem. This trend also reflects retrenchment of the public sector and �scal consolidation in ROLE OF GOVERNMENT BANK some countries.9 LENDING IN MITIGATING Despite the fact that the activity and ECONOMIC CYCLES AND CRISES importance of state-owned banks declined sharply over time, they still play a substantive The recent global �nancial crisis has brought role in some countries. For instance, state- to the fore the potential role of state-owned owned banks still dominate the process of banks in stabilizing aggregate credit.11 His- � nancial intermediation in Algeria, Belarus, torical evidence shows that the failure of China, the Arab Republic of Egypt, India, credit flows to recover may drag down the and the Syrian Arab Republic, where the asset recovery of real economic activity (Biggs, market share of these banks exceeded half the Mayer, and Pick 2010). In other words, a pre- assets of the banking system in 2010. In other condition for the recovery of the corporate countries, state-owned banks do not lead the sector may be the recovery of the � nancial process of credit creation but still play an sector.12 Though a strand of the empirical lit- important role. For example, state-owned erature argues that recoveries in real output banks in Argentina, Brazil, Indonesia, Korea, without a recovery in credit flows—credit- Poland, the Russian Federation, and Turkey less recoveries or “Phoenix miracles�—are GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 105 not rare events, and become more frequent banks. Governments injected capital into after a banking crisis or credit boom (Abiad, their state-owned banks to roll over exist- Dell’Ariccia, and Li 2011; Bijsterbosch and ing loans or provided new credit to SMEs or Dahlhaus 2011; Calvo, Izquierdo, and Talvi exporting � rms (this has been the case, for 2006a, 2006b), the evidence at the macro- example, in Canada, Chile, Korea, and Tuni- economic level is far from conclusive and sia), raised credit ceilings of their state-owned displays a great deal of heterogeneity. Fur- banks (in Finland and Korea, among others), thermore, firm-level evidence (Ayyagari, and established credit facilities for banks Demirgüç-Kunt, and Maksimovic 2011a) (for example, in India and Tunisia). In Bra- shows that real sector recoveries are not cred- zil, a state-owned development bank (Banco itless, and that firms substitute short-term Nacional de Desenvolvimento Econômico credit with long-term external �nance, either e Social, BNDES) played an important role through long-term borrowing or capital issu- in expanding credit during the recent cri- ance. In sum, Phoenix miracles are not sup- sis. However, the fact that most loans seem ported by �rm-level data either in the United to have gone to large firms and that credit States or among emerging markets. continued to expand even after the economy The global recession in 2008–09 involved recovered in 2010–11 calls into question the unprecedented conditions: real output fell ability of this bank to behave countercycli- sharply and in a synchronized fashion across cally (box 4.1). State-owned development the world (with more than 80 percent of the banks in Mexico expanded their loan portfo- countries sharing a recessionary phase in the lios and also supported key markets through period from the third quarter of 2008 to the special guarantee programs (box 4.2). Non- �rst quarter of 2009). In addition, real credit performing loans (NPLs) and profitability also fell in a synchronized manner across seem stable so far, perhaps because most of countries. Aisen and Franken (2010) show the lending was through tier II operations in that 95 percent of countries experienced which credit decisions are left to private inter- a contraction in bank credit in at least one mediaries.13 However, because credit kept month between September 2008 and May growing even after the economy recovered, 2009. The dire financial conditions in the it is still unclear whether state-owned devel- world economy reignited the debate on the opment banks in Mexico will be able to play countercyclical role of government banks. an effective countercyclical role rather than The fact that recoveries may not take place crowd out private banks. without credit growth may be reason enough In Poland, state-owned bank PKO Bank for the state to use their government bank Polski has played a role in credit stabiliza- infrastructure in order to keep the credit flow- tion by lending to the corporate and SME ing. State-owned banks may facilitate credit sectors (box 4.3). Though so far the bank stabilization for several reasons. First, stabi- has remained pro�table and NPLs have not lization of credit markets may be a part of changed dramatically, time will tell whether state-owned banks’ mandate. In this context, the bank’s credit allocation was efficient. state-owned banks may �ll the gap of credit The case of Poland is not representative of caused by underprovision by their private all countries in Eastern Europe. Research counterparts in times of crisis (Rudolph 2009, comparing the behavior of state-owned and 2010). Second, state-owned banks could be foreign-owned banks in relation to private perceived as safer during recessions or crises, domestic banks in Eastern Europe and Latin and their more stable deposit base could allow America before and after the 2007–09 crises them to have more stable lending activity dur- (box 4.4) shows that in Eastern Europe, state- ing crises (Micco and Panizza 2006). owned banks in general did not play a coun- During the recent crisis, many countries tercyclical role. On the other hand, state- deployed their state-owned banks to counter- owned banks in Latin America increased act the credit contraction by privately owned loan growth during the crisis relative to their 106 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 4.1 Intervention Using State-Owned Banks in Brazil The Brazilian government actively used its state bank cent of GDP. The continued expansion of credit by infrastructure to engineer a rapid countercyclical public banks, and especially BNDES, in 2010 and response to mitigate the contagion effects from the 2011 illustrates the dif�culties of behaving counter- global � nancial crisis. New liquidity assistance lines cyclically during upswings in real economic activity. were created by the Central Bank of Brazil (BCB) to BNDES played an important role: it extended enable the country’s state commercial banks, Banco special credit facilities thanks to a generous capital do Brasil (BB) and Caixa Econômica Federal (CEF), injection by the government (R$100 billion in 2009). to acquire ownership interest in private and public Loans from BNDES were extended at a subsidized � nancial institutions—including insurance compa- rate—the long-term interest rate was set at 6 percent, nies, social welfare institutions, and capitalization which is 7.5 percentage points lower than the market companies—with or without the acquisition of the rate (Lazzarini and others 2011). In fact, credit by capital stock control. Moreover, the government used BNDES surged from R$160 billion (at 2005 prices) in its public bank infrastructure—BB, CEF, and Banco September 2008 to R$277 billion in December 2010. Nacional de Desenvolvimento Econômico e Social The fact that most of this credit is being extended to (BNDES), the state-owned development bank—to large � rms that are likely to get loans elsewhere (Laz- play a countercyclical role in credit markets during zarini and others 2011) indicates that private banks this period of global � nancial turmoil. Figure B4.1.1 are crowded out in credit markets. illustrates the expansion of credit by Brazilian state- The portfolio expansion of BNDES was accompa- owned banks from 13.4 percent of GDP in Septem- nied by a shift in the composition of its funding. Fig- ber 2008 to 18.1 percent of GDP by end-2009. As ure B4.1.2 shows that resources channeled from the of December 2011, credit extended by public banks national treasury rather than those coming from the represented 21.4 percent of GDP, while overall credit Workers’ Assistance Fund became the main source of operations in the � nancial system totaled 49.1 per- funding beginning in 2009. The transfer from Trea- FIGURE B4.1.1 Ownership and Credit in Brazil Credit as percentage of GDP 60 50 40 30 20 10 0 20 ary 20 ary 20 ary 20 ary 20 ary 20 ary 20 ary 20 ary 20 ary 20 ary uly uly uly uly uly uly uly uly uly uly nu u u u u u u u u u -J -J -J -J -J -J -J -J -J -J an an an an an an an an an Ja 02 03 04 05 06 07 08 09 10 11 -J -J -J -J -J -J -J -J -J 2- 03 04 05 06 07 08 09 10 11 0 20 20 20 20 20 20 20 20 20 20 Government-owned banks Private domestic banks Foreign-owned banks Source: Central Bank of Brazil. Note: Government-owned banks refer to institutions in which federal, state, or municipal governments hold more than 50 percent of the voting capital. Private domestic banks include those institutions in which individuals or corporate entities domiciled and resident in the country hold more than 50 percent of the voting capital. Foreign-owned banks include those institutions that have—under external control and either directly or indirectly—the majority of voting capital. They also include those banks established and headquartered abroad with agencies or branches in the country. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 107 BOX 4.1 Intervention Using State-Owned Banks in Brazil (continued) FIGURE B4.1.2 BNDES: Sources of Funding consequences. Figure B4.1.3 breaks down BNDES dis- bursements by � rm size: 85 percent of disbursements R$ billions by BNDES went to large � rms in 2009 (R$115.9 bil- 500 lion)—the worst year of the crisis. In the midst of a 450 strong postcrisis recovery, BNDES continued lending 400 at the same pace of the crisis years (R$139.7 billion in 350 2011 from R$136.4 billion in 2009). Large � rms still 300 represented a big share of BNDES disbursements in 2011 (64 percent); however, the participation of micro 250 and small �rms as well as individuals has increased to 200 26 percent (R$36.3 billion). 150 Recent research also shows that loan activities 100 deployed by BNDES are consistent with the politi- 50 cal view of public lending: an increasing amount of funds are typically channeled to � rms in regions 0 where allied incumbents are facing political competi- 2007– 2008– 2009– 2010– 2011– December December December December September tion (Carvalho 2010), or to � rms that donate to can- FAT Bonds Multilateral National didates that won an election (Claessens, Feyen, and institutions treasury Laeven 2008; Lazzarini and others 2011). Economi- Source: Central Bank of Brazil. cally, employment expands among (manufacturing) Note: FAT = Workers’ Assistance Fund. firms that are eligible to BNDES loans in regions sury through bond issuances may crowd out private where allied incumbents face political competition credit, keep interest rates at high levels, and thus (Carvalho 2010). However, pro�tability, market val- reduce the provision of credit in the overall economy. uation and investment appear to remain unchanged The transfers also involve a �scal cost that reduces in � rms that received funding from BNDES—either the scope of direct public investment. through loans or equity provision. Only their � nan- As pointed out in the report, state-owned bank cial expenses were reduced considerably (Lazzarini interventions in credit markets also carry economic and others 2011). FIGURE B4.1.3 Distribution of BNDES Disbursements by Company Size (percentage) a. 2009 b. 2011 9% 6% 26% 64% 10% 85% Large Medium Micro/small and individual Source: BNDES Management Report, various years. 108 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 4.2 The Recent Global Crisis and Government Bank Lending in Mexico Mexican development banks (DBs) rapidly expanded cover about 200 percent of NPLs. Pro�tability has operations during the critical moments of the global remained stable or even improved in some cases. Pre- � nancial crisis by providing short-term credit to well- liminary estimates of the �scal costs (in terms of cap- established private sector � rms and nonbank � nan- ital allocations to DBs) appear modest because banks cial institutions with problems, both to refinance were adequately capitalized before the crisis and did their debt and to mitigate the sharp deceleration in not need additional funds to increase their lending private bank lending (�gure B4.2.1).a Credit to the portfolios. Moreover, the bulk of total interven- private sector (instead of public sector) expanded tions took place through credit guarantees and tier fastest, and overall credit stabilized or began declin- II operations, in which credit decisions are left to pri- ing after fourth quarter 2009. Despite the DBs’ vate intermediaries. Tier I lending expanded, but to strong lending expansion, the portfolio share of DBs a large extent (especially loans by Bancomext) those relative to that of private banks remained below the loans carried high rates and were highly collateral- 2006 level, as its share in total lending had been con- ized, providing � rms with incentives for early repay- tracting prior to the crisis. ment as market conditions stabilized (which many Mexican DBs were credited with restoring oper- � rms have done). However, it is still early to assess ations in key markets such as commercial paper the full impact on the quality of intermediation. through special guarantee programs—with most of Overall, Mexican DBs played an important coun- these programs being temporary. Overall, DBs have tercyclical role during the downturn. However, the extended Mex$71 billion (approximately US$6.1 bil- portfolio of credit and, especially, guarantees has lion) in guarantees as of June 2011—about 20 per- decreased very slowly during the upturn. Expanded cent of the total loan portfolio balance—and each operations have so far not compromised the �nancial guaranteed peso induced 2.8 pesos in credit (�gure position of DBs. Going forward, it may be useful to B4.2.2). establish mechanisms to facilitate the expansion and Nonperforming loans have remained at reason- contraction of the DBs’ balance sheet—and, in par- able levels—1.1 percent compared to 3.0 percent for ticular, of capital—to mitigate incentives to compete private banks as of November 2011—and provisions with the private sector during the upswing. FIGURE B4.2.1 Gross Loan Portfolio Growth FIGURE B4.2.2 Partial Credit Guarantees Year-on-year loan growth Mex$, billions 40 90 80 30 70 20 60 34.8 27.4 26.4 27.1 25.3 17.0 50 18.4 10 4.8 5.0 6.1 6.0 40 6.5 5.8 10.8 2.9 10.5 10.2 11.2 10.1 0 30 24.0 25.6 20 30.3 28.2 29.2 28.3 –10 26.7 10 9.9 10.7 0 –20 r r r r e be e e ne e e mb mb mb 20 q4 20 4 20 q4 20 q2 20 q4 20 2 20 q4 20 2 20 -q4 -O e r un un be em -q -q -q n u - - - - - 20 1-Ju -J -J -J te ce ce 05 06 07 08 08 09 09 10 10 cto ec 09 10 11 ep e e 20 -D -D -D 1 20 20 20 -S 08 09 10 11 08 20 20 20 20 Gross portfolio: developing banks Nafin Bancomext Gross portfolio: private banks Banobras Sociedad Hipotecaria Federal Source: Calculations based on data from the Mexican Banking and Securities Commission. a. The Mexican �nancial system has six development banks along with several trust funds and development institutions that provide credit. DBs provide tier I and tier II �nancing and guarantees and undertake loan portfolio sales and special development programs besides administrative functions, such as technical assis- tance and training. Banobras and Na�n account for about a third of total development bank assets each, while Banobras alone accounts for 44 percent of total development banks loans. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 109 BOX 4.3 State Commercial Banks in Action during the Crisis: The Case of Poland Poland was the only economy in the European Union major drivers. First, PKO BP’s management made an that avoided a recession in 2009 thanks to timely �s- arm’s-length commercial decision to capitalize on the cal stimulus and growth in lending, which were sup- decreased participation of foreign-owned banks in ported by policy measures undertaken by the Polish loan markets and to create attractive business oppor- Financial Supervisory Authority and the National tunities for households and corporations. Second, Bank of Poland (NBP), as well as by an expansion of the government supported PKO BP’s lending activi- the loan portfolio by the country’s state-owned com- ties during the crisis through public announcements mercial bank, PKO Bank Polski (PKO BP). by its supervisory board. Third, PKO BP’s conserva- PKO BP expanded credit at a faster pace during tive funding structure reduced the bank’s dependence the crisis than Polish subsidiaries of foreign banks, on wholesale �nancing (domestically and abroad), as which control almost three-fourths of the total assets opposed to foreign banks that relied partly on external of the system. Despite high credit spreads, credit by funding from parent banks. Finally, PKO BP’s capital PKO BP grew by more than 1 percent of GDP per adequacy ratio—above the regulatory minimum of 8 year during 2009–11. PKO BP’s share in total lend- percent—was further raised following a 5 billion Pol- ing increased from 15.6 percent of GDP in 2008 ish zlotys rights and new share issuance in late 2009. Q3 to 17.2 percent in 2010 Q4 (figure B4.3.1). In So far, PKO BP’s nonperforming loans have turn, its loan portfolio increased for all market seg- increased at a roughly similar pace to that of the ments, including, most important, the corporate and average of the largest foreign-owned banks, despite SME sectors—in which the value of new loans that the substantial increase in lending during the crisis were extended to those sectors throughout the crisis (�gure B4.3.2). Credit overdue by more than 90 days totaled 0.5 percent of nominal GDP in 2011. amounted to 4.2 percent for PKO BP (as opposed Rapid credit expansion by PKO BP relative to to the 5.8 percent market average). Moreover, PKO that of foreign-owned banks is attributed to four BP remained profitable during the crisis period, FIGURE B4.3.1 PKO BP’s Loan Share, 2008–11 FIGURE B4.3.2 Nonperforming Loans for PKO BP, 2008–11 Loans as % of GDP 17.5 Percent 10 17.0 9 8 16.5 7 16.0 6 5 15.5 4 15.0 3 08 08 09 09 09 09 10 10 10 10 11 11 11 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 14.5 08 08 09 09 09 09 Q2 0 10 10 10 11 11 11 1 PKO BP Largest foreign banks Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q3 Q4 Q1 Q2 Q3 Banking sector Sources: PKO BP, Polish Financial Supervisory Authority, and publicly available data on large foreign banks. as reflected by its high return on equity. However, ies. Its relative success in propping up credit may these indicators are largely backward looking, and be tied to the fact that it is a commercially oriented more time will be needed to assess the full impact of bank, open to free-market competition, and with increased lending on loan quality. conservative lending and funding policies. Stock Overall, PKO BP played a countercyclical role market–induced transparency, market discipline, for- during the crisis, partly offsetting the decline in the midable budget constraints, and professional man- available credit from foreign-owned bank subsidiar- agement have also played a role. Source: Piatkowski 2012. 110 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 4.4 Bank Ownership and Credit Growth during the 2008–09 Crisis: Evidence from Eastern Europe and Latin America Using bank-level data from 2004 to 2009, Cull and not contract their credit flow at a faster pace than Martínez Pería (2012) examined the impact of bank their domestic counterparts. On the other hand, the ownership on credit growth in developing economies behavior of state-owned banks in Eastern Europe before and during the crisis (table B4.4.1). They ana- (in terms of credit growth) did not differ from that lyzed the growth of banks’ overall loan portfolios, of private domestic banks. In general, state-owned as well as changes in corporate, consumer, and resi- banks in Eastern Europe did not mitigate the impact dential mortgage loans. They compared the determi- of the crisis on aggregate credit in the economy. The nants of credit growth for banks in eight countries in opposite is true in Latin America. Lending by state- Eastern Europe (Bulgaria, Croatia, the Czech Repub- owned banks grew during the crisis at a faster pace lic, Hungary, Poland, Romania, the Slovak Repub- than domestic and foreign banks (�gure B4.4.1). lic, and Slovenia) and six countries in Latin America Complementary evidence by de Haas and others (Argentina, Brazil, Chile, Colombia, Mexico, and (2012) from 1,294 banks for 30 Eastern European Peru). They found that the decline of credit growth countries over the period 1999–2009 shows that for- among foreign banks in Eastern Europe was larger eign bank credit grew at a faster pace than domes- than that of their domestic private counterparts tic (public and private) banks before the crisis, and during the crisis. In Latin America, foreign banks it sharply decelerated in 2008. On the other hand, did not fuel loan growth prior to the crisis and did credit extended by both state and private domestic TABLE B4.4.1 Determinants of the Growth of Total Gross Loans Variables Latin America Eastern Europe Foreign banks (Fgn) –11.098*** –10.403*** –9.320*** 2.651 3.83 4.04 [–6.293] [–6.014] [–4.705] [0.558] [0.986] [0.956] Government banks (Govt) –8.66 –8.576 –9.833 –1.338 0.174 1.207 [–1.779] [–1.694] [–1.765] [–0.248] [0.030] [0.203] Crisis2008 (dummy) –39.508** –22.540** [–3.102] [–3.360] Crisis2009 (dummy) –11.75 –19.179* [–0.867] [–2.063] Fgn ‫ ן‬Crisis2008 11.431* 11.375* 8.793 2.499 1.909 1.83 [2.267] [2.091] [1.450] [0.574] [0.384] [0.377] Fgn ‫ ן‬Crisis2009 –7.336 –10.269 –10.006 –14.394** –13.504* –15.562** [–0.508] [–0.754] [–0.994] [–3.048] [–2.120] [–2.648] Govt ‫ ן‬Crisis2008 27.569*** 27.648*** 27.926*** 4.677 2.303 0.318 [8.882] [5.441] [4.778] [0.605] [0.302] [0.057] Govt ‫ ן‬Crisis2009 14.954 14.831* 20.421 –1.249 –3.244 –4.989 [1.411] [2.222] [1.945] [–0.153] [–0.386] [–0.599] Constant 53.932*** 48.121*** 70.439*** 47.285*** 17.877** 20.015 [13.125] [11.527] [15.072] [7.362] [2.372] [1.606] Country–time interactions No Yes Yes No Yes Yes Bank characteristics Yes Yes Yes Yes Yes Yes Bank characteristics interacted with crisis No No Yes No No Yes Observations 878 878 878 770 770 770 R-squared 0.17 0.311 0.326 0.21 0.54 0.544 Number of countries 6 6 6 8 8 8 Source: Cull and Martínez Pería 2012. Note: The dependent variabloe is the annual percentage in total gross loans. t-statistics are in brackets. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 111 BOX 4.4 Bank Ownership and Credit Growth during the 2008–09 Crisis: Evidence from Eastern Europe and Latin America (continued) banks declined in 2009. However, the behavior of banks in some Eastern European countries may have state banks was less procyclical than that of (domes- partly mitigated the contraction of credit when pri- tic and foreign) private banks. Hence, state-owned vate banks began to deleverage. FIGURE B4.4.1 Growth of Gross Loans and Bank Ownership in Latin America and Eastern Europe, 2004–2009 Percentage change 60 50 40 30 20 10 0 Foreign banks— Private Government banks— Foreign banks— Private Government banks— Eastern Europe domestic banks— Eastern Europe Latin America domestic banks— Latin America Eastern Europe Latin America Precrisis Crisis Source: Calculations based on Cull and Martínez Pería 2012. private counterparts (box 4.4 and Cull and (2006) � nd that state-owned bank lending Martínez Pería 2012). is less procyclical than lending by private In sum, the experiences mentioned above domestic banks. However, one important regarding the role of state-owned banks dur- caveat is that their paper does not analyze the ing the recent crisis show that they played a general equilibrium effect of the smoothing countercyclical role in some instances but not activity of state-owned banks. State-owned others. Furthermore, although it is too early bank lending may merely crowd out lending to assess the quality of intermediation based from private banks; hence, the presence of on information on NPLs, there is evidence, state-owned banks would not affect aggregate in some cases, of (a) dif�culty unwinding the lending during the business cycle. Comparing expanded portfolios of development banks; the lending behavior of Western European (b) politically motivated lending; and (c) neg- state-owned and private banks during 2000– ligible effects on revenues, investment, and 09, Iannotta, Nocera, and Sironi (2011) �nd employment for eligible �rms. no difference in the behavior of both types of The evidence is also mixed regarding the banks across the business cycle. Nonetheless, cyclicality of lending by government banks unlike private banks, loan growth of state- in previous crises and economic cycles. owned banks is signi�cantly more sensitive to Using bank-level data for 119 countries over the election cycle, thus indicating the politi- the 1995–2002 period, Micco and Panizza cal role of the government as a bank’s owner. 112 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 In a recent paper, Bertay, Demirgüç-Kunt, capital is scarce, and public distrust is sig- and Huizinga (2012) use an approach simi- ni�cant, government banks can play a signi�- lar to Micco and Panizza (2006) and Ian- cant role in jump-starting �nancial and eco- notta, Nocera, and Sironi (2011), but unlike nomic development.15 According to this view, these studies they control for the endogene- governments have adequate information ity of GDP growth to credit growth by using and incentives to promote socially desirable the system-generalized method of moments projects, and ownership of banks enables estimation. Using a large worldwide sample the government both to collect savings and of banks for the recent period from 1999 to to direct them toward strategic longer-term 2010, they also show that state bank lend- projects. Through such project � nance, the ing is less procyclical than lending by private government overcomes institutional failures banks, especially in countries with good gov- that are undermining private capital markets ernance. Furthermore, they show that lend- and generates aggregate demand and other ing by state banks in high-income countries externalities that foster growth (Armendáriz is even countercyclical. Finally, using mac- de Aghion 1999; Bruck 1998). rolevel data on bank credit and government In stark contrast, the political view bank participation, Calderón (2012) shows argues that governments do not have suf- that cycles in real credit per capita are more �cient incentives to ensure socially desirable volatile in countries with large participation investments and acquire control of banks to of state-owned banks (as opposed to those provide employment, subsidies, and other with low participation) and that the recov- bene�ts to supporters, who return the favor ery of credit—although stronger because of in the form of votes, political contributions, a larger rebound effect—is slower (box 4.5). and bribes (Kornai 1979; Shleifer and Vishny 1998).16 In this view, government ownership politicizes resource allocation, softens budget LONG-RUN IMPACT OF constraints, and hinders economic ef�ciency. GOVERNMENT BANK The development and political views of OWNERSHIP government ownership make contrasting The countercyclical role of state banks can- predictions about the impact of government not be evaluated independently of their ownership on �nancial development. Accord- long-term performance in credit allocation. ing to the development view, government In theory, economists hold different views ownership should result in deeper, more ef�- about the merits of government bank own- cient, inclusive, and stable credit markets. On ership for long-run financial development the other hand, the political view predicts and economic growth.14 Box 4.6 summarizes that government ownership of banks will competing views on the role of state-owned result in inef�cient, noninclusive, underdevel- banks in promoting financial development oped, and unstable credit markets. While the and access. development view predicts that government Advocates of state presence argue that gov- ownership will enhance economic growth ernment ownership in banking is justi�ed by through its positive impact on � nancial sec- market failures and development goals. They tor development, the political view maintains point out that � nancial markets are differ- that government ownership could be perni- ent from other markets and that government cious for the real economy because it leads to intervention can remedy market failures (such credit misallocation. as the underprovision of information or the Theoretical arguments notwithstanding, undersupply of capital in the case of projects the bulk of the empirical evidence suggests that have externalities). The development that government bank ownership in develop- view, associated with Alexander Gerschen- ing economies has had negative consequences kron (1962), argues that in countries where for long-run � nancial and economic devel- economic institutions are not developed, opment. A large number of cross-country GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 113 BOX 4.5 Macroeconomic Evidence on the Impact of Government Banks on Credit and Output Cycles Calderón (2012) studies the impact of government 9-quarter period in countries with high GOB share, bank participation on credit and output cyclicality while it declined at 2.9 percent per quarter over an using a country-level quarterly database of 66 coun- 11-quarter period in those with low GOB share. tries over 1980–2010. Aggregate credit cycles are Credit recoveries following the banking crises were deeper in � nancial systems of countries with high shorter: it took approximately 5 quarters for the participation of government-owned banks (GOBs), trough in real credit to reach the previous peak in especially during times of banking crises. In fact, developing economies with high GOB share com- credit contractions are almost twice as deep in coun- pared with the 9 quarters that it took a country with tries with high GOB participation when compared low GOB share. Also, credit recoveries after crises with those with a low participation—14.3 versus 7.7 occurred at a slightly faster pace in high-GOB-share percent, respectively (table B4.5.1). Credit recoveries developing economies (3.2 percent per quarter). are faster and have a stronger rebound effect in coun- When looking at upswings in real economic tries with high GOB participation (8.4 percent) than activity, credit appears to behave procyclically. Fig- those with low participation (5.1 percent). ure B4.5.1 shows that credit grows in tandem with During � nancial crises, credit contractions tend real GDP as real economic activity starts its recov- to be shorter but deeper in developing economies ery (period T), and the recovery in real output and with high GOB share relative to those with low credit in countries with high GOB share appears to share. Calderón (2012) shows that credit dropped at be stronger because of a larger rebound effect. How- an annualized rate of 6.4 percent per quarter over a ever, credit growth is above trend after 10 quarters of TABLE B4.5.1 Credit Cycles and Government Ownership of Banks Main feature of credit contractions and recoveries All episodes Episodes coinciding with crisis Episodes with no crisis GOB Region Sharea Duration Amplitude Slope Duration Amplitude Slope Duration Amplitude Slope 1. Credit contractions All economies Low 6.7 –7.7% –1.7% 9.3 –16.2% –2.0% 6.4 –6.9% –1.6% High 6.8 –14.3% –2.3% 9.2 –37.4% –5.3% 6.3 –11.8% –2.1% Developed economies Low 6.7 –4.9% –1.1% 18.0 –25.1% –1.4% 6.4 –4.6% –1.1% High 6.2 –6.7% –1.6% 7.3 –6.7% –1.7% 6.0 –6.6% –1.3% Developing economies Low 7.0 –11.3% –2.0% 11.1 –22.8% –2.9% 6.6 –10.0% –2.0% High 6.7 –15.4% –2.8% 8.6 –45.0% –6.4% 6.0 –13.6% –2.5% 2. Credit recoveries All countries Low 4.9 5.1% 1.2% 8.8 9.2% 1.3% 4.6 4.9% 1.2% High 4.5 8.4% 2.0% 5.8 9.4% 2.8% 4.3 8.0% 1.8% Developed economies Low 5.3 4.4% 0.9% 30.0 4.5% 0.8% 4.9 4.4% 0.9% High 4.6 4.4% 0.6% 6.6 1.8% 0.4% 4.3 4.8% 0.6% Developing economies Low 4.7 5.7% 1.5% 9.2 19.5% 2.9% 4.4 5.5% 1.5% High 4.5 9.6% 2.5% 4.9 9.6% 3.2% 4.4 9.3% 2.3% Source: Calderón 2012. Note: The duration of contractions is de�ned as the period (in quarters) between the peak in real credit per capita and its subsequent trough. Recoveries, on the other hand, are de�ned as the early stages of expansion. Duration of contractions elapses the time that the corresponding variable goes from its trough to the previous peak level. The table reports the average duration of the different cyclical phases (downturns and upturns) for real credit per capita. The statistics for amplitude and slope refer to sample median across episodes. The amplitude of the downturn is the distance between the peak in real output and its subsequent trough, and that of the upturn is computed as the four-quarter cumulative variation in real output following the trough. The slope of the downturn is the ratio of the peak-to-trough (trough-to-peak) phase of the cycle to its duration. These calculations were made using a quarterly database of 66 countries over the period 1980–2010 (21 developed economies and 45 developing economies). a. The �nancial system of a country with high (low) participation of state-owned banks is de�ned as the system with an asset share of government-owned banks that is above the median sample. (Continued on next page) 114 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 4.5 Macroeconomic Evidence on the Impact of Government Banks on Credit and Output Cycles (continued) the start of the recovery in countries with low GOB with low GOB share when comparing crisis-related share, while it is still converging to trend growth in upturns with other upturns. It can be inferred from high-GOB-share countries.a this � nding that countries may not be equipped with In general, credit starts to recover in the period government infrastructure or may � nd ways other following the start of the upswing in real GDP— than government banking to prop up � nancial con- period T + 1—regardless of the magnitude of the ditions. For instance, they may choose to expand shock that led the economy to a recession. Figure the balance sheet of their central bank. In countries B4.5.2 shows that the dynamics of credit around with high GOB share, real credit per capita fluctuates real economic recoveries is similar among countries more intensely in recoveries that follow crisis-related FIGURE B4.5.1 Evolution of Real GDP and Credit around Recoveries in Economic Activity a. Countries with low participation of GOBs Deviations from trend 0.05 0.00 –0.05 –0.10 –0.15 –0.20 T-12 T-11 T-10 T-9 T-8 T-7 T-6 T-5 T-4 T-3 T-2 T-1 T T+1 T+2 T+3 T+4 T+5 T+6 T+7 T+8 T+9 T+10 T+11 T+12 Real output Real credit per capita b. Countries with high participation of GOBs Deviations from trend 0.05 0.00 –0.05 –0.10 –0.15 –0.20 T-12 T-11 T-10 T-9 T-8 T-7 T-6 T-5 T-4 T-3 T-2 T-1 T T+1 T+2 T+3 T+4 T+5 T+6 T+7 T+8 T+9 T+10 T+11 T+12 Source: Calderón 2012. Note: T represents the trough in real GDP (start of the recovery period). GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 115 BOX 4.5 Macroeconomic Evidence on the Impact of Government Banks on Credit and Output Cycles (continued) recessions when compared with regular upswings in its own recovery. Also, the stronger surge in credit economic activity. Following a crisis-related reces- in countries with GOB share is related to a greater sion, credit takes more time to hit a trough and start rebound effect from the crisis. FIGURE B4.5.2 Evolution of Real GDP and Credit around Recoveries in Economic Activity a. Countries with low participation of GOBs Deviations from trend 0.15 0.10 0.05 0.00 –0.05 –0.10 –0.15 –0.20 –0.25 T-12 T-11 T-10 T-9 T-8 T-7 T-6 T-5 T-4 T-3 T-2 T-1 T T+1 T+2 T+3 T+4 T+5 T+6 T+7 T+8 T+9 T+10 T+11 T+12 No crisis Crisis b. Countries with high participation of GOBs Deviations from trend 0.15 0.10 0.05 0.00 –0.05 –0.10 –0.15 –0.20 –0.25 T-12 T-11 T-10 T-9 T-8 T-7 T-6 T-5 T-4 T-3 T-2 T-1 T T+1 T+2 T+3 T+4 T+5 T+6 T+7 T+8 T+9 T+10 T+11 T+12 No crisis Crisis Source: Calderón 2012. Note: T represents the trough in real GDP (start of the recovery period), and the dating of the banking crisis corresponds to that of Laeven and Valencia (2010). a. The analysis conducted in �gure B4.5.1 and B4.5.2 follows Beck, Levine, and Levkov (2010). It consists of regressing the (year-on-year) growth in real credit on a 25-quarter window (12 quarters before and after the peak or trough in real GDP) centered on a dummy variable at time T that takes the value of 1 when there is a trough in real GDP at quarter T (that is, the starting period of a recovery). The remaining variables are leads and lags of the trough taking place at time T. These regressions also include dummy variables to control for other country-speci�c characteristics. The plot of the time evolution of credit along this window shows deviations of credit relative to the average credit growth in tranquil times. 116 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 4.6 Two Views on the Role of State-Owned Banks In a recent debate published in the World Bank’s All by the Chilean Government. It is the country’s third About Finance blog, two academics expressed con- largest lender and operates in all major segments of trasting views about the role of state-owned banks in the banking market. The fact that it has to compete promoting �nancial stability and access. with private banks ensures it is well run. Banco del Franklin Allen, Nippon Life Professor of Finance Estado also has a long history of promoting access in and Economics, Wharton School, University of Penn- all parts of the country and to all people. Many other sylvania, argues that, despite being outperformed countries might bene�t from this type of bank.� by their private counterparts in terms of long-term In contrast, Charles Calomiris, Henry Kaufman resource allocation, “public banks may enjoy an Professor of Financial Institutions at Columbia Uni- advantage over private banks in times of crisis and, versity, argues that academic work “indicates power- hence, their merits need to be reassessed.� He goes fully the negative effects of state-controlled banks on on to say that “the real advantage would come when the banking systems of the countries in which they there is a crisis. Rather than having central banks operate and that the winding down of state-con- intervene in commercial credit markets, where they trolled banks was rightly celebrated in many coun- have little expertise, the state-owned commercial tries in the 1990s as creating new potential for eco- bank can temporarily expand its role both in terms nomic growth and political reform.� He goes on to of assets and loans. This should considerably improve say that there are three main reasons that explain the the functioning of the economy and overcome credit dismal performance of state-controlled banks. “First, crunch problems.� government of�cials do not face incentives that are “The � nancial system can be safeguarded during conducive to operating well-functioning banks. They times of crisis through a mixed system with mostly are typically not incentivized to maximize economic private banks but one or two are state-owned com- effectiveness and they tend not to be trained in credit mercial banks. They would compete with private analysis as well as private bankers. They face incen- banks in normal times to ensure a competitive cost tives that reward politically rather than economi- structure and prevent corruption, and they would cally motivated allocations of credit. Second, the provide useful information to regulators by signaling politically motivated allocation of funds to crony excess risk-taking or exercise of monopoly power by capitalists has adverse consequences for the politi- private banks. However, their real advantage would cal and social system of a region or country. State- become evident during a � nancial crisis. State-owned controlled banks are a breeding ground for corrup- commercial banks would be a safe haven for retail tion of elected and appointed government of�cials, and interbank deposits, act as a � re break in the pro- the � nancial regulatory authorities, and the courts. cess of contagion, and provide loans to businesses— Not only do they stunt the growth of the economy, particularly small and medium size enterprises— they also weaken the core political and bureaucratic through the crisis. They could expand and take institutions on which democracy and adherence up the slack in the banking business left by private to the rule of law depend. Third, state-controlled banks. Listing such banks will ensure full informa- banks are ‘loss-making machines.’ Because they tion on them is available and their stock prices will are not geared toward pro�tability or the aggressive indicate how well they are performing.� enforcement of loan repayment, but rather toward “Public banks can play another important role rewarding political cronies with funding, the losses in increasing access to � nancial services. If the gov- of state-controlled banks pose a major �scal cost for ernment wishes it to pursue this agenda then it may governments. Those �scal costs crowd out desirable be helpful to subsidize this kind of activity. In many government initiatives, and given the large size of the European countries in the nineteenth and twentieth losses, can be a threat to the solvency of government centuries, the post of�ce provided access to savings and a source of inflationary de�cit � nancing.� accounts and other kinds of � nancial services that “The crisis has only recon�rmed the extreme dam- many customers would not otherwise have had. A age that politically motivated lending can inflict. The good example of a public bank that plays these roles quasi-state-controlled U.S. entities Fannie Mae and is Chile’s Banco del Estado, which is entirely owned Freddie Mac accounted for more than half of the GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 117 BOX 4.6 Two Views on the Role of State-Owned Banks (continued) funding of subprime and Alt-A mortgages leading up “The huge crisis-related losses of equity capital in to the crisis. There is evidence that political motiva- the banking system and the subsequent stepping up tions drove the intentional risk taking and deteriora- of regulatory oversight over banks have resulted in tion of underwriting standards at those institutions a short-term contraction in the supply of credit. This after 2003—a crucial ingredient in the subprime credit crunch magni�ed the decline of GDP during boom of 2004–2007. Government quotas dramati- the recession, and slowed the pace of the recovery. In cally increased the funding that Fannie and Freddie such an environment, it may seem appealing to pass had to supply to low-income and underserved borrow- a law creating a state-owned bank with the goal of ers, but the supply of creditworthy low-income and re-starting the rapid flow of loanable funds. But such underserved borrowers was limited. Inevitably, lend- an initiative would be short-sighted. Rather than ing standards were relaxed. The U.S. experience is not promoting sustainable growth, it would slow growth unique. Political motivations drove Spanish cajas to over the medium or long run, as funds would be support a real-estate boom that ended in a massive channeled to low-productivity users. A move to sup- bust. In Germany, state-controlled banks also made port state-controlled banks would also raise systemic horrible investment decisions, thus reflecting incom- risk (as Fannie and Freddie, and the Spanish cajas petence more than corruption or political motives clearly show), promote corruption of government for channeling funds. Looking back historically, it is of�cials and institutions, and lead to �scal losses that clear that state-controlled lending has been a major could threaten the solvency of government and lead contributor to unwise and politically motivated risk to high inflation.� taking that has ended badly over and over again.� Source: All About Finance (blog), World Bank. studies show that greater government par- and Yañez 2007).18 State-owned banks also ticipation in bank ownership is associated tend to display lower z-scores, thus showing with lower levels of financial development greater instability than their private counter- (Barth, Caprio, and Levine 2001, 2004; La parts (Ianotta, Nocera, and Sironi 2007). Porta, López-de-Silanes, and Shleifer 2002), Interpreting cross-country evidence is dif- more politically motivated lending (Dinç �cult because of the potential for endogeneity 2005; Micco, Panizza, and Yañez 2007), biases resulting from reverse causation and lower banking-sector outreach (Beck, Demir- omitted factors.19 In other words, it is feasible güç-Kunt, and Martínez Pería 2007), wider that the negative association of government intermediation spreads and slower economic bank ownership with the different dimen- growth (La Porta, López-de-Silanes, and sions of �nancial development and with eco- Shleifer 2002), and greater financial insta- nomic growth could arise from the need for bility (Caprio and Martínez Pería 2002; La more government intervention in countries Porta, López-de-Silanes, and Shleifer 2002).17 with lower financial and economic devel- The evidence also suggests that state-owned opment or from some omitted factor. Simi- commercial banks operating in develop- larly, analyzing differences in the pro�tabil- ing economies have lower pro�tability than ity of state-owned versus private banks does comparable private banks, as well as lower not necessarily provide conclusive evidence interest margins, higher overhead costs, and regarding their impact and performance. a higher fraction of nonperforming loans State-owned banks may exhibit lower pro�t- (Berger, Hasan, and Zhou 2009; Farazi, ability because they maximize broader social Feyen, and Rocha 2011; Iannotta, Nocera, objectives, investing in � nancially unpro�t- and Sironi 2007; IDB 2005; Micco, Panizza, able projects with positive externalities. 118 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 However, detailed within-country stud- to target rents to voters by subsidizing mul- ies that are less susceptible to endogeneity tilocation firms to shift employment and concerns and are better able to identify the capital toward politically attractive regions impact of government ownership provide evi- and away from nonattractive regions. Politi- dence consistent with the bulk of the cross- cal manipulation of public bank lending does country studies. For instance, Cole (2009a) not seem to be restricted to developing econ- analyzes the expansion of government bank omies, as shown by Sapienza (2004), who ownership through nationalization in India presents evidence that state-owned banks and finds that, although areas with more serve as a mechanism to provide political nationalized banks showed a large increase patronage in Italy. in credit to rural borrowers, this increase did Not only are state-owned banks more not result in improved agricultural outcomes. likely to be engaged in political lending, Moreover, his results suggest that govern- but they also do not serve � rms that require ment bank ownership was associated with a more active participation of the state—that lower quality of �nancial intermediation and is, firms with deeper asymmetric informa- a misallocation of resources. Using detailed tion problems, such as SMEs. In a study of loan-level data from Pakistan, Khwaja and bank lending relationships in India, using Mian (2005) � nd that politically connected �rm-level and bank-level data during the late �rms were able to obtain larger and cheaper 1990s, Berger and others (2008) find that loans from state-owned banks (but not from state-owned banks have tighter relations with private ones) and defaulted on these loans state-owned enterprises than with SMEs. more frequently than nonconnected bor- Similarly, Ongena and Sendeniz-Yuncu rowers. They estimate that the economy- (2011) � nd that Turkish state banks fail to wide costs of the resulting misallocation of engage small opaque �rms. resources could be as high as 1.9 percent of State-owned banks can also exist for pop- GDP every year. Cole (2009b) also presents ulist reasons. Acharya (2011) demonstrates evidence of political manipulation of lending that the government can threaten the stability in India, with state-owned banks increasing of the � nancial system with the formulation agricultural lending substantially in tightly and implementation of �nancial policies that contested districts during election years. 20 focus on short-term populist goals. So-called The election-year increase in government short-termist governments may extend guar- lending is associated with higher default rates antees, weaken capital requirements, provide and does not have a measurable effect on direct lending, and encourage competition to agricultural output. generate greater entry into the � nancial sec- Carvalho (2010) provides support for the tor and expand economic activity. Excess view that politicians use government bank risk-taking arising from these policies may lending to influence the real behavior of �rms fuel credit booms and threaten the stability using plant-level data for the universe of Bra- of the financial system down the line. The zilian manufacturing � rms with at least 50 boom-bust cycle of U.S. housing prices in the employees. He shows that � rms eligible for run-up to the 2007–09 � nancial crisis is an government bank lending expand employ- illustration of this type of policies.21 ment in regions with allied incumbents near Not all state-owned banks are alike, reelection years. The effects represent persis- however. Depending on whether they aim tent and economically important increases in for profit maximization, take deposits, or the local employment of � rms and are asso- have a clear mandate, state-owned banks ciated with greater borrowing from govern- can be classi�ed as state commercial banks, ment banks. However, they are not associ- state development banks, and development ated with persistent expansions of the overall financial institutions (Scott 2007). State- employment and capital of �rms. The analy- owned commercial banks are institutions sis suggests that politicians use bank lending that behave similarly to private commercial GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 119 banks: they are deposit-taking institutions features right tends to be most challenging that may have a pro�t-maximizing objective. in weak institutional environments, where State development banks can take deposits the potential bene�ts of these institutions are and have a clear mandate, while development likely to be the greatest. � nancial institutions are state-owned � nan- Few success stories can be found for the cial institutions that have a speci�c mandate role of state-owned banks in promoting and are not funded through deposits. State- financial development when these institu- owned development banks and financial tions are engaged directly in the allocation institutions further distinguish between tier and pricing of credit. However, some exam- 1 and tier 2 institutions—those that lend ples of targeted government interventions directly to the public and those that lend to in the � nancial sector show positive results. private banks that in turn lend to the pub- Development banks in Mexico have played a lic. A large part of the evidence on the impact positive role in the provision of more complex of state-owned banks either focuses entirely (noncredit) � nancial services despite heavy on commercial banks or does not distinguish setup costs and uncertainty on financial between commercial and development banks. returns. Partnership of state-owned finan- De Luna and Vicente (2012) have recently cial intermediaries with their private coun- undertaken a global survey of development terparts has allowed them to overcome � rst- banks, which include state development mover disincentives, coordination failures, banks and development financial institu- and obstacles to risk sharing and distribution tions. They � nd large differences in the way (Demirgüç-Kunt, Beck, and Honohan 2008). these institutions are run and perform. The For instance, NAFIN (Nacional Financiera) authors show that although there are some developed an online platform called Cadenas bright spots, in general, most development Productivas to provide reverse factoring ser- banks and institutions need institutional vices to SMEs. This framework allows small reform (box 4.7). The literature identi- suppliers to use their accounts receivable �es a set of good practices to follow so that from large creditworthy �rms to get working state-owned banks institutions are well-run capital � nancing (Klapper 2006). BANSEFI (Gutiérrez and others 2011; Rudolph 2009; (Banca de Ahorro Nacional y Servicios Finan- Scott 2007). To avoid the problems of credit cieros) implemented an electronic platform to misallocation discussed earlier, state-owned help semiformal and informal �nancial insti- banks need a clear and sustainable mandate. tutions reduce their operating costs by pro- To ful� ll their mandate, state-owned banks viding centralized back-of�ce operations (for should target strategic sectors and work as example, clearinghouse services and liquidity complements to rather than substitutes for management, among others). Finally, FIRA private sector efforts to allocate resources. (Fideicomisos Instituidos en Relación con la They also require adequate risk management Agricultura) has brokered structured � nan- systems to guarantee � nancially sustainable cial products to align incentives and curtail business and to obtain funding from markets adverse selection problems between �nancial without explicit guarantees from the gov- intermediaries and firms in different parts ernment. Sound corporate governance plays of the supply chain of several industries. For a key role in explaining good performance example, FIRA has arranged collateralized of state-owned banks. It requires the trans- loan obligations for shrimp producers and parent nomination of board members and asset-backed securities for sugar mills (de la the selection of senior management by the Torre, Gozzi, and Schmukler 2007).22 These board. Drafting an ownership policy is rec- examples suggest that the government has ommended, with principles associated with a role, to be directly involved in the � nan- sound commercial practices, good corporate cial sector through short-run interventions governance, and competitive neutrality (Scott that address specific market failures and 2007). Unfortunately, getting these design that seek to complement private financial 120 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 4.7 Development Banks: What Do We Know? What Do We Need to Know? The world economy has witnessed a revival of inter- practices in governance and risk management. For est in development banks (DBs) in recent years. New instance, DBs remain vulnerable to undue political DBs have been formed in low- and middle-income interference and capture by interest groups because countries (for example, Bosnia and Herzegovina, about 75 percent of them do not have independent India, Malawi, Mexico, Mongolia, Mozambique, members on their boards. In addition, several DBs Serbia, and Thailand) as well as in some high-income have not adopted criteria with regard to the minimum countries (for example, the U.K.’s Green Investment quali�cations that their board members and senior Bank). DBs, along with other state-owned � nancial management should meet. There is concern about the institutions, account for one-quarter of banking real � nancial situation of DBs—about one-quarter of assets in developing economies—and this share is the surveyed institutions are not supervised with the even larger among Brazil, Russia, India, and China. same accounting and prudential standards applicable The renewed debate and interest in DBs has been to private commercial banks. fueled by their potential countercyclical role in times Financial performance of DBs is mixed. Although of crisis. Typically, DBs have fulfilled long-term 15 percent of DBs report nonperforming loans development roles by filling market gaps in long- exceeding 30 percent of their total loan portfolio, term credit and agriculture � nance and by promoting 78 percent of them admit the need to improve their access to � nance for SMEs. More recently, in light risk management framework. Going forward, DBs’ of the global � nancial crisis, DBs are now being per- major challenges include the need to reduce their reli- ceived by national authorities as an important part of ance on government budget transfers and to improve the policy toolkit to mitigate contractions in aggre- their own pro�tability: 59 percent of them indicate gate credit during crises. In fact, the loan portfolio that self-sustainability is a major challenge. Success of DBs grew by 36 percent globally during the period of DBs would also require major improvement in cor- 2007–09—that is, four times as fast as the growth porate governance and transparency, with one out of in credit provided by private commercial banks. two institutions surveyed responding that they are However, there is historical evidence that the bene�ts still far from having best practices in these areas. from DBs’ intervention in credit markets may come Successful stories of government bank owner- with potentially large costs, if these institutions have ship are not abundant. However, some DBs have unsound operating practices—that is, if they have proved more effective than others in achieving their neither appropriate risk management practices nor goals while ensuring financial sustainability. The sound corporate governance. more effective DBs have de� ned a clear and sustain- A global survey conducted by De Luna-Martínez able mandate and have adopted better corporate and Vicente (2012) aims at building new knowl- governance practices. A clear mandate (including a edge on activities, funding, business models, lend- target sector, positioning with regard to the private ing instruments, government arrangements, and sector and other DBs, and financial sustainability challenges faced by development banks. The survey, objectives) helps to focus the activity of the DBs and which de� nes development banks as � nancial insti- avoids the tendency to engage DBs in business where tutions with more than 30 percent of their shares the private sector has a comparative advantage. A owned by the state and with a public mandate, cov- clear mandate also complements the accountability ers 90 DBs from 61 countries—with their combined of the board of directors and management and facili- assets and their loan portfolio as of December 2009 tates performance monitoring. being approximately US$2 trillion and US$1.6 tril- Good practices in DB governance call for clear lion, respectively. de� nition of the roles of shareholders, the board of The global survey suggests that the performance directors, and management, as well as a separation of most DBs still needs substantial improvement. of their functions to avoid confl icts of interest. The Although one can identify some bright spots among shareholder representative should be clearly identi- development banks (for example, Mexico’s NAFIN �ed (for example, the minister of � nance), provide and Germany’s Kreditanstalt fuer Wiederaufbau broad policy guidelines, and appoint the board of [KfW]), many DBs still are far from adopting best directors, but it is advisable to avoid the presence of GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 121 BOX 4.7 Development Banks: What Do We Know? What Do We Need to Know? (continued) ministers on the board of DBs. The board should be have also adopted legislation that prevents them from professional and independent. The board provides a bailing out DBs in case of failure. In other jurisdic- strategic vision to ensure compliance with the policy tions, DBs lend only through tier II operations and objectives and establishes indicators to monitor per- share with the private sector the risk of lending to formance. It should also ensure that the �nancial sus- underserved segments of the market. Some DBs tainability of the institution is preserved and should are also governed by boards with only independent appoint and dismiss the chief executive officer. members. It is worth exploring all these innovations Finally, more effective DBs are supervised and regu- because they may be part of the solutions needed to lated as any other bank by the � nancial supervisory strengthen the weak institutions covered in the sur- authorities. vey, in particular those operating in dif�cult institu- The survey found that innovative procedures have tional environments. been put in place to help DBs operate effectively in Good practice recommendations are particularly various jurisdictions. For instance, some DBs are difficult to apply in countries with weak institu- legally obliged to achieve a minimum return on tional settings. Development banks could add most capital, measured in terms of the inflation rate or the value in � nancially underdeveloped countries where government’s cost of borrowing. Moreover, certain there are signi�cant market failures, but it is in those DBs have been partially privatized, and the man- environments in particular that the design can go agement has been transferred to the private sector bad. That is why, in practice, many DBs tend to under management contracts. Certain governments underperform. Source: De Luna-Martínez and Vicente 2012. intermediation, not replace it. Of course, can also diversify risk by guaranteeing loans once again the design and governance of such across different sectors or geographic areas. interventions are key factors. Credit guarantee schemes are designed to enable lenders to learn about the creditwor- thiness of constrained borrowers without GOVERNMENT INTERVENTION incurring the initial risks involved and to IN CREDIT MARKETS BEYOND allow these borrowers to establish a repay- BANK OWNERSHIP ment reputation and in time graduate to non- Credit guarantee schemes are a popular guarantee loans. Public guarantee schemes intervention tool during crises. For example, refer to those funded or managed with gov- during the 2007–08 crisis, many govern- ernment resources. ments (for example, Canada, Chile, Finland, Even before the crisis, credit guarantee France, Germany, Greece, Japan, Korea, schemes have become one of the most popu- Malaysia, the Netherlands, the United King- lar mechanisms of intervention to expand dom, and the United States) extended new the use of � nancial services for credit-con- and special schemes or refueled existing ones strained firms, such as SMEs. According to alleviate the impact of the credit crunch on to Green (2003), there are over 2,000 such SMEs. 23 Credit guarantee schemes typically schemes in place in almost 100 countries. As serve as risk transfer and risk diversi�cation discussed by Honohan (2010), credit guar- mechanisms. By replacing part or all the risk antee schemes can emerge for three main of the borrower with that of the issuer of the reasons: to mitigate asymmetric information guarantee (depending on the coverage ratio), problems between lenders and borrowers they tend to lower the lender’s risk. They that can result in credit rationing, as a means 122 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 of spreading and diversifying risk, and to income and quality of life of the borrowers exploit regulatory arbitrage if the guarantor and, in general, an increase in the amount is not subject to the same regulatory require- of commercial and economic activity in the ment as the lender. None of these three rea- country, as measured in terms of employment sons per se imply government involvement in and economic growth. guarantee schemes. Measuring the bene�ts of credit guaran- Government involvement in guarantee tees through financial and economic addi- schemes can be primarily justified on the tionality is far from easy, and most of the grounds of coordination failures among pri- existing evidence examines the ability of vate parties. 24 Coordination failure among credit guarantees to overcome problems in private parties and � rst mover disadvantage access to credit—that is, �nancial additional- can prevent private providers from entering ity. For instance, Riding, Madill, and Haines the market for credit guarantees or prevent (2007) found that 75 percent of guarantees lenders from pooling resources for such a generated additional loans thanks to the scheme. As Green (2003) argues, because implementation of a credit scoring methodol- banks cannot exclude the free riding of other ogy on loan applications to Canadian banks. � nancial institutions, they have little incen- In Chile, there is evidence that microenter- tive to produce information on constrained prises are more likely to get a bank loan using borrowers such as SMEs. A similar reluc- the State-Owned Guarantee Fund for Small tance applies to developing lending technolo- Entepreneurs, or FOGAPE (Larraín and gies suitable for such borrowers. Anginer, Quiroz 2006), and credit guarantees have de la Torre, and Ize (2011) argue that gov- helped expand lending to insurance-intensive ernment guarantees may have an edge over sectors (Cowan, Drexler, and Yañez 2009). private guarantees because the government There is also evidence of �nancial additional- is better at solving collective action fric- ity among G-7 countries. For example, �rms tions or coordinating failures among private that participated in the Italian credit guaran- parties—rather than solving agency fric- tee scheme during 1999–2004 bene�ted from tions (such as informational asymmetries or an increase in credit and a reduction in bor- adverse selection). rowing costs (Zecchini and Ventura 2006). Political factors also justify the existence In Japan, SMEs using the Special Credit of public guarantee schemes (for example, Guarantee Program for financial stability Beck and others 2010). Theory shows that received more credit than nonusers (Uesugi, guarantee funds are more effective and less Sakai, and Yamashiro 2010; Wilcox and Yas- costly in expanding access than directed uda 2008). Finally, guaranteed loans by the lending schemes (Arping, Lóránth, and Mor- U.S. Small Business Administration were less rison 2010). Also, guarantee schemes might affected by economic shocks than nonguar- be easier to justify politically because they anteed loans—thus providing some evidence resemble market-friendly instruments, in of countercyclicality of guarantee schemes which the lending decision typically stays (Hancock, Peek, and Wilcox 2007). with the lender, and because they imply small There is also some evidence—although initial costs of funding (losses accumulate rather scant—of economic additionality. For over time as defaults materialize). instance, the rate of employment was higher Rigorous evaluations of the total costs and in U.S. districts that received more guaran- bene�ts of these schemes are rare. Guaran- teed lending (Craig, Jackson, and Thompson tee schemes strive to attain financial and, 2007). In addition, sales performance and ultimately, economic additionality. Financial productivity improved among �rms �nanced additionality refers to greater provision of by KOTEC, a Korean government guaran- credit to credible clients for whom credit was tor that provides credit guarantees to new previously rationed. Economic additionality, technology-based enterprises (Kang and Hes- by comparison, refers to improvements in the hmati 2008; Roper 2009). GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 123 Credit guarantees can generate � nancial Honohan 2010). For instance, credit risk and economic bene�ts; however, because the assessment practices should be outsourced programs are generally targeted to speci�c to the private sector (rather than being con- sectors, they are unlikely to have large mac- ducted by the government) to improve the roeconomic effects. Also, they are not truly quality of risk decisions and to minimize loan countercyclical tools since they do not tend to losses. However, outsourcing risk manage- contract during periods of economic booms. ment in the case of public guarantee schemes At the same time, they can bring about siz- could potentially lead the lender to assign to able displacements and deadweight losses. the guarantee the worst eligible risks in the For instance, there is evidence that a large portfolio. This can be mitigated by penal- and growing share of guarantees granted by izing lenders that have high claims and by FOGAPE have been allocated to the same imposing higher future premium payments. � rms (Benavente, Galetovic, and Sanhueza Targeting for the credit guarantee scheme 2006). In addition, approximately half of should be broad (for example, speci�c sector the guaranteed loans in the Philippines went and areas) if the focus remains on credit-con- to borrowers with suf�cient collateral, thus strained groups, whereas too-speci�c target- generating signi�cant deadweight loss (Sal- ing may involve high bureaucratic costs that dana 2000). In Pakistan, half of the subsi- might distort lending decisions. The coverage dized credit for exporters went to �nancially ratios determined by the scheme should pro- unconstrained firms that did not need the vide incentives for lenders to properly assess funds, and the diversion in unneeded credit to and monitor borrowers. Most practitioners bene�ciary �rms could have held GDP below argue that lenders should retain a signi�cant its potential by 0.75 percent (Zia 2008). part of the risk—for example, from 30 to 40 There is also evidence of signi�cant costs percent. However, in practice, 40 percent of in credit guarantee schemes among devel- the 76 credit guarantee schemes analyzed in oped nations. The massive credit guarantee Beck and others (2010) offer guarantees of program implemented by the government up to 100 percent. The median coverage is of Japan between 1998 and 2001 rendered 80 percent, which is certainly not in line with only a temporary availability of funds for providing incentives for lenders to properly the intended program participants, and their assess and monitor borrowers. Guarantee ex post performance deteriorated relative programs with coverage ratios between 90 to nonusers of the guarantee. Also, major and 100 percent have been shown to bring banks often used the guarantee scheme to about large losses. For instance, the rural replace nonguarantee loans with guaran- credit guarantee fund established by the Lith- teed ones to minimize their exposure to uanian government, which offered 100 per- risky assets (Uesugi, Sakai, and Yamashiro cent coverage for loans aimed at �nancing the (2010). Another scheme set up by the Japa- purchase of tractors and other agricultural nese government in October 2008, the Emer- equipment, brought about a large number of gency Credit Guarantee Program, failed to bad loans within three years of its starting translate a greater availability of funds into date (Rute 2002). higher investment and employment among Regarding the pricing of credit guaran- user �rms and, in addition, deteriorated their tees, the fees charged by the scheme should creditworthiness (Ono, Uesugi, and Yasuda be high enough to ensure � nancial sustain- 2011). ability of the fund and low enough to secure The success of credit guarantees is tightly adequate participation by lenders and bor- linked to the design of the scheme. The lit- rowers. The payout of the guarantee should erature identi�es and discusses a set of good take place after the bank initiates legal action practices that contribute to the success- following default in order to reduce moral ful implementation of a credit guarantee hazard on the side of the lender, who might scheme (Beck and others 2010; Green 2003; be too quick to write off a loan after default. 124 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 However, there is a dearth of schemes that laws and enforcement mechanisms is prefer- structure the payout so as to maximize incen- able to government interventions in address- tives for lenders to minimize loan losses ing inadequacies of the legal framework asso- (Beck and others 2010). Finally, among risk ciated with the credit system (Holden 1997; management practices, guarantee schemes Vogel and Adams 1997). Although rigorous can reduce lenders’ own ex post exposure to impact evaluations of their costs and bene�ts loan defaults through reinsurance, loan sales, are still scarce, their performance hinges on or portfolio securitization. Risk diversi�ca- good design, which is more challenging to tion abilities are tied to the development of get right in weak institutional environments. local capital markets and �nancial products. Best-practice lessons suggest that the most Beck and others (2010) � nd that 76 percent successful credit guarantee schemes are those of guarantee schemes use risk management that move to broad eligibility and other cri- tools. This figure is encouraging since the teria, reduce subsidies, and make greater use authors also �nd that schemes that do not use of the portfolio and wholesaling approach in risk management tools exhibit higher inci- preference to case-by-case evaluation by the dence of default losses. guarantor of retail loans. Finally, recent research shows that 49 Finally, deploying credit through state- percent of the 76 credit-guaranteed schemes owned banks and extending credit guaran- covered in Beck and others (2010) are funded tees were not the only means used by gov- by governments. The government has a ernments to prop up financial conditions much more limited role in management, during the recent crisis. 25 Some countries risk assessment, and recovery: less than 20 implemented alternative strategies to offset percent of schemes are managed by govern- the credit crunch using their central banks. In ments, and credit risk assessment and recov- addition to traditional policies such as inter- ery are conducted by governments in only est rate cuts to support aggregate demand, approximately 10 percent of the schemes. developed economies tried to revive the ailing Government-backed guarantee schemes with financial markets by implementing uncon- responsibilities in credit risk and recovery ventional monetary policies that led to the are typically older, are more prone to guar- expansion and change in the composition antee loan portfolios, pay out after the bank of their central bank’s balance sheets. The initiates recovery, and lack a risk manage- wide array of measures implemented include ment program. These results are consistent the U.S. Federal Reserve Board’s purchase of with the notion that guarantee schemes with long-term Treasury bills, the European Cen- greater government involvement are less tral Bank’s purchase of covered bonds, relax- likely to manage risk and losses. Consistent ation of the collateral framework to access with these � ndings, Beck and others (2010) the discount window, changes in funding � nd that government involvement in credit terms or auction schedules, support of money decisions is associated with higher incidence markets, and foreign currency swaps.26 of default losses. Using the central bank as a countercycli- Overall, although government-backed cal tool in times of crisis has its advantages. credit guarantee schemes might help jump- This type of intervention not only has smaller start lending to certain borrowers in certain implementation lags but also is easier to sectors, these schemes are not likely to have unwind. Generating mechanisms to retrench large macroeconomic effects nor are they lending activity by state-owned banks is more likely to work as truly countercyclical tools. cumbersome, especially during the expan- Furthermore, they cannot substitute for sion that follows a crisis. However, legal reform of the underlying institutional require- constraints and credibility issues due to high ments of an effective credit system and should inflation episodes may make it dif�cult for not diminish the focus on these long-term emerging markets to use a more unconven- reforms. For instance, improving collateral tional approach to central banking and credit GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 125 stabilization. Also, emerging market econo- that the recovery of credit—although stron- mies do not have an international reserve ger because of a larger rebound effect—is currency, so quantitative easing measures slower. Microlevel evidence for the Latin may jeopardize the stability and value of their America and the Caribbean and Central and currencies. Finally, the monetary authority in Eastern Europe regions suggests that state- emerging markets can actively participate by owned banks were not universally respon- setting up open bank assistance lines to the sible for propping up credit during the recent � nancial sector. However, as in the case of crisis. Banks in Latin America and the Carib- bank ownership, this approach also requires bean seem primarily to have played a coun- strong corporate governance so as to avoid tercyclical role. political interference. Furthermore, the benefits of this short- A comprehensive examination of the con- term credit stabilization may come at a cost. duct of monetary policy, conventional or From experience, the associated costs of unconventional, during times of financial the intervention are likely to be steep, with stress is beyond the scope of this report. A agency and political economy problems lead- more detailed account of the response of ing to credit misallocation and economic the monetary authority during the recent inefficiency. Lending and investment deci- crisis can be found in Aït-Sahalia and oth- sions by state-owned banks (typically moti- ers (2010); Brave and Gesnay (2011); Gian- vated by political connections or allegiances) none and others (2011); and Lenza, Pill, and lead to a deterioration of the quality of loans Reichlin (2010).27 and, hence, to a greater misallocation of resources. So far, the �nancial and economic costs of recent state interventions are dif�cult CONCLUSION to ascertain since the overall effects are still Governments actively responded to restore unfolding. Hence, the potential bene�ts and credit conditions during the recent global costs require careful consideration by policy �nancial crisis. Emerging markets used their makers. The track record of state banks in state-owned banks mainly to inject liquidity credit allocation remains generally unimpres- in their �nancial markets, while many devel- sive, questioning the wisdom of using state oped economies effectively used their central banks as a countercyclical tool. bank to prop up the � nancial sector. Credit Ideally, the performance of state-owned guarantees were also actively used to foster banks can be improved by adopting a set of credit in some underserved segments of the good practices in their institutional design. A economy. clear and sustainable mandate is needed so Having state-owned banks may have that state-owned banks complement rather facilitated the flow of credit in some coun- than substitute private efforts in the credit tries during the recent global �nancial crisis. markets. An adequate risk management It helped stabilize aggregate credit in some system is required to guarantee the finan- emerging markets and restored credit condi- cial sustainability of the institution. Finally, tions. State-owned banks partially compen- sound corporate governance—and, more sated for the slowdown (and, in some cases, speci�cally, the elaboration of an ownership decline) in credit provided by (domestic and policy—is also a key factor for the optimal foreign) private banks. However, the evidence function of these banks. Unfortunately, these on the short-term benefits of government design features are challenging to implement bank lending is far from conclusive. Some in weak institutional environments, precisely macrolevel evidence, including the recent as where state-owned banks can make the well as previous crises, shows that credit fluc- greatest contributions—but often fail to do tuations are more volatile in countries with so because of these design weaknesses. high participation of state-owned banks (as Governments also devised strategies to opposed to those with low participation) and prop up the financial sector that did not 126 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 directly involve the use of state-owned unconventional monetary and �scal policies, banks. Both developing and developed coun- see Aït-Sahalia and others (2010), Claessens tries actively used credit guarantees during and others (2011), and Lenza, Pill, and Reich- the crisis to enhance credit to the sectors of lin (2010). 4. The main focus of this chapter is on direct the economy underserved by private finan- interventions using state-owned banks rather cial intermediaries (for example, small and than direct interventions in nonbank �nancial medium enterprises). Analogous to interven- institutions (NBFIs) and markets. The choice tions through state-owned banks, credit guar- to focus primarily on state-owned banks is antees may generate some short-term bene�ts, driven by the wider evidence among develop- but they often come with �scal and economic ing economies and the greater availability of costs in the form of contingent government data on banking rather than on nonbanking liabilities and misallocation of resources. The institutions. success of guarantee schemes hinges on their 5. Laeven and Valencia (2010) document the design features. Best practices include (a) leav- increase in public debt and direct �scal costs ing credit assessments and decision making to arising from the recent global �nancial crisis. 6. Financial additionality refers to whether the the private sector, (b) capping coverage ratios guarantee allows previous nonborrowers to and delaying the payout of the guarantee access credit. Economic additionality refers to until recovery actions are taken by the lender the real effects from a larger number of bor- so as to minimize moral hazard problems, rowers having access to credit as a result of (c) pricing guarantees to take into account guarantees (for example, more jobs created, the need for �nancial sustainability and risk �rms being able to invest more). minimization, and (d) encouraging the use of 7. Bonin, Hasan, and Wachtel (2005) provide risk management tools. However, as in the a detailed examination of the trends and case of state banks, these best-practice design consequences of bank privatization in transi- features are more challenging to get right in tion economies—namely, Bulgaria, Croatia, weak institutional environments. the Czech Republic, Hungary, Poland, and Romania). Finally, the monetar y authorit y in 8. The flagship report points out that the pen- advanced countries provided unprecedented etration of foreign banking in Africa wit- amounts of liquidity to the � nancial sector nessed another development over the past by expanding their balance sheets. Although 15 years: a marked increase in the share of easier to unwind than the expansion of the African banks among foreign banks, espe- balance sheets of government banks, the cially the expansion throughout the region of method’s successful implementation among South African banks (for example, Absa and emerging countries may be dif�cult because Standard Bank), West African banks (Bank of legal obstacles and policy credibility issues. of Africa and Ecobank), as well as Moroccan and Nigerian banks (Beck, Maimbo, and oth- ers 2011). 9. Clarke, Cull, and Megginson (2005) provide NOTES extensive evidence on the trends, determi- 1. State-owned banks can also be deployed for nants, and consequences of bank privatiza- populist reasons (Acharya 2011). tion both at the cross-country level and in 2. During the crisis, Germany’s KfW introduced country case studies. a “special countercyclical program� that pro- 10. Farazi, Feyen, and Rocha (2011) point out vided lines of credit and loan guarantees to that there is a great deal of heterogeneity in banks to keep credit to the economy flowing. terms of government ownership in the Europe For more details on the actions undertaken and Central Asia region. by governments in the �nancial markets, see 11. Recent evidence suggests that much of the Laeven and Valencia (2010, 2011), Rudolph decline in new lending reflects changes in the (2010), among others. supply as opposed to the demand for credit 3. For a more detailed look at the policy (Huang and Stephens 2011; Ivashina and responses of advanced countries in terms of Scharfstein 2010). GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 D I R E C T S TAT E I N T E R V E N T I O N S 127 12. Biggs, Mayer, and Pick (2010) argue that of government bank ownership on private recoveries in the real sector typically come credit growth. Galindo and Micco (2004) try along with a recovery in the flow of credit. to address the problem of causality by ana- The authors build a model in which an lyzing within-country differences in industry upturn in real GDP coincide with a recovery growth. They �nd that the development of in the flow of credit even as the stock of credit private �nancial intermediaries is associated decreases. Data from developed economies with a higher growth rate of industries that and emerging market economies con�rm that rely more on external �nance and have less premise of their model. collateral, while public bank ownership has 13. Tier 2 credit operations refer to lending by no effect on the growth of these industries. state-owned banks through other private 20. Dinç (2005) shows that increased lending �nancial institutions to reach agents in the by government-owned banks during elec- economy. tion years is not speci�c to India but is also 14. De la Torre, Gozzi, and Schmukler (2007) pro- observed in a sample of 19 emerging markets vide a good summary of the different views on (but not in developed economies). Micco, the role of the state in the �nancial sector. Panizza, and Yañez (2007) show that these 15. The development view is also referred to as results hold for a much larger sample of “the interventionist view� (de la Torre, Gozzi, developing economies and that the increased and Schmukler 2007). lending by government-owned banks during 16. The political view is included in what de la election years is associated with a decrease in Torre, Gozzi, and Schmukler (2007) refer to their interest rate margins and pro�tability. as “the laissez faire� view. 21. Acharya (2011) shows that government poli- 17. IDB (2005) reviews the empirical evidence cies in the housing market that enhanced the on the impact of government-owned banks ability of the government-sponsored enter- and �nds that, although the result that these prises Fannie Mae and Freddie Mac to take banks have a negative impact is not as strong riskier mortgages led to the deterioration of as previously thought, there is no indica- the quality of their assets and a competitive tion that government ownership has a posi- “race to the bottom� among �nancial inter- tive effect. It concludes that public banks, at mediaries. best, do not play much of a role in �nancial 22. De la Torre, Gozzi, and Schmukler (2007) development. Andrianova, Demetriades, and provide an extensive description of the dif- Shortland (2010) and Körner and Schnabel ferent advances in the provision of noncredit (2010) conducted cross-country analyses of services by Mexican state development banks. the effects of government bank ownership on Other well-established state-owned banks �nancial development and macroeconomic include Japan Development Bank, Kredi- outcomes, which conclude that government tanstalt fuer Wiederaufbau (KfW) in Ger- bank ownership might not be as harmful to many, Kommunalbank in Norway, as well the development as suggested by earlier studies. multilateral development banks, all of which However, these studies cannot convincingly have remained pro�table for decades. dispel concerns about omitted variables and 23. Podpiera (2011) also documents the use of reverse causality biases. credit subsidies as a countercyclical crisis 18. In contrast with the observed lower pro�t- measure in Serbia. ability and ef�ciency of government-owned 24. This is what Anginer, de la Torre, and Ize commercial banks in developing economies, (2011) call collective action frictions. Fur- the empirical evidence suggests that in devel- thermore, they argue that in a world devoid oped countries there are no signi�cant dif- of risk aversion and collective action frictions, ferences in performance between public and agency frictions (such as informational asym- private banks (see, for example, Altunbas, metries or adverse selection) cannot justify Evans, and Molyneux 2001; Micco, Panizza, government guarantees. They argue that only and Yañez 2007). partial equilibrium models that ignore the 19. For example, Levy Yeyati, Micco, and Panizza welfare effects of the taxes needed to �nance (2007) try to deal with the endogeneity prob- the guarantees come to the flawed conclusion lem by using panel data and by conducting that guarantees are not justi�ed by agency GMM System estimations and �nd no impact frictions. 128 D I R E C T S TAT E I N T E R V E N T I O N S GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 25. It should be noted that though many coun- 27. Countercyclical �scal policy can complement tries did not have a signi�cant presence of the actions from the central bank and state- state-owned banks before the crisis, measures owned banks by containing the response of taken during the crisis, such as massive bail- aggregate expenditure to changing �nancial outs and nationalizations, amounted to the conditions. However, discretionary counter- socialization of risks and losses that might cyclical deployment of �scal policy often faces also bring negative consequences to the econ- considerable delays, which limit its ability to omy down the line. counteract a sudden crash in a timely manner. 26. For instance, the Federal Reserve Board of the This limitation underscores the need to build United States aggressively expanded its bal- up self-deploying automatic stabilizers, which ance sheet in two different stages. First, it tried are still weak in most developing economies, to contain the stress in �nancial markets by including Latin America (Claessens and oth- providing loans to �nancial institutions and ers 2010; Debrun and Kapoor 2010). injecting liquidity in key markets. Second, it launched a massive asset purchase program. 5 The Role of the State in Financial Infrastructure • The global �nancial crisis has highlighted the importance of a resilient �nancial infra- structure and reignited the debate on what role the state should play in its develop- ment, particularly in (a) promoting the availability and exchange of reliable credit information and (b) supporting the development of institutions to better manage counterparty risk in interbank markets and securities transactions. • Transparent credit information is a prerequisite for sound risk management and �nan- cial stability. However, due to the prevalence of monopoly rents in the market for credit information, information sharing among private lenders may not arise natu- rally. This creates an important rationale for the involvement of the state. • Existing credit reporting systems contain extensive information on credit risks in the �nancial sector. There is signi�cant potential for improving their use for risk manage- ment and prudential supervision. • Many credit reporting systems cover only risks in the traditional � nancial sector. This limits their effectiveness in supporting credit market ef�ciency and stability. An important role of the state is to help extend the coverage of credit reporting systems to include nonregulated lenders, such as nonbank �nancial institutions and micro�nance lenders, in existing credit reporting systems. • The state can help establish market infrastructure that helps to manage and mitigate counterparty risk. This includes robust large-value payment systems and, potentially, support for the development of collateralized interbank markets. • There is signi�cant scope for state involvement in the development of a robust infra- structure for securities and derivatives settlements. The state can further reduce counterparty and settlement risks by monitoring these transactions and their clearing and settlement arrangements. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 129 130 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 he global � nancial crisis has high- the important role of the state in establish- T lighted the importance of a resilient � nancial infrastructure and reignited the debate on what role central banks and ing a legal and regulatory framework that allows open and transparent credit reporting to emerge. The chapter emphasizes the chal- other state agencies should play in its develop- lenges posed by market segmentation and ment. Financial infrastructure, as de�ned in the prevalence of monopolies in the market this report, consists of credit reporting insti- for credit information, which may provide a tutions (credit registries and bureaus), pay- rationale for the measured involvement of the ment and settlement systems, and the legal state. The chapter then turns to the lessons framework that governs � nancial transac- of the �nancial crisis for credit reporting and tions.1 A well-developed �nancial infrastruc- discusses how existing credit information ture makes credit markets more ef�cient by systems can be used as a tool for prudential reducing information asymmetries and legal oversight and regulation. uncertainties that may hamper the supply of The second part of the chapter discusses new credit. This improves the depth of credit the role of the state in ensuring the stability market transactions and broadens access to of payment and securities settlement systems. �nance. The global �nancial crisis has also The chapter argues that large-value payment renewed interest in the role of �nancial infra- systems around the world have demonstrated structure in supporting systemic stability. remarkable stability during the global �nan- Financial infrastructure promotes � nancial cial crisis, thanks in large part to the wide- stability in several ways. Transparent credit spread adoption of modern real-time settle- reporting can support the internal risk man- ment systems over the past two decades. The agement of �nancial institutions and supply crisis nonetheless revealed several areas for �nancial regulators with timely information policy improvements. In particular, the chap- on the risk pro�le of systemically important ter argues that the state can further reduce �nancial institutions. Similarly, well-designed counterparty risk by supporting the develop- payment and security settlement systems ment of collateralized interbank markets and enhance �nancial stability by reducing coun- derivatives settlement systems, particularly in terparty risk in interbank markets and com- countries where the development of a modern plex securities and derivatives transactions. settlement infrastructure has lagged the rapid The role of the state in financial infra- growth of equity and securities markets. structure has varied over time and across countries. This chapter examines how state CREDIT REPORTING agencies and central banks can operate, reg- ulate, and oversee financial infrastructure. Introduction The focus is on two areas: � rst, the state’s role in developing and using credit informa- Transparent credit information is a prerequi- tion systems, and second, the state’s role in site for sound risk management and �nancial improving payment and securities settlement stability. Credit reporting institutions sup- systems. The chapter does not examine, for port �nancial stability and credit market ef�- example, retail payment systems or the legal ciency and stability in two important ways. framework that governs financial transac- First, banks and nonbank � nancial institu- tions. Reflecting the report’s focus on the tions (NBFIs) draw on credit reporting sys- � nancial crisis, these areas will be explored tems to screen borrowers and monitor the in future editions. risk pro� le of existing loan portfolios. Sec- The �rst part of the chapter focuses on the ond, regulators rely on credit information to role of the state in credit reporting. It reviews understand the interconnected credit risks the evidence on credit information and �nan- faced by systemically important borrow- cial stability and the public good nature of ers and �nancial institutions and to conduct information sharing. The chapter highlights essential oversight functions. Such efforts GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 131 reduce default risk and improve the ef�ciency developing the capabilities of private credit of �nancial intermediation. In a competitive reporting institutions through public private credit market, these efforts ultimately bene�t partnerships and institutional innovation. consumers through lower interest rates. The World Bank’s General Principles for Effective credit reporting systems can Credit Reporting (2011a) reviews best prac- mitigate a number of market failures that tices and makes policy recommendations for are common in financial markets around developing credit reporting systems. the world, and most severely apparent in less developed economies. The availability of Credit information as a public good high-quality credit information, for example, reduces problems of adverse selection and The open and transparent exchange of credit asymmetric information between borrowers information has several characteristics of a and lenders (Stiglitz and Weiss 1981; Jappelli public good that bene�ts both borrowers and and Pagano 2002; Pagano and Jappelli 1993). lenders. However, because lenders can use This reduces default risk and improves the the information advantage over their exist- allocation of new credit. Information shar- ing clients to extract monopoly rents, credit ing can also promote a responsible “credit information sharing does not always arise culture� by discouraging excessive debt and naturally. The state therefore plays an impor- rewarding responsible borrowing and repay- tant role in promoting the exchange of credit ment (de Janvry, McIntosh, and Sadoulet information and in protecting open and equal 2010; Padilla and Pagano 2000).2 access to the market for credit information. Perhaps most important, credit report- There are at least three areas in which a ing allows borrowers to build a credit his- well-functioning credit reporting infrastruc- tory and to use this “reputational collateral� ture performs the role of a public good. First, to access formal credit outside established credit reporting bene�ts banks and nonbank lending relationships. This is especially ben- lenders by mitigating problems of moral haz- e�cial for small enterprises and new borrow- ard and adverse selection. Detailed infor- ers with limited access to physical collateral mation on the credit history of individual (Djankov, McLiesh, and Shleifer 2007; Love borrowers allows banks to improve the ex and Mylenko 2003; see also Padilla and ante screening of prospective clients as well Pagano 2000). Stylized evidence from the as the ex post monitoring of credit risks in recent �nancial crisis also suggests that posi- their existing loan portfolios. This, in turn, tive credit information helped to safeguard reduces the cost of � nancial intermediation the �nancial access of creditworthy borrow- and allows banks to price, target, and moni- ers that would have otherwise been cut off tor loans more effectively. from institutional credit (Simovic, Vaskovic, Second, credit reporting supports � nan- and Poznanovic 2009). This �nding is consis- cial stability by making it easier for � nancial tent with evidence from the literature on rela- regulators to assess and monitor systemic tionship banking (Berger and others 2003; risks. Although traditional approaches to Petersen and Rajan 1995), which emphasizes financial oversight have focused on risks how access to more detailed client informa- at the level of individual � nancial institu- tion can facilitate pro�table lending to infor- tions, a key advantage of comprehensive mationally opaque borrowers, such as start- credit information systems is that they allow ups and small enterprises (Mian 2006).3 regulators to monitor the interconnected The World Bank Group has supported risks of systemically important financial the development of credit reporting systems institutions. around the world for more than a decade. While the recent �nancial crisis has under- The International Finance Corporation’s scored this important function of credit Credit Bureau Knowledge Guide (IFC 2006) reporting, it has also revealed a number of provides an overview of experiences in limitations of current credit information 132 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 systems. In many countries, regulators have Public and private credit reporting access to credit information only from regu- around the world lated �nancial institutions but lack access to similarly comprehensive data on nonregu- The role of the state in credit reporting has lated lenders. Extending the reach of credit varied widely across countries and over time. information systems to nonregulated lenders Two main types of credit reporting insti- to better capture systemic risks outside the tutions can be found around the world: (a) traditional banking sector is an important credit registries, which are public entities that policy prescription that has emerged from the are managed by bank supervisors or cen- recent � nancial crisis. At the same time, the tral banks and typically collect information unfolding of the �nancial crisis in the United from supervised financial institutions, and States and around the globe has shown that (b) credit bureaus, which are privately owned the availability of high-quality credit infor- enterprises that tend to cover smaller loans, mation is a necessary but not a sufficient often collect credit information from bank condition to promote financial stability: and nonbank lenders, and provide a range of even where a well-developed information value-added services, such as credit scores, to infrastructure exists, it needs to be accom- banks and nonbank lenders. panied by regulatory incentives that reward Historically, public and private credit the appropriate use of available information reporting institutions have evolved to serve in the evaluation and management of credit different purposes. Credit registries gener- risk. ally developed to support the state’s role as Third, open and transparent credit report- a supervisor of � nancial institutions. Where ing benefits bank customers by promoting credit registries exist, loans above a certain credit market competition. The exchange of amount must, by law, be registered in the credit information enables customers to build national credit registry. In some cases, credit reputational collateral and to access credit registries have relatively high thresholds for outside established lending relationships. loans that are included in their databases. This reduces the ability of established lend- Credit registries tend to monitor loans made ers to exploit their privileged knowledge of by regulated � nancial institutions and usu- clients’ credit histories. Because open access ally do not offer value-added services, such to credit information erodes the information as credit scores or collection services. Against monopoly of individual lenders, banks are the backdrop of the financial crisis, many often reluctant to share such information— countries have made efforts to optimize the especially positive information that may use of credit registry data for prudential over- empower borrowers with good credit histo- sight and regulation. ries to seek credit elsewhere. Similarly, where Credit bureaus, by contrast, are privately credit reporting institutions exist, larger owned commercial enterprises catering to �nancial institutions often have an incentive the information requirements of commer- to prevent equitable access to credit informa- cial lenders. Though there is variation in the tion through anticompetitive pricing or the type and extent of information they collect, formation of closed user groups, despite the credit bureaus generally strive to collect very positive ef�ciency implications that improved detailed data on individual clients. They access to credit information would have on therefore tend to cover smaller loans than the �nancial system as a whole. registries and often collect information from Taken together, the positive implications a wide variety of � nancial and non� nancial of credit reporting for �nancial stability and entities, including retailers, credit card com- credit market ef�ciency create an important panies, and micro� nance institutions. As a rationale for an active role of the state in pro- result, data collected by credit bureaus are moting the development of an effective credit often more comprehensive and better geared reporting infrastructure. to assess and monitor the creditworthiness of GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 133 individual clients. Compared to credit regis- To provide an overview of the state of tries, private credit bureaus are a relatively public and private credit reporting around recent institution. Although credit bureaus the world, this section presents data on the have existed in Germany, Sweden, and ownership structure and extent of informa- the United States for nearly a century, they tion collected by credit bureaus and regis- emerged in many other high-income coun- tries. Map 5.1 shows the prevalence of credit tries, including France, Italy, and Spain, as reporting institutions around the world. The recently as the 1990s.4 maps and summary statistics in table 5.1 MAP 5.1 Credit Information Systems around the World a. Global distribution of credit registries at least one credit registry no credit registry b. Global distribution of credit bureaus at least one credit bureau no credit bureau Source: Doing Business Indicators database. 134 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 TABLE 5.1 Credit Reporting, Coverage by Region Credit registry coverage Credit bureau coverage Region % of population % of GDP % of population % of GDP East Asia and Paci�c 8.2 60.5 17.3 20.2 Eastern Europe and Central Asia 13.1 38.9 21.3 35.8 Latin America and the Caribbean 10.1 19.5 31.5 18.1 Middle East and North Africa 5.3 53.2 7.0 13.2 OECD 8.0 157.1 61.1 36.6 South Asia 0.8 46.2 3.8 108.4 Sub-Saharan Africa 2.7 16.6 4.9 7.9 Source: Calculations based on Doing Business Indicators database. Note: GDP = gross domestic product. show that there is some striking geographic credit information infrastructure in place. variation in the existence of public and pri- The first credit registries emerged in the vate credit reporting institutions. Overall, United States in the 1830s in response to credit registries are more prevalent in coun- recurring episodes of defaults and � nancial tries with a French legal tradition, whereas instability, but credit reporting institutions private credit reporting is more widespread in did not arise in many other countries until countries of British legal origin. In a number much later. Especially in countries with a of countries—primarily in Latin America— European common law tradition, legal bar- private and public credit reporting systems riers to disclosing credit information have coexist, often catering to distinct segments of often constrained the development of private the credit market. credit reporting (Djankov, McLiesh, and Figure 5.1 looks at the evolution of credit Shleifer 2007; Olegario 2003). Despite these reporting institutions over time. As late as the obstacles, which persist in many countries, early 1980s, few countries had a signi�cant credit information has expanded rapidly. The number of credit markets covered by either private or public credit reporting systems FIGURE 5.1 The Development of Credit Reporting Institutions, (or both) almost tripled over the past two 1980–2012 decades. The effectiveness of a credit reporting sys- Percent tem is determined by the quality and depth 90 of information it makes available to market participants. To assess the quality of infor- 80 mation sharing, this report focuses on three 70 important characteristics of a country’s 60 credit reporting system: (a) the coverage of 50 the credit reporting system, measured by the 40 number of borrowers or the volume of credit listed in the credit reporting system (see sum- 30 mary �gures in table 5.1 and �gure 5.2); (b) 20 the extent of institutional participation (that 10 is, which types of �nancial and non�nancial 0 institutions exchange information through 1980 1984 1988 1992 1996 2000 2004 2008 2012 the credit reporting system); and (c) the depth Countries with credit registry Countries with credit bureau of credit information (that is, what kind of information on borrowers and credit risk is Source: Calculations based on Doing Business Indicators database. tracked). GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 135 FIGURE 5.2 Prevalence of Credit Reporting payments. Each information item can be clas- by Income Group si�ed as personal information, loan informa- tion, or information on a client’s repayment history. Each index sums the range of infor- Percent mation items contained in the credit registry 80 or bureau and normalizes the resulting score 70 so that the summary index lies between 0 60 (poor information content) and 1 (high infor- mation content). 50 Figure 5.4 reveals some striking differ- 40 ences in the type of information collected 30 by credit bureaus and credit registries. On average, credit registries and credit bureaus 20 collect approximately the same extent of 10 information on the personal or identify- 0 ing information of borrowers. In line with Low income Middle income High income their historical role as a supporter of the Credit registry Credit bureau state’s supervisory function, registries tend to record more-detailed information about Source: Calculations based on Doing Business Indicators database. the type, terms, and structure of individual loans. The information collected by credit bureaus, on the other hand, is much more Figure 5.3 presents evidence on the reach geared toward tracking the repayment his- of credit reporting institutions by summa- tory of individual borrowers in order to rizing which types of financial institutions provide commercially viable data to market participate in the exchange of credit informa- participants. tion. The �gure shows that credit registries are less likely than credit bureaus to contain FIGURE 5.3 The Reach of Credit Reporting: Who Contributes data from nonregulated �nancial institutions. Information? Nearly all credit registries collect information from banks, whereas only 67 percent of reg- istries contain information from any unregu- Percent lated lender. Credit bureaus are more likely 100 to cover NBFIs such as leasing and retail 90 finance companies and microfinance lend- 80 ers—and may therefore be better suited to 70 promote �nancial access of new borrowers. 60 Turning to the quality and depth of avail- 50 able credit information, �gure 5.4 compares 40 the type of credit information collected by credit registries and credit bureaus, respec- 30 tively. To do so, the �gure presents four infor- 20 mation indexes based on the Doing Business 10 data. For each credit reporting institution, the 0 Regulated Nonregulated Banks Nonbanking Microfinance data set provides information on the different lenders lenders financial lenders types of information collected and reported institutions by the credit registry or bureau. Examples of Credit registries Credit bureaus information items include customer age, total liabilities, or data on previous defaults or late Source: Calculations based on Doing Business Indicators database. 136 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 5.4 The Depth of Credit Reporting: What Information Is credit reporting infrastructure. First, state Collected? actors, such as central banks and � nancial regulators, can both operate and use credit reporting systems. Second, the state can act Percent as a regulator of credit reporting systems, 70 compelling private lenders to exchange high- 60 quality credit information and ensuring open and equal access to credit reporting systems. 50 Finally, the state can act to promote the 40 development of a private credit reporting infrastructure that can complement the role 30 of public credit registries in supporting credit 20 market ef�ciency. This section reviews each of these roles in turn and provides examples 10 of the challenges and opportunities of the 0 state’s involvement in credit reporting, with Overall Personal Loan Repayment a special focus on the state’s changing role Credit registry Credit bureau in light of the global � nancial crisis. Source: Calculations based on Doing Business Indicators database. The state as a user of credit information: Risk management and supervision The global financial crisis has generated Despite the growth of credit reporting renewed interest in the use of information institutions, their development has been from national credit registries for pruden- highly uneven across and within regions. In tial oversight and regulation (Girault and many emerging credit markets, the develop- Hwang 2010). Although many countries’ ment of credit reporting systems remains registries collect detailed data on loans, the constrained by the lack of an appropriate crisis has highlighted the need for improve- legal infrastructure needed for the volun- ments in the use of existing credit infor- tary exchange of credit information. In many mation for � nancial oversight and regula- countries, privacy laws have no provision for tion. The new Basel III accords, which are credit reporting or, in some cases, prohibit being adopted around the world, present the disclosure of vital information to third an important window of opportunity to parties altogether. Finally, as the chapter improve the use of credit information for the discusses in greater detail in the following purpose of identifying and managing threats sections, the structure of competition in the to � nancial stability. banking sector can pose a signi�cant obsta- Currently, financial regulators use data cle to the emergence of comprehensive credit from credit registries primarily for the off- reporting systems. This creates an important site monitoring of credit risks. Credit reg- role for the state in establishing an appropri- istry data can support this task in several ate legal framework in which transparent ways. First, they allow regulators to esti- credit reporting can evolve. mate the portfolio credit risk and calculate loan loss provisions for individual � nancial institutions. Second, they enable regulators CREDIT REPORTING to compare loan credit risk across banks, to AND THE STATE conduct stress tests, and to detect anomalies The state can play three main roles in sup- in lending patterns, portfolio structure, and porting the development of a transparent loan performance. Third, registries track GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 137 and monitor the development of credit risk themselves, credit registries can provide a by type of borrower or type of credit. This wealth of data for this purpose (see box 5.1). allows regulators to detect credit risks con- Since credit registries generally contain infor- centrated in a speci�c sector, loan category, mation on all loans above a given threshold, or region. making full use of this information allows The Basel III accords represent a shift regulators to obtain a more comprehensive from a microprudential approach to �nancial picture of interconnected risks in the � nan- regulation, centered on the risk of individual cial sector. �nancial institutions, toward a macropruden- The main challenges to the use of existing tial approach that is focused on systemically credit registry data for prudential supervision important �nancial institutions. Understand- are the limited coverage of unregulated �nan- ing and containing these threats requires cial institutions and the often high minimum detailed understanding of interconnected loan sizes of existing credit registries, which credit risks in the traditional � nancial sys- further limit the range of loans captured by tem as well as risks outside it. There is much the registry. The unfolding of the � nancial room to leverage data from credit registries crisis in Europe and the United States has for this purpose. Credit registry data can pro- shown that � nancial stability is increasingly vide the basis for affected by the risks taken on by nonregu- lated � nancial entities. In developed econo- • Evaluating the systemic importance mies this market includes hedge funds, money of financial institutions. The Basel III market funds, and structured investment accords require more stringent provisions vehicles, sometimes referred to as the shadow for systemically important �nancial insti- banking system. In many emerging markets, tutions. Comprehensive credit informa- nonbanking financial institutions play an tion is required to assess and monitor their important role in consumer credit, which has interrelated exposures. often remained outside the scope of informa- • Informing countercyclical buffer deci- tion available to �nancial regulators. sions. The countercyclical capital frame- In its role as a user and regulator of credit work introduced by the Basel III accords information, the state can therefore play an tries to reduce the cyclicality of bank lend- important role in extending the coverage of ing. It introduces a conservation buffer, existing credit reporting systems to new and set at 2.5 percent of banks’ risk-weighted systemically important borrower groups, and assets, and a countercyclical capital buffer in incentivizing regulators to make appropri- to address potentially excessive risk-taking ate use of credit reporting systems for iden- as a result of cyclical credit growth. Bet- tifying and monitoring threats to systemic ter credit information can increase the stability. accuracy of risk weighting in banks’ loan portfolios. The state as a regulator of credit • Building more reliable early warning sys- reporting: Removing barriers to tems. Early warning systems used by bank information sharing regulators often only capture portfolio risk for individual � nancial institutions. Another important role of the state is to Better use of credit registry data for off- ensure that the market for credit informa- site monitoring allows for more nuanced tion remains transparent, open, and ef�cient. tracking of links in credit risk exposures Because banks can extract information rents across institutions. from proprietary credit information, lenders may try to retain monopolistic knowledge Though some of these tasks can be accom- of their clients’ creditworthiness by sharing plished with data from �nancial institutions only limited, inaccurate, or incomplete credit 138 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 5.1 Argentina: Using Credit Registry Information for Prudential Supervision Argentina’s central bank, the Banco Central de la tify and contain a deterioration in the quality of loan República Argentina, has been operating a credit portfolios, (b) facilitate provisioning and supervision reporting system since 1991. The system is mostly of provisioning requirements for credit risk, and (c) focused on large loans and has been increasingly facilitate the debt refinancing process. Monitoring used for supervisory purposes. In 1995, the system credit risk and informing banks of these risks remain was reformed, and access to data was granted to primary objectives of the Argentine credit registry. financial institutions. After Argentina’s financial I n it ia l ly, A rgent i na’s publ ic cred it reg is- crisis in 2003, several reforms of Argentina’s credit try included only information on debts above reporting system were undertaken to facilitate its use US$200,000. As a means of broadening the coverage for prudential regulation and to support greater sta- of the system, the minimum threshold was reduced bility in the � nancial sector. to US$50,000 in 2002. To further improve the qual- Currently, the Argentine credit reporting indus- ity of recorded credit information, the central bank try comprises a public credit registry and several has taken important steps toward the establishment private sector credit bureaus. The legal and regula- of a � nancial statements database and has begun to tory framework covering credit reporting activities collect new information on loans already covered by in Argentina is limited to data protection compli- the system (including credit lines, currency denomi- ance, which places some limitations on the authori- nation, and maturity structure of outstanding liabil- ties’ ability to adopt a holistic approach to credit ities). These improvements in data collection have reporting. The Argentine central bank currently acts been particularly helpful in facilitating the imple- as an operator of databases. Regulated entities are mentation of portfolio models of credit risk based on required to report their credit exposures to the cen- data from the registry. In particular, the expanded tral bank, which makes this information available data can be used to check that provisions for credit to private credit reporting agencies.a Several private risk properly cover banks’ expected and unexpected credit bureaus operate in Argentina, with the largest losses and serve as the foundation of scenario analy- credit bureau, Veraz, holding data on approximately ses and stress tests. For example, credit registry data 90 percent of all credit lines in the market. can be used for simulations to test for the effects In the aftermath of the Argentine economic crisis, of new internal ratings–based capital standards as the central bank focused on a strategy to enhance the envisaged in the Basel III accords. availability of credit information for risk management Taken together, these changes have signifi- and prudential supervision. This strategy included cantly enhanced the central bank’s ability to lever- extending the coverage of the system and the collec- age Argentina’s existing credit reporting system for tion of new information. In addition to providing data the purpose of prudential oversight and regulation. useful for prudential oversight and regulation, these Some opportunities for improving the capabilities reforms also aimed to facilitate the restructuring of of the system nonetheless exist, for example, with the banking sector. The aim of these reforms was to regard to tracking the risk pro� le of securitized loans make it easier for banks and regulators to (a) iden- and monitoring loan portfolios after origination. a. Initially, this arrangement came into existence because Argentina’s leading banks failed to agree on a mechanism for the voluntary exchange of credit information (see Berger and others 2003). In response, the Argentine authorities required banks to share information and made the data available to private credit reporting agencies on an equal basis. Private credit reporting agencies may access data from the central bank’s databases and offer value-added services. information. The state can act to overcome As a � rst step, the government can help such barriers to information sharing in a lay the legal foundations for effective infor- variety of ways, ranging from the establish- mation sharing. This includes a sound legal ment of an appropriate legal framework to framework that maintains consumer pro- direct interventions, mandating the exchange tection but at the same time permits banks of credit information. and nonbank lenders to share relevant GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 139 information with credit reporting institu- in the information-sharing framework. In tions. The legal framework that governs addition, capturing information from non�- the exchange of credit information should nancial entities can provide information that include provisions that limit the data that gives greater insight into borrowers’ pay- can be shared among lenders to just infor- ment behavior, for example, vis-à-vis utility mation that is relevant for the lending deci- companies and retailers. Enabling compa- sion (for example, to prevent discrimination nies outside the traditional banking sector to in lending). 5 It should also grant borrowers share credit information is advantageous for the right to appeal incorrect information. borrowers because it facilitates the establish- At the same time, government and regula- ment of a credit history, but it may require tors need to ensure that consumer protec- additional changes to existing laws. The col- tion laws are clear and that the administra- lection of such additional information may, tive burden for compliance does not in fact however, also place greater consumer protec- reduce incentives for information sharing. tion responsibilities on the regulator. Brazil’s data protection law, the Consumer Although having a sound legal framework Protection and Defense Code, for example, in place is crucial for allowing the market requires lenders to notify consumers when- for credit information to develop, it may not ever their data are updated. It has been be enough to ensure a level playing �eld for argued that this procedure is so costly that it the exchange of credit information. As high- prevents lenders from sharing positive credit lighted in this section and in Bruhn, Farazi, information and constrains credit reporting and Kanz (2012), market failures such as more generally (OECD 2010). monopoly rents and coordination problems In some countries, the existing legal are prevalent in the market for credit infor- framework may already balance the goals mation and can create important barriers to of protecting consumers while allowing the development of a private credit reporting for broad and comprehensive credit report- infrastructure. ing. However, in many countries, the exist- Since credit information is a public good, ing legal infrastructure may constrain the it is most effective when contribution and exchange of credit information. This was the access to credit reporting systems are widely case in Egypt (see box 5.2). Until recently, shared. For an individual lender, the ben- Egypt’s existing banking and data protec- e�ts of joining a credit bureau depend on the tion laws were highly restrictive and did not number of other members. Setting up a credit allow lenders to disclose client information to bureau thus requires extensive coordination the market. Amendments to the legal frame- and collaboration among lenders. In practice, work for credit information sharing had to this coordination may be dif�cult to achieve. be made to allow the exchange of informa- In fact, experience has shown that commit- tion among lenders, credit bureaus, and the ment by lenders can be a major problem in central bank without obtaining borrower establishing credit bureaus in developing consent for each report. The Central Bank of economies. In addition, the private interests Egypt was instrumental in helping to bring of banks and other information providers about these changes and helped to create a may also get in the way of sharing deep and framework conducive to the operation of a comprehensive information among a large credit bureau covering both banks and non- number of entities, since individual banks bank lenders. can capture monopoly rents by not sharing When considering whether any changes information. That is, lenders benefit from to existing laws are necessary, policy makers having information on borrowers from other should keep in mind that information shar- lenders, but at the same time they can pro�t ing may involve not only banks. Other lend- from not sharing their own information with ers, such as mortgage � nance, leasing, and other lenders. Large banks may therefore be credit card companies may also be included particularly reluctant to share proprietary 140 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 5.2 Egypt: Removing Regulatory Barriers to the Development of a Private Credit Bureau The � rst private credit bureau in Egypt—“I-Score�— that offers services in Arabic and English. I-Score’s was established in September 2005 and became data center has been vastly expanded to include 9 operational in March 2008. I-Score’s sharehold- million data records, a 13-fold increase from the ers consist of 25 banks and Egypt’s Social Fund for baseline of 0.9 million facilities initially held by Development. In Egypt, data secrecy laws posed the Central Bank of Egypt’s Public Credit Regis- a major obstacle to the establishment of a private try. The data pertain to over 4 million small and credit reporting infrastructure. The Central Bank medium enterprises and consumer borrowers. of Egypt was highly instrumental in creating a leg- I-Score currently services the credit information islative framework conducive to the operations of a needs of 55 institutional subscribers, which include private credit bureau. 41 banks, eight mortgage � nance companies, four The World Bank Group, through the Interna- leasing companies, the Egyptian Social Fund for tional Finance Corporation, provided implementa- Development, and one retailer. All banking insti- tion support accompanying the launch of I-Score in tutions and the Social Fund for Development have 2008. On the legislative front, existing laws were completed the credit data migration process to amended to allow the exchange of information I-Score. Mortgage � nance companies have submit- among banks, mortgage � nance and � nancial leas- ted approximately 65 percent of their data records, ing companies, credit bureaus, and the central bank and the four leasing companies have submitted 35 without obtaining individual borrower consent. percent of their data. I-Score has devised a speci�c The new legislation also specifies which users— package for MFIs to help them join in the credit that is, subscribers of I-Score—have a legitimate information–sharing scheme; that is, special prices purpose to inquire, obtain credit reports, and use have been agreed upon for MFI lenders, technical the services provided by the bureau. In September support is being offered, a free trial period for new- 2006, I-Score contracted with an international comers is being granted, and the development of ad partner, Dun & Bradstreet, to provide software hoc services is part of the package tailored to the solutions, enhance operational know-how, and speci�c needs of Egyptian MFIs. build the technological capability for the manage- The role of I-Score is to provide Egyptian facil- ment of its database. ity grantors with accurate, factual information Since the beginning of 2011, I-Score has been relevant to the history and payment habits of their proactive in winning the confidence of microfi- existing or prospective clients, enabling them to nance institutions (MFIs) to participate in the better assess their clients’ creditworthiness. To Egyptian information-sharing scheme. Initially, date, the effects on � nancial access and risk man- Egypt’s MFIs envisioned a separate credit bureau agement have been very impressive. For example, for microfinance clients. However, a pilot study since 2008 the consumer loans have been on an highlighted how much relevant borrower infor- ascending scale while the nonperforming loans mation would remain invisible to MFIs in a seg- have decreased by signi� cant percentages. I-Score mented credit information system. This convinced also aims to educate the general public of the val- the country’s leading MFIs to defer their decision ues, bene�ts, and consequences of owning a good to establish a separate micro� nance credit bureau credit � le. Therefore, it plays a major role in chang- and to join I-Score instead. Including MFI clients ing and modifying the behavior and culture of bor- in the credit reporting system will prevent the nega- rowers in the Egyptian credit market. Reforms to tive effects of data fragmentation and enhance the the legal and regulatory framework governing the use of credit bureau data for risk management and exchange of credit information were essential in � nancial inclusion. removing barriers to the establishment of a well- Since its inception, I-Score has managed to functioning credit reporting system in Egypt and establish a transparent and advanced credit bureau laid the foundation for I-Score’s success. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 141 credit information. Empirical evidence (see competition in the market for credit infor- box 5.3) shows that credit bureaus are indeed mation. In doing so, the state as a regulator less likely to emerge in markets where the has to balance the need to counter monopo- banking sector is highly concentrated and listic tendencies, while avoiding excessive dominated by a small number of lenders. market fragmentation. Although no consen- This � nding highlights that policies or reg- sus exists about the optimal degree of com- ulatory interventions intended to support petition in credit reporting, there is agree- the development of a comprehensive credit ment among regulators and policy makers reporting system need to be informed by an that regulatory oversight should (a) ensure a understanding of the underlying structure of level playing �eld for new entrants into the credit market competition. credit information market, (b) ensure open Where a well-functioning private credit and equal access to credit information sys- reporting infrastructure is in place, the state tems for regulated and unregulated lenders, plays an important role in safeguarding (c) identify and eliminate anticompetitive BOX 5.3 Monopoly Rents, Bank Concentration, and Private Credit Reporting Although the existence of a comprehensive credit tration is negatively associated with the probability reporting system is bene�cial for the � nancial mar- that a credit bureau emerges. Table B5.3.1 illustrates ket as a whole, individual lenders may profit from that 80 percent of countries with low bank concen- sharing only limited information with other market tration have a credit bureau, whereas only 39 per- participants. If only one lender has credit information cent of countries with high bank concentration have on � rms or individuals, this lender faces less com- a credit bureau. This difference is smaller for credit petition in lending to these borrowers because other registries (56 percent versus 37 percent), which may institutions may be reluctant to offer them credit. In reflect the fact that banks are required to report to a economic terms, a lender can capture monopoly rents credit registry while participation in a credit bureau from not sharing information. This issue may be par- is often voluntary. ticularly pronounced when the market for credit is This result is robust for controlling for confound- dominated by a few large banks. These banks each ing factors that could bias the analysis. In addition, have a broad customer base already and may try to the data also show that higher bank concentration is maintain their large market share by holding onto associated with lower coverage and quality of infor- information. Not making information available can mation being distributed by credit bureaus. These also prevent entry from new banks. findings suggest that market failures can prevent Bruhn, Farazi, and Kanz (2012) study the rela- the development of effective credit-sharing systems, tionship between bank concentration and the emer- implying that the state may have to intervene to help gence of private credit reporting. Using data for close overcome these obstacles. to 130 countries, the authors � nd that bank concen- TABLE B5.3.1 Bank Concentration and Credit Reporting Countries with low Countries with high bank concentration bank concentration Credit registry? 0.56 0.37 Credit bureau? 0.80 0.39 Credit bureau or registry? 0.92 0.53 Source: Bruhn, Farazi, and Kanz 2012. 142 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 pricing policies, and (d) prevent the forma- Finally, the state also plays a role in moni- tion of closed user groups. toring the quality of credit information. This If market failures prevent the development is important since lenders, when required to of a transparent credit reporting system alto- share information, may try to retain market gether, the state can play a productive role by power by reporting inaccurate, incomplete, creating incentives for information sharing or lower-quality information (Semenova or—in extreme cases—by requiring banks to 2008). provide credit information to public or pri- vate providers of credit information. The state as a promoter of private There are several cases in which states credit reporting have mandated information sharing among private lenders to overcome monopolies Public and private credit reporting systems and coordination problems in the market ful� ll distinct and at times complementary for credit information. Argentina, as one roles. Aside from acting as an operator and example, managed to extend its credit report- regulator of credit reporting systems, the ing infrastructure in this manner, despite a state can therefore enhance credit market backdrop of high and increasing bank con- ef�ciency by promoting the development of a centration. Because banks were reluctant to private credit reporting infrastructure. share credit information directly, the central In many cases, the type of credit infor- bank made information sharing mandatory mation required by private lenders differs for all loans above 50 pesos, which essen- significantly from that required by central tially meant that the national credit registry, banks and � nancial regulators. Regulators the Central de Riesgo, covered all loans in require information allowing them to moni- the market. To promote the development of tor the loan portfolio of �nancial institutions a private credit reporting infrastructure, the and to estimate associated risks. Such infor- Argentine central bank then made these data mation is often available from public credit available to private credit information pro- registries. Private lenders, by contrast, have viders, which provide client-level risk assess- to assess the creditworthiness of individual ments and other value-added services (see borrowers. This assessment may require Berger and others 2003). more detailed information on indebtedness Elsewhere, state interventions to over- and repayment behavior, including informa- come barriers to competition in credit report- tion on utility payments, debt with credit ing have met with greater challenges. Even card companies and retailers, or individual after a private credit reporting infrastructure credit scores. This generally goes beyond the emerges, the information can be captured information available from national credit by closed user groups. As discussed in box registries and is more readily available from 5.4, this has been a challenge in the case of private credit bureaus that routinely provide Mexico, where credit information is frag- such data through their value-added services. mented among multiple credit bureaus, each This functional differentiation explains why controlled by a distinct subset of lenders that public and private credit information systems cover different segments of the market. Mexi- often coexist, and it makes a case for the role can regulators have made attempts to over- of the state in promoting the development of come these barriers to the emergence of a uni- a private credit reporting infrastructure even versal credit reporting system. These attempts where a credit registry is already in place. have, however, been challenging because the Governments can promote private credit fragmentation of the market for credit infor- reporting by working closely with lenders to mation reflects a similar underlying segmen- help them overcome the coordination fail- tation of the Mexican credit market. ures discussed in the previous section and GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 143 BOX 5.4 Mexico: State Interventions to Prevent Market Fragmentation and Closed User Groups Until the early 1990s, very little credit information During the early 2000s consumer credit from was available and shared in Mexican credit markets. non� nancial institutions grew signi�cantly in Mex- The only information-sharing mechanism available ico and accounted for 40 percent of total credit to at that time was the National Banking Credit Infor- the private sector after 2000. Partly in response to mation Service (Servicio Nacional de Información this development, in 2005 a new credit bureau, Cir- de Crédito Bancario), a public credit registry estab- culo de Crédito, was established, which covered pri- lished by Banco de Mexico in 1964. In the mid- marily retail and mass-market loans. For example, 1990s, several segments of the Mexican credit mar- Circulo de Crédito collects information from major kets were experiencing fast growth, particularly the micro� nance institutions, credit cooperatives, sav- consumer lending and residential mortgages sectors. ings and loans � rms, retailers, grocery stores, and The 1994 Tequila Crisis and ensuing large wave of two banks that specialize in consumer credit to defaults prompted lenders, as well as the Mexican poorer segments of the population. regulators, to pay greater attention to background The Buró de Crédito and the Circulo de Crédito, checks based on credit information to facilitate however, did not exchange information, leading to sound credit decisions. data fragmentation within the credit reporting sys- In reaction to this situation, Mexico’s largest tem, because each bureau � lled a niche in the mar- private credit reporting agency, Buró de Crédito, ket for credit information. In an effort to address was established in 1995, as a collaboration between these issues, Mexican regulators amended the laws TransUnion Mexico (covering consumer credit) and governing credit reporting in 2004, 2008, and more Dun & Bradstreet (covering SME lending). The Buró recently in 2010. One legal change introduced a new de Crédito includes information on loans from the governance structure within the Buró de Crédito, banking sector, and commercial banks have been limiting the extent to which users of the system holding a stake of up to 70 percent in the credit could simultaneously act as owners of credit report- bureau. At present, Buró de Crédito, with 2,800 ing service providers. The reforms also established users and 900 data contributors, contains informa- new rules for the exchange of databases and cre- tion on 52 million individuals and 6 million � rms. ated a mechanism to avoid exclusivity agreements Two other credit bureaus existed in Mexico, but between lenders and providers of credit information. both have ceased operations: Experian failed to These legal changes have facilitated access to obtain suf�cient data from the banking sector, and the Buró de Crédito by smaller banks, an impor- Equifax sold its database to TransUnion so that it tant achievement in Mexico’s concentrated banking was subsumed in Buró de Crédito in 2000. market, where the �ve largest banks hold more than To foster competition in the credit information 80 percent of total assets. The Buró de Crédito and industry, the Mexican authorities passed a new legal the Circulo de Crédito have also started exchanging framework in 2002.a The framework mandated all some information, but this exchange remains lim- regulated institutions to have access to a report as ited because fragmentation of the market for credit part of their underwriting practices, based on the information refl ects a corresponding segmentation data subject’s consent. In addition, it allowed for in the credit market (with banks serving higher- credit reporting service providers to exchange data- income clients and NBFIs catering to lower-income bases between themselves on a for-pro�t basis. households). a. The main components of this legislation were the Law to Regulate Credit Information Societies (Ley para Regular las Sociedades de Información Crediticia), the Operational Rules for the Functioning of Credit Reporting Systems (Reglas a las que Deberán Sujetarán las Operaciones y Actividades de las SIC y Sus Usuarios), and the Law to Regulate Financial Groups (Ley de Grupos Financieros). 144 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 to provide legislation that makes it easy for CONSUMER CREDIT AND lenders to make credit information avail- MICROFINANCE: EXTENDING able to a credit bureau. Several successful THE REACH OF CREDIT country cases illustrate how public-private REPORTING SYSTEMS partnerships can be used to improve a coun- Credit reporting is most effective in support- try’s credit reporting infrastructure. Box 5.5 ing � nancial access, ef�ciency, and stability describes the case of Morocco, where the cen- when participation in the exchange of credit tral bank worked with lenders from different information is as widely shared as possible. market segments to build a comprehensive information-sharing network. Nonetheless, many credit reporting systems An important point, highlighted by the cover only risks in the traditional banking case of Morocco, is that—where public and sector. This limits their effectiveness in sup- private credit reporting institutions coexist— porting market efficiency and creates an the state needs to ensure that public credit important rationale for an active role of the registries and private credit bureaus do not state in promoting the inclusion of nonregu- duplicate services. After a well-functioning lated lenders (such as NBFIs and microfi- credit bureau is in place, the state may, for nance lenders) into existing credit reporting example, scale back the services provided by systems. the credit registry. Over the past decade, many emerging Providing support for the development of markets have witnessed a dramatic growth in a private credit reporting infrastructure can consumer and micro�nance lending. In Indo- also allow central banks to focus on their nesia, for example, consumer lending grew at core competencies of collecting credit infor- an average of 36 percent annually between mation for the purpose of prudential over- 2001 and 2007 (Santoso and Sukada 2009). sight and regulation. In fact, there is no clear In many countries, this expansion in con- economic reason why the state should act as sumer credit has been spearheaded by non- a distributor of comprehensive data on indi- regulated lenders, such as micro�nance insti- vidual borrowers for commercial purposes or tutions and NBFIs. Bringing information on why it should provide value-added services, this rapidly growing market segment into the such as credit scores, if these can be provided fold of the credit reporting infrastructure is by a private entity instead, after market fail- important not only for the risk management ures have been addressed successfully. As the of individual � nancial institutions but also previous section has argued, a more produc- for systemic stability. Many regulators have tive role for the state is to establish the legal begun to recognize this and have been pro- framework that allows for the transparent active in extending the coverage of credit exchange of credit information, to forestall reporting systems to lenders outside the tra- the formation of information monopolies and ditional banking sector. closed user groups, and to provide technical The challenges of extending the reach of assistance for the establishment of private existing credit reporting systems to new bor- credit reporting institutions where such assis- rower groups and lenders outside the tradi- tance is required. tional financial system are twofold: First, To summarize, in addition to its supervi- established lenders often view micro� nance sory and regulatory responsibilities, the most institutions and NBFIs as potential com- adequate role for the state may be to remove petitors and are therefore reluctant to grant legal and institutional obstacles to the devel- them access to existing information-sharing opment of private credit reporting rather arrangements. This gives rise to the well- than to provide comprehensive and value- documented concerns of data fragmentation added credit information services to lenders and closed user groups highlighted in the pre- where such information can be provided by vious section. Second, nonregulated �nancial the private sector. institutions, such as NBFIs and microlenders, GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 145 BOX 5.5 Morocco: Public Support for the Development of a Private Credit Bureau At the onset of 2005, Morocco was characterized So that overlap in the services provided by the by unfavorable credit market conditions a and an public and private credit reporting systems could be inadequate credit reporting system.b The country prevented, lenders could no longer access the public lacked a stable regulatory framework for credit credit registry once the � rst private credit bureau reporting, and there was no cross-sector informa- became operational. (Lenders were, however, still tion sharing by key lenders (that is, banks, micro- required to provide information.) To promote the finance institutions, and nonbank institutions. use of available credit information, the system Responding to these shortcomings, lenders were also required lenders to consult at least one credit considering plans to create separate informational bureau prior to making any credit decision. The databases, one for each sector. This would have led central bank also introduced several innovative to a fragmented, partial credit reporting system, measures to achieve a more effective credit report- which would not allow lenders to check the com- ing infrastructure. Nonregulated lenders would be plete � nancial pro� le of credit applicants. Against able to provide information directly to any private this background, Morocco’s central bank, the Bank credit bureau on consumer consent. Nonregulated of Morocco, decided to take a leadership role in lenders that provided data to the credit bureau were introducing a best practice private credit reporting then given the right to consult the bureau on the system in the country. basis of reciprocity principles. This step ensured Initially, the Bank of Morocco planned to that participation in the private credit reporting upgrade Morocco’s existing public credit regis- system would gradually extend to nonregulated try. However, in a successive stage, the Bank of lenders outside the traditional banking system. Morocco was open to consider other viable credit Finally, to provide the legal framework for the reporting models, including the participation of effective operation of private credit bureaus in the private sector partners. Given the limited capabili- country, the government and central bank also ties of Morocco’s existing public credit reporting undertook some reforms to the legal and regulatory system, and drawing on the results of a market framework governing the exchange of credit infor- assessment study, the International Finance Corpo- mation. The current legislative and operational ration suggested a public-private partnership model framework counteracts the creation of information similar to that operating in several Latin Ameri- monopolies and closed user groups, because all can countries (such as Bolivia, Ecuador, and Peru). existing and future private credit bureaus will be In this model, the central bank � rst upgraded its supplied with the same information from the cen- technological capabilities necessary to receive and tral bank. This, in return, will allow for competi- process credit information (positive and negative) tion on prices and service quality. from the entire universe of regulated lenders. Build- Through this initiative, Morocco’s central bank ing on the trusted leadership of the central bank has established a transparent, competitive, and among � nancial institutions, the public credit reg- advanced credit information sharing infrastructure istry would then consolidate this information and that incorporates private credit bureaus to pro- make it available to any private credit bureau estab- vide value-added services that are more effectively lished and licensed by the central bank. In Septem- supplied by the private sector. At the same time, ber 2007, the central bank issued the � rst private Morocco’s central bank has been building a wealth credit bureau license to Experian Morocco, which of information that assists in its supervisory role as provides both positive and negative credit reporting a regulator of the � nancial system, as well as the information. regulator of private credit reporting institutions. a. Speci�cally, this included limited, collateralized, and selective credit access, elevated rejection rates, and extremely high debt rates. b. Bank Al-Maghrib’s public credit registry (established in 1978) was the only entity providing, albeit limited, credit information. 146 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 may also lack the technological capabilities market into a comprehensive credit informa- necessary to participate in comprehensive tion infrastructure. information sharing. Extending the reach of credit reporting In both instances, the state can support to new client groups can also pose techno- market efficiency through regulation and logical challenges. Participation in advanced capacity-building initiatives that ensure that credit reporting systems can be expensive information on the rapidly growing mar- and technologically complex, requiring lend- ket for micro� nance and consumer credit is ers to adopt new data management and tele- brought into the view of established credit communications solutions. In addition, for reporting systems. There are several models a credit reporting system to function effec- of how this can be achieved. In China, where tively, it must be possible to identify borrow- consumer lending has grown at double-digit ers with reasonable certainty. Many develop- rates since the deregulation of consumer ing economies lack a national identi�cation credit in 1999, the state has taken an active system, particularly for borrowers at the role in the establishment of a comprehensive bottom of the � nancial pyramid. This con- credit reporting infrastructure. The People’s straint has given rise to some innovative solu- Bank of China is currently establishing the tions using, for example, biometric technol- world’s largest credit registry, covering more ogy to uniquely identify borrowers (see, for than 600 million consumers. Information example, Giné, Goldberg, and Yang 2011). sharing is mandatory, and the registry col- Many initiatives in this area are state led, and lects data from regulated as well as nonreg- the state can play an important role in sup- ulated lenders, so that it covers more than porting both the technological and regula- 90 percent of all consumer loans (Jentzsch tory infrastructure required to extend credit 2008). reporting to important new borrower groups, Attaining similarly high coverage rates such as consumer �nance clients and micro�- for mass-market and micro�nance loans has nance borrowers. been more challenging elsewhere, despite regulatory interventions. In Latin America, INTRODUCTION TO PAYMENT the signi�cant presence of micro�nance lend- AND SECURITIES SETTLEMENT ers has required regulators to incorporate SYSTEMS nonregulated lenders into credit reporting systems much earlier than in other world Payment and securities settlement systems are regions. However, both the involvement of a key part of a country’s �nancial infrastruc- the state and the success in creating a level ture. They can also be a major transmission playing �eld for the exchange of credit infor- channel of shocks during times of crisis. The mation have varied. In Guatemala, the state state can reduce potential threats to systemic limited its intervention to the creation of a stability through the regulation of payment legal framework that allowed the country’s and securities settlement systems or through leading micro� nance lenders to form a pri- direct interventions to reduce counterparty vate credit bureau. In Bolivia, by contrast, risk. State agencies and central banks can regulators responded to a large micro�nance also support systemically important par- default crisis in the late 1990s by promoting ticipants of payment and settlement systems the development of credit reporting institu- in times of financial distress. This section tions outside the traditional banking sector. explores the role that the state has played in Despite such reforms, many credit reporting the development and operation of payment systems in the region remain highly frag- and securities settlement infrastructure. mented and characterized by closed user There are two ways in which state inter- groups. As a result, these credit reporting sys- ventions before and during the �nancial crisis tems have often lagged behind in their abil- helped to promote the stability of payment ity to incorporate new segments of the credit and settlement systems. First, large-value GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 147 payment systems survived the crisis relatively through the development of a sound legal unaffected, in large part because of efforts by and regulatory framework. central banks to introduce robust payment systems (such as real-time gross settlement Payment and securities settlement systems that reduce counterparty risk) well systems before the onset of the crisis. Second, once the crisis spread, many central banks acted deci- Payment and securities settlement systems are sively to step up liquidity provision to stabilize the infrastructure that enables the transfer of large-value payment systems and the markets monetary value between parties discharging that rely on this infrastructure, such as the mutual obligations (World Bank 2011b). This interbank money market. Without support infrastructure consists of several components, from central banks, the failure of one or more which include the legal, regulatory, and over- participants to settle their exposures in these sight frameworks for payment transactions, markets could have created credit or liquid- large-value funds transfer systems, retail pay- ity problems for other participants and posed ment systems, foreign exchange settlement a systemic risk to the � nancial system. The systems, and securities settlement systems. timely provision of liquidity by and large miti- Payment systems are a key part of a coun- gated these concerns about counterparty risk. try’s �nancial infrastructure. They can, how- The global � nancial crisis did, nonethe- ever, also be a major transmission channel less, reveal some limitations in existing pay- of shocks during times of financial crisis: large-value payment systems are particularly ment and securities settlement systems. For important for systemic stability because they instance, stress emerged in securities settle- have the potential to generate and transmit ment systems, in particular for over-the- disruptions between the financial and the counter (OTC) derivatives markets. In this real sector of an economy. Since the cur- area, the crisis highlighted limitations in the rent report focuses on �nancial stability and transparency and legal framework govern- the role of the state, this chapter focuses on ing securities settlement systems.6 Building large-value payment systems and the steps a stronger infrastructure for securities and that central banks have taken to strengthen derivatives settlement systems to minimize these systems. The chapter also touches on counterparty risk is an important concern for securities settlement systems. In particular, it many emerging markets, where the develop- reviews the performance of OTC derivatives ment of robust securities settlement systems settlement systems, which is one of the areas has sometimes lagged behind the rapid devel- in which the crisis highlighted the need for opment of equity and derivatives markets. proactive oversight and development support. The remainder of this section reviews the Robust payment and security settlement role that state agencies and central banks systems promote economic activity by con- have played in supporting payment and set- trolling the counterparty risk inherent in the tlement systems through the � nancial crisis. transfer of high-value funds and by helping It also highlights some of the challenges to with the implementation of monetary policy. the development and operation of payment Payment systems are essential for � nancial and securities settlement systems that have sector development because they contribute become apparent during the crisis. The chap- to the innovation and development of new ter argues for an active role of the state in � nancial products and facilitate functioning the development, regulation, and oversight of � nancial markets (List�eld and Montes- of payment and securities settlement sys- Negret 1994). The smooth functioning of tems, particularly in (a) mitigating counter- payment systems can mitigate �nancial crises party risk through interventions in interbank by reducing or eliminating counterparty risk markets and (b) mitigating counterparty and related to �nancial market transactions and is settlement risks in securities transactions therefore vital for ensuring �nancial stability 148 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 (Afonso and Shin 2009; Cirasino and García Large-value payment systems are impor- 2009; Flannery 1996). Finally, improvements tant for systemic stability because they are in payment infrastructure can result in signif- able to generate and transmit disturbances icant ef�ciency gains, reductions in transac- of a systemic nature to the � nancial sector. tion costs, and increased economies of scale Bank-to-bank transactions in the interbank in � nancial intermediation (Hasan, Schmie- money market, in particular, depend strongly del, and Song 2009; Humphrey and others on large-value payment systems in two ways. 2006; Lindquist 2002). Greater efficiency First, the interbank money market, as the dis- of interbank payment systems can therefore tributor of liquidity throughout the �nancial have a wider positive impact on credit cre- sector, relies heavily on payment systems to ation and �nancial development (Merrouche transmit funds across the financial system and Nier 2010). rapidly and safely. Second, the interbank The World Bank, through its Payment money market is a key source of liquidity Systems Development Group, has been pay- for the operation of payment and settlement ing close attention to payment and securities systems. Because of their central role in allo- settlement systems as a key component of cating liquidity, ef�cient large-value payment the �nancial infrastructure of a country. For systems are also indispensable for the swift example, country assessments done by the implementation of monetary policy. Because World Bank and the IMF under the Finan- of the signi�cant amount of funds channeled cial Sector Assessment Program (FSAP) fre- through the system, large-value payment sys- quently include a component that assesses tems can also transmit and, where inef�cien- the design, safety, and ef�ciency of payment cies exist, amplify disturbances in the �nan- and securities settlement systems.7 Some cial sector. concrete recommendations on payment sys- The turnover of large-value payment sys- tems that have come out of recent FSAPs tems, as measured by the number of times include (a) strengthening the overall legal an amount equivalent to the value of the framework for payment system oversight to GDP in each country is settled in a year, is ensure safety and soundness of the system, typically several times a country’s GDP. Fig- (b) linking large-value payment systems with ure 5.5 shows that the size of these flows, securities depositories to address potential however, varies quite substantially across settlement risks for securities transactions, regions. It tends to be higher in countries and (c) ensuring business continuity of pay- with active securities markets that are sup- ment systems by having backup servers to ported by large-value payment systems, with avoid any loss of transaction data. The World several national payment systems handling Bank also actively works with international transactions more than 30 to 40 times their standard-setting bodies to establish the set of country’s GDP. These high turnover rates standards and best practices.8 underscore the importance of well-function- ing payments infrastructure for financial stability more broadly.9 Large-value payment systems Large-value payment systems are the � nan- The role of central banks in large-value cial infrastructure used to process time- payment systems sensitive high-value payments. Banks and other � nancial institutions use these systems Central banks have historically performed to transfer funds to each other. They are the function of payment intermediaries and also the platform on which the interbank remain centrally important for the ef�cient money market operates. As such, stable and operation of payment systems (Johnson resilient large-value payment systems are an and Steigerwald 2008; Millard and Saporta essential prerequisite for the ef�cient opera- 2005). In particular, central banks pro- tion of a � nancial market. vide banks with the physical infrastructure GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 149 FIGURE 5.5 GDP Turnover of Large-Value Payment Systems by Region, 2009 Turnover (as ratio to GDP) 45 41.7 40 35 30 25 20 19.6 19.6 15 12.6 10.7 10 5.6 5 0 East Asia Europe and Latin America Middle East and South Asia Sub-Saharan and Pacific Central Asia and the Caribbean North Africa Africa Sources: World Bank 2011b and Global Payment Systems Survey 2010. Note: Weighted averages. needed to make interbank payments, as well cooperation and coordination among pay- as overdraft and deposit facilities (central ment system participants. bank money10). Central banks have also been actively involved in promoting safety and ef�- Instead of discussing each role separately, ciency of these systems to maintain �nancial this chapter focuses on the recent contribu- stability (Bech, Preisig, and Soramäki 2008). tions central banks have made to make large- The involvement of the central bank in pay- value payment infrastructure more resilient ment systems can take several non–mutually and on the role payment systems can play in exclusive forms: possible future crises. • As an operator of payment systems, the Evolution of real-time gross settlement central bank acts as the settlement and systems and the role of central banks sometimes clearing agent in payment sys- tems. In this role, the central bank also Around the world, large-value payment sys- at times provides the technological infra- tems have rapidly evolved over the past two structure for settlement systems, as well decades. Two key risks faced by such systems as intraday and overnight credit (lender are liquidity and credit risk, and most coun- of last resort), to preserve liquidity in the tries have upgraded existing large-value pay- system. ment systems to more effectively address these • As a regulator of payment systems, the risks.11 Over the past two decades, countries central bank enforces regulatory rules and have largely moved away from deferred net monitors safety and ef�ciency of transac- settlement (DNS) systems to real-time gross tions taking place through the system. It settlement (RTGS) systems. The key advan- also safeguards the legal rules governing tage of RTGS is that each payment is settled interactions in interbank money markets immediately (in real time) and processed and other transactions using the large- individually (on a gross basis). On the other value payment system. hand, in a DNS system, processing of funds • As a promoter of the development, it acts is delayed (usually at the end of the business as a catalyst of innovation and supports day) and is done on net basis (see Martin 150 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 5.6 The Adoption of Real-Time Gross Settlement Systems countries have followed suit and have shifted over Time, 1990–2010 to real-time gross settlement systems in the past decade. More recent adopters include, for example, El Salvador and the West Bank Number of countries and Gaza, where RTGS systems did not 120 become operational until 2010, and Ethiopia 100 and Rwanda, which established RTGS sys- tems in 2011 (World Bank 2011b).12 Accord- 80 ing to the World Bank’s 2010 Global Payment 60 Systems Survey, 116 out of 139 countries surveyed (83 percent) have RTGS systems 40 (World Bank 2011b). In 112 of these coun- tries, central banks act as the operator of the 20 system (private entities operate the system in 0 the remaining four countries). In virtually all 1990 1995 2000 2005 2010 of those countries (115), the central bank also acts as settlement agent for the system. The Sources: World Bank 2011b, Global Payment Systems Survey 2010, and Bech and Hobijn 2007. vast majority of RTGS participants are com- mercial banks: the payment systems survey shows that 95 percent of RTGS systems give 2005 for more detail). The main drawback access to only commercial banks, 60 percent of conventional DNS systems is that they are give access to noncommercial banks, and 69 vulnerable to counterparty credit risk: a pay- percent give access to NBFIs. ment system participant can become insol- In addition to managing liquidity risk, cen- vent between the time a payment is made tral banks and regulators have also played an and the time it is settled. The credit risk in important role in reforming large-value pay- deferred settlement systems is thus generally ment systems to mitigate systemic risk. Before borne by market participants. we turn to the role of central banks in man- Real-time gross settlement systems elimi- aging liquidity risk, box 5.6 reviews coun- nate this risk, because all transactions are try cases highlighting a number of reforms settled instantaneously on an individual designed to make large-value payment systems basis. However, the central bank now has a more resilient to systemic �nancial risks. 13 greater responsibility in providing liquidity in the system. The flow of funds provided by THE STATE’S ROLE IN PAYMENT central banks consists of both collateralized AND SECURITIES SETTLEMENT and uncollateralized current account over- SYSTEMS drafts and central bank credit, either in the form of a loan or a repo. As the next section At the peak of the �nancial crisis, large-value shows, the shift toward real-time gross settle- payment systems showed some signs of dis- ment systems has increased the responsibili- tress but generally proved to be resilient.14 ties of central banks in providing liquidity This resilience was mainly due to the swift for better management of counterparty risks, and decisive action of central banks pro- particularly in times of crisis, when credit viding liquidity to the system. This section lines between banks may dry up. looks at the role of central banks in distrib- Figure 5.6 traces the rapid adoption of uting liquidity in the system to address ris- real-time settlement systems around the ing concerns about counterparty risk during world over the time period between 1990 and the recent financial crisis. In particular, it 2010. Developed economies were the early considers the role of central banks in stabiliz- adopters; for example, most European coun- ing interbank money markets and highlights tries adopted RTGS systems by the late 1990s some new measures that have been proposed (Bech and Hobijn 2007). Many developing to improve the management of liquidity risk. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 151 BOX 5.6 Reforming Large-Value Payment Systems to Mitigate Systemic Risk PAKISTAN mitigate liquidity risk. In addition, PRISM generates and sends participants liquidity alerts, followed up The national payment system (NPS) in Pakistan has by phone calls to the banks to arrange the required undergone major reforms during the past �ve years funds from the interbank market. In the second year under the strong leadership of the State Bank of Pak- of PRISM, the SBP introduced intraday liquidity istan (SBP). This has been a complex project com- facilities, a credit from the SBP to participants in prising the modernization of the infrastructure for PRISM for the settlement of payments. It is available wholesale and retail payments, the modernization of only on a fully collateralized basis against approved the legal and regulatory framework for payment sys- government securities as speci�ed by the SBP. tems and instruments, and the introduction of inno- As an operator of the large-value system, the SBP, vative solutions in the provision of payment services during the reform process, paid increasing atten- for the underserved banking population. tion to the operational reliability and security of This box � rst describes the new features the SBP PRISM, � xing the security policies and operational introduced in the large-value payment system as a procedures in a number of normative documents and result of the reform and then highlights various roles instructions. Business continuity–related activities the SBP performs to ensure the smooth and secure for PRISM have been established and documented functioning of the country’s national payment as well. Procedures are in place for periodic back- system. ing-up and storing of data. A secondary processing site has been established, and full system recovery is Launch of PRISM expected in 30 minutes to four hours with no data In July 2008, the SBP launched the Pakistan Real loss. In addition, the SBP has developed require- Time Interbank Settlement Mechanism (PRISM) to ments and recommendations for system participants better cope with systemic risk associated with large- to have in place the necessary business resumption value payment systems. PRISM is a real-time gross and recovery tools. settlement (RTGS) system. Before the introduction of PRISM, � nancial market transactions and large- Various roles of the SBP value government payments used to be processed Clearly, the SBP played a vital role in reforming the through checks and settled on a multilateral net national payment system and continues to play vari- basis once a day. In such a system, if multilateral ous roles to support the system’s functioning. First, credit exposures are not managed properly, the sys- the SBP is the operator and owner of the RTGS sys- tem’s participants can be exposed to considerable tem. It also operates the securities settlement system credit risk. and the central depository for government securities The shift from end-of-day to intraday settlement (which is part of PRISM). Second, the SBP provides and the gradual introduction of time-critical pay- safe settlement assets for all interbank payments set- ments has considerably increased liquidity needs in tled through the RTGS system by maintaining and PRISM. Opening balances and funds received from managing settlement accounts for all participants in other participants during the operational day are the PRISM. main sources of liquidity for participants in PRISM. Third, the SBP is responsible for the regulation In addition, the participants can use the cash reserve and oversight of all recognized payment systems and requirements and statutory liquidity requirements payment instruments in the country. As part of its they hold in accounts with the SBP to make pay- oversight role, the SBP has been working continu- ments during the day in the RTGS system. These ously to ensure that the overseen systems comply funds are available free of charge. with international standards and best practices, such Several mechanisms have been put in place to as the Committee on Payment and Settlement Sys- better manage liquidity risks in PRISM. The sys- tems’ “Core Principles for Systemically Important tem offers centralized queuing whereby payments Payment Systems� and the Committee on Payment that cannot be settled immediately because of lack and Settlement Systems, and International Organi- of suf�cient funds in the settlement accounts of par- zation of Securities Commissions’ “Recommenda- ticipants are placed in a central queue. A gridlock tions for Securities Settlement Systems.� In addition, resolution mechanism is another tool used in the sys- the SBP supported the government in establishing tem to facilitate timely settlement of payments and the Pakistan Remittance Initiative, and played a (Box continues next page) 152 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 BOX 5.6 Reforming Large-Value Payment Systems to Mitigate Systemic Risk (continued) catalyst role by encouraging the commercial banks gross basis, that is, according to model 1 of deliv- to leverage the PRISM system to support faster pro- ery versus payment, with the STR processing the cessing of international remittances. cash leg of the securities transactions. As de� ned Fourth, the SBP has been playing a leading role in by the Committee on Payment and Settlement Sys- establishing a sound legal and regulatory framework tems, delivery versus payment is a link between a for payments to reduce legal risk. The SBP had initi- securities transfer system and a funds transfer sys- ated the drafting of the Payment System and Elec- tem that ensures that delivery occurs if, and only if, tronic Fund Transfer Act (PSEFT Act), which was payment occurs. Currently, delivery versus payment introduced in 2007 and deals with a broad range of is observed in all securities settlement systems,a and risk-related issues such as irrevocability of payments almost all securities are dematerialized. and settlement � nality, validity and enforceability of The Brazilian payment system reform, however, netting arrangements, and � nality of settlement of went beyond the launch of the STR and the SELIC’s government securities. The act empowers the SBP to modus operandi changes. To reduce the systemic risk issue rules, guidance, circulars, bylaws, standards, and vulnerability to shocks, changes to the legal and or directions with respect to such systems or instru- regulatory framework were also necessary, in terms ments in pursuing its objectives to promote mone- of clarifying the rights and obligations of partici- tary stability and a sound � nancial structure. pants in payment transactions. Prior to the reform, Brazil was lacking legal validation of multilateral BRAZIL netting, protection of assets pledged as collateral in In 2002, the Banco Central do Brasil (BCB) launched case of failure of a participant, and a sound legal the new Brazilian Payment System (Sistema de Paga- basis for the central bank’s oversight function. mento). Brazil had a fairly sophisticated payment In this complex process of reform, the BCB system even prior to the launching of the reform pro- assumed a leading role. In particular, the reform gram. However, to better manage systemic risk, the of the Brazilian payment system had three impor- Reserves Transfer System (Sistema de Transferência tant outcomes: (a) the reduction of systemic risk in de Reservas —STR) was launched. With this system, the settlement of � nancial transactions; (b) a more interbank fund transfers can be settled irrevocably appropriate sharing of the risks associated with set- and unconditionally, that is, with � nality, on a real- tlement of payment transactions between the central time basis. This allows for settlement risk reduc- bank and private market players; and (c) the compli- tion for interbank transactions and, consequently, ance of the systemically important payment systems systemic risk reduction. As of 2010, the STR had of the country with international standards and best 151 participants, including the BCB, the National practices. Treasury Secretariat, three clearinghouses, and 137 In particular, two elements of the Brazilian banks. In 2010, the STR processed 12.7 million reform process have been notable, contributing to transactions for a total value of R$132,318.9 billion. its breadth, scope, and complexity. First, the central With the launch of the new central bank pay- bank conducted a comprehensive diagnostic study ment system, a direct link was established between before de� ning the reform, which sought to identify the STR and the central depository for federal pub- all forms of risks present in the system. Second, the lic securities (Sistema Especial de Liquidação e de central bank consistently involved key stakeholders Custódia —SELIC). The link made it possible for (banks, other � nancial institutions, clearinghouses, SELIC to settle all transactions in real time on a other regulators, and so forth) in the reform debate. a. Federal government bonds are traded by telephone (in the traditional OTC market) or on a BM&FBOVESPA-operated electronic trading platform (SISBEX). In this market, repurchase agreements predominate over outright transactions. Traditional OTC is also the main trading method for corporate bonds, state government bonds, non-standard derivatives and most securities relating to the National Treasury’s special responsibilities. Some of the National Treasury’s securities can also be traded at organized OTC markets operated by CETIP and BM&FBOVESPA. Stocks, standardized deriva- tives, and commodities are traded at BM&FBOVESPA, the only Brazilian stock and derivatives exchange. Two electronic trading platforms are used: MEGABOLSA for equities and equity derivatives, and Global Trading System (GTS) for com- modities and other derivatives. GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 153 The section then turns to the performance FIGURE 5.7 Sources of Intraday Liquidity for Participants of of securities settlement systems, an area in Real-Time Gross Settlement Systems which the crisis highlighted a number of challenges and possibilities for reform and Percent improvement. 100 Management of liquidity risk15 80 In most countries, the primary source of 60 intraday liquidity for financial institutions having access to real-time gross settlement systems is funds received from other partici- 40 pating institutions (see figure 5.7).16 How- ever, during times of stress it can become 20 dif�cult for banks to obtain suf�cient funds from banks holding excess reserves. As mar- 0 ket participants become more cautious, they Other RTGS operator Reserve Lines of credit participants (current account requirements between banks tend to hold onto liquid assets or target them overdraft, loan, balance toward investments perceived to be less risky, or repo) rather than extending loans in the interbank market.17 Signi�cant shocks to the willing- Sources: World Bank 2011b and Global Payment Systems Survey 2010. ness of � nancial institutions to lend to each other—for example, in the form of a default or bank run—can result in a liquidity crunch have a stronger effect on the uncollateralized in interbank markets that is potentially or unsecured segment of the money market. severe enough to trigger a payment system In secured (or repo) transactions, such con- crisis. The collapse of Lehman Brothers in cerns are mitigated to some extent by the the United States is a prominent example of a presence of collateral.20 Therefore, an impor- recent event that triggered a liquidity crisis in tant policy measure that has been proposed interbank markets.18 to stabilize interbank markets—especially To revive interbank markets and to stop in emerging economies—is to collateralize a contagion spreading through the payment system,19 central banks worldwide actively transactions in the interbank market. provided liquidity to interbank markets, for To illustrate the rationale behind this example, by extending access to their liquid- proposition, figures 5.8 and 5.9 compare ity facilities to a wider set of payment system the volatility of interbank money market participants, by widening the range of col- rates around the collapse of Lehman Broth- lateral accepted in their operations (both the ers in economies with secured and unse- Federal Reserve and the Bank of England cured interbank money markets. Figure extended their list of eligible collateral), and 5.8 shows the high volatility of interbank by lending at longer maturities (the European money market rates in the United States and Central Bank used the open market opera- the United Kingdom, two of the economies tions, and the Federal Reserve used the term most severely affected by the � nancial crisis auction facility). These proactive interven- that use an unsecured interbank market. Fig- tions in interbank markets were by and large ure 5.9 compares the volatility of interbank successful, and they highlighted the impor- money market rates in the Russian Federa- tant, and possibly increased, role of central tion and Ukraine, 21 where interbank money banks in liquidity management. markets are largely unsecured, and Brazil Liquidity crises ensuing from deterioration and Mexico, where for historical reasons the in the perceived creditworthiness of coun- interbank money market operates mainly on terparties in the interbank market generally a secured basis. Although these economies 154 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 FIGURE 5.8 Interbank Money Market Rates in the United States takes advantage of existing infrastructure, and United Kingdom and proved to be successful in reviving Ita- ly’s interbank market. The case of Italy is reviewed in box 5.7. a. Federal funds rate, United States Percent Although collateralized interbank markets 3.5 can improve � nancial stability, introducing 3.0 collateralized transactions also comes with a number of legal and technological require- 2.5 ments, which can be challenging to achieve, 2.0 particularly in low- and middle-income 1.5 countries. In terms of legal requirements, the 1.0 introduction of collateralized interbank mar- ket necessitates (a) a sound legal framework 0.5 that ensures � nality of funds and securities 0 transfers; (b) the legal protection to pledged 8 8 8 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 10 8 10 8 10 8 08 collateral from third-party claims; (c) the /0 /0 /0 /0 /0 0 0 /0 /0 /0 /0 /0 /0 /0 /0 2/ 6/ 2/ 13 17 21 25 29 10 14 18 22 26 30 /4 /8 /1 8/ 8/ 8/ 8/ 8/ possibility of seizing the pledged collateral ef�ciently in case of a default by the debtor; and (d) reliable, universally accepted, and b. LIBOR (one week), United Kingdom enforceable accounting standards for valuing Percent collateral. In terms of technology, a collat- 6.1 eralized interbank market relies on the exis- 5.9 tence of a large-value payment system that is 5.7 fully integrated with an electronic book-entry 5.5 system that enables the recording of pledges 5.3 on securities and changes in their ownership. 5.1 Integrating payment systems and collat- 4.9 eral registries and securities depositories can 4.7 also make sense in view of some of the steps 4.5 central banks took to stabilize the interbank 8 8 8 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 10 8 10 8 10 8 08 markets during the recent financial crisis. /0 /0 /0 /0 /0 0 0 /0 /0 /0 /0 /0 /0 /0 /0 2/ 6/ 2/ 13 17 21 25 29 10 14 18 22 26 30 /4 /8 /1 8/ 8/ 8/ 8/ 8/ For example, central banks began to accept a broader range of collateral in all classes of Sources: Federal Reserve Bank of New York and Bank of England. lending operations (accepting private sec- Note: U.S. and U.K. interbank money market rates are shown for the periods August 13–Septem- ber 12, 2008 (before failure of Lehman Brothers), and September 15–October 14, 2008 (increased tor collateral and allowing counterparties to volatility). LIBOR = London interbank offered rate economize on the use of government securi- ties, which is often the only collateral that counterparties can still use in secured, or were arguably affected by different shocks, repo, markets). Expanding the pool of eligible the stylized evidence suggests that collateral- collateral would generally mean that central izing transactions in interbank markets can banks are willing to accept securities other reduce volatility during times of financial than those issued by the government or by crisis. itself as collateral in lending operations. In The potential bene�ts of encouraging col- practice, this can be done quickly, safely, and lateralization in interbank markets can be ef�ciently only if the funds transfer system is illustrated by some recent country experi- interconnected with the securities depository ences. In Italy, for example, in an effort to that holds such other, newly eligible securities. help restart activity in the interbank market Although many RTGS systems have some during the peak of the crisis, the Bank of interface with securities depositories and the Italy introduced an innovative collateral- related settlement systems, in many cases ized interbank market (MIC) segment that (especially in the case of private securities GLOBAL FINANCIAL DEVELOPMENT REPORT 2013 T H E R O L E O F T H E S TAT E I N F I N A N C I A L I N F R A S T R U C T U R E 155 depositories) this interface is available solely FIGURE 5.9 Interbank Money Market Rates in Emerging Markets for the final settlement of securities trades and is not operational for any other purpose, Percent a. Interest rates on interbank market, Ukraine such as for collateralization in interbank 22 markets (World Bank 2011b). Also, typically, 20 though not in all cases, when an RTGS sys- 18 16 tem is interconnected with a securities depos- 14 itory, the latter holds the records for govern- 12 10 ment and central bank securities only. This 8 type of securities depository is usually owned 6 and operated by the central bank, just as in 4 8/ 8 8/ 8 8/ 8 8/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 10 8 10 8 10 8 08 the case of RTGS systems. Central banks can /0 /0 /0 /0 /0 0 0 /0 /0 /0 /0 /0 /0 /0 /0 2/ 6/ 2/ 13 17 21 25 29 10 14 18 22 26 30 /4 /8 /1 8/ therefore support better integration of funds transfer systems with securities settlement systems in their country. Percent b. MIBOR (one day), Russian Federation 12 Securities settlement systems 10 Financial transactions can take place either 8 through organized exchanges or as “over- the-counter� (OTC) transactions. Exchanges, 6 such as the stock market, offer a centralized 4 way of transacting, where one party facili- 8 8 8 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 9/ 8 10 8 10 8 10 8 08 /0 /0 /0 /0 /0 0 0 /0 /0 /0 /0 /0 /0 /0 /0 2/ 6/ 2/ 13 17 21 25 29 10 14 18 22 26 30 /4 /8 /1 tates transactions by connecting buyers and 8/ 8/ 8/ 8/ 8/ sellers. They also offer greater regulatory oversight and transparency, since only mem- Percent c. TIIE (28 days), Mexico bers of the exchange can trade the products 8.80 that are listed on the exchange. There is no counterparty risk in transactions settled 8.70 through exchanges, because the exchange 8.60 acts as the regulator and the counterparty to 8.50 each transaction. By contrast, OTC markets are generally decentralized, with numerous 8.40 mediators trying to connect buyers and sell- 08 08 08 08 9/ 8 9/ 8