WORLD BANK LATIN AMERICAN 65744 AND CARIBBEAN STUDIES Financial Development in Latin America and the Caribbean THE ROAD AHEAD Augusto de la Torre Alain Ize Sergio L. Schmukler FINANCIAL DEVELOPMENT IN LATIN AMERICA AND THE CARIBBEAN i Financial Development in Latin America and the Caribbean FINANCIAL DEVELOPMENT IN LATIN AMERICA AND THE CARIBBEAN iii Financial Development in Latin America and the Caribbean The Road Ahead By a team led by Augusto de la Torre, Alain Ize, and Sergio L. Schmukler © 2012 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org 1 2 3 4 15 14 13 12 This volume is a product of the staff of The World Bank with external contributions. The � ndings, interpretations, and conclusions expressed in this volume 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 judg- ment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to the work is given. For permission to reproduce any part of this work for commercial purposes, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-8847-1 eISBN: 978-0-8213-8936-2 DOI: 10.1596/978-0-8213-8847-1 Library of Congress Cataloging-in-Publication data have been requested. Cover design: Naylor Design, Inc. Cover photo: SVLuma and Ash Design/Shutterstock.com Contents Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv 1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 The conventional view on � nancial development and its limitations . . . . . . . . . . . . . . . . . . 2 Revisiting � nancial development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Where is LAC? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Promoting the bright side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Dealing with the dark side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2 Financial Development: Bright Side, Patterns, Paths, and Dark Side . . . . . . . . . . . . . . . . . 15 The bright side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 The patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 The paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 The dark side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Can there be too much � nance? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3 Domestic Financial Development: Where Does LAC Stand? . . . . . . . . . . . . . . . . . . . . . . . 29 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Main � ndings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Banking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Bond markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Equity markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 The new players . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 v vi CONTENTS Alternative markets and products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4 Financial Globalization: Where Does LAC Stand? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 The two dimensions of � nancial globalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Financial diversi�cation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Financial offshoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 5 Financial Inclusion: Where Does LAC Stand? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Supply-side evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Demand-side evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 The role of government . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 6 The Banking Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Is the banking gap real? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Supply or demand? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Are these the ghosts of a turbulent past? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Where is LAC now? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 7 The Equity Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Effects of globalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Free float . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Market concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Institutional investors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Corporate governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Other factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 8 Going Long . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 What are the issues? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 What are the options? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Where is LAC? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 What are the key policy challenges? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 9 Risk Bearing by the State: A Collective Action Perspective . . . . . . . . . . . . . . . . . . . . . . . 141 The role of the state in the pure agency paradigms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Adding collective action frictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Adding risk aversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 The question of private or state guarantees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Systemic risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 LAC’s policy swings—a play in four acts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Toward a rebalanced policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 10 Prudential Oversight: Where Does LAC Stand? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 The progress and the remaining gaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 The systemic oversight challenges ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 CONTENTS vii Annex 10.A Methodology for the econometric analysis of BCP Ratings . . . . . . . . . . . . . 171 Annex 10.B Highlights of the 2007 World Bank survey of supervisory practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 Annex 10.C Joint World Bank–ASBA survey on systemic oversight . . . . . . . . . . . . . . . . . 175 Annex 10.D Financial concentration in LAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 11 Macroprudential Policies over the Cycle in LAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Macroprudential policy and the dynamics of the dark side . . . . . . . . . . . . . . . . . . . . . . . 182 LAC’s cycles and vulnerabilities: Lessons from the past? . . . . . . . . . . . . . . . . . . . . . . . . . 184 Some key macroprudential policy design issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 The path ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 12 Microsystemic Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 The rationale for prudential regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 The outer boundaries: Illuminating the shadows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 The inner boundaries: Silos versus universal licensing . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 The SIFI problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 Systemic liquidity: Norms and access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 Financial innovation: Did LAC have it right? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 The regulatory agenda ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 Annex 12.A Challenges posed by large, complex � nancial conglomerates: The case of CL Financial in the Caribbean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 13 Systemic Supervision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 The new connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 A new approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 New tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 Getting there . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 The agenda ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 14 Summary of Policy Directions for the Road Ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 Where is LAC? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 The tough issues to think about . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 Policy directions: The bright side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 Policy directions: The dark side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Boxes 3.1 The bond market in Chile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3.2 The equity market in Brazil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 4.1 Costs and bene�ts of the participation of foreign banks in domestic � nancial systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 6.1 A brief history of housing � nance in LAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 7.1 Corporate governance and equity market development . . . . . . . . . . . . . . . . . . . . . . . 115 8.1 Covered bonds versus mortgage-backed securities: LAC’s recent experience . . . . . . . 131 8.2 LAC’s indexation experiences in housing � nance. . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 9.1 Welfare criteria in the theoretical literature on state guarantees . . . . . . . . . . . . . . . . 142 viii CONTENTS 12.1 The Mexican Sofoles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 12.2 Reforming the regulatory perimeter: United States versus the European Union. . . . . 207 Figures 2.1 Frictions, paradigms, and failures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.2 Appearance, convexity, and returns to scale of FD indicator paths . . . . . . . . . . . . . . . 20 2.3 Paths for government debt: External and domestic . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.4 Banking indicators’ paths: Retail and wholesale funding and private credit . . . . . . . . 21 2.5 Financial depth indicators: Dynamic and cross-section development paths . . . . . . . . . 23 3.1 Domestic � nancial systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.2 Banking indicators relative to global benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 3.3 Bank credit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 3.4 Foreign ownership and concentration of banking systems . . . . . . . . . . . . . . . . . . . . . . 36 3.5 Total banking assets as a percentage of GDP, within LAC. . . . . . . . . . . . . . . . . . . . . . 37 3.6 LAC-7 � nancial indicators against benchmark . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3.7 Primary bond markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.8 Domestic bond market turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 3.9 Average maturity of bonds at issuance in local markets . . . . . . . . . . . . . . . . . . . . . . . . 42 3.10 Currency composition of bonds at issuance in local markets . . . . . . . . . . . . . . . . . . . . 43 3.11 Equity market size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.12 Equity markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.13 Stock market turnover relative to global benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . 46 3.14 Concentration in equity markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.15 Within LAC: Trading activity and number of listed � rms in domestic equity markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.16 Equity market capitalization as a percentage of GDP, within LAC . . . . . . . . . . . . . . . 49 3.17 Pension fund, mutual fund, and insurance company assets . . . . . . . . . . . . . . . . . . . . . 51 3.18 Composition of pension fund portfolios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 3.19 Mutual fund portfolio holdings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 3.20 Derivative and factoring markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 3.21 Financial cooperatives, credit unions, and exchange-traded funds . . . . . . . . . . . . . . . 55 4.1 Foreign assets and liabilities and gross capital flows . . . . . . . . . . . . . . . . . . . . . . . . . . 61 4.2 Within LAC: Foreign assets and liabilities and gross capital flows. . . . . . . . . . . . . . . . 62 4.3 Valuation effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 4.4 De jure and de facto � nancial globalization measures . . . . . . . . . . . . . . . . . . . . . . . . . 64 4.5 Net foreign assets as percentage of GDP: Equity and bonds . . . . . . . . . . . . . . . . . . . . 66 4.6 Within LAC: Net foreign assets as percentage of GDP: Equity and bonds . . . . . . . . . . 67 4.7 New capital-raising issues in foreign markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 4.8 Relative size of foreign capital markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 4.9 Average maturity of bonds at issuance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 4.10 Within LAC: Average maturity of public and private bonds in foreign markets at issuance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 4.11 Ratio of foreign currency bonds to total bonds at issuance . . . . . . . . . . . . . . . . . . . . . 72 4.12 Equity trading in domestic and foreign markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.13 Concentration in foreign bond and equity markets . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.14 Foreign-owned bank assets as a percentage of total banks assets . . . . . . . . . . . . . . . . 75 5.1 Number of branches, ATMs, and deposits and loan accounts . . . . . . . . . . . . . . . . . . . 79 5.2 Actual versus predicted number of deposits per 1,000 adults . . . . . . . . . . . . . . . . . . . 82 5.3 Actual versus predicted number of loans per 1,000 adults . . . . . . . . . . . . . . . . . . . . . . 82 5.4 Annual fees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 CONTENTS ix 5.5 Firms’ use of bank accounts and credit products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5.6 Fixed assets and working capital � nanced by banks . . . . . . . . . . . . . . . . . . . . . . . . . . 86 5.7 Percentage of � rms that consider access to � nance as a severe obstacle . . . . . . . . . . . . 87 5.8 Governments’ commitment to � nancial inclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 5.9 Governments’ commitment to � nancial inclusion: LAC-7 versus other LAC comparators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 5.10 The adoption of correspondent banking and mobile branches . . . . . . . . . . . . . . . . . . . 91 5.11 The adoption of correspondent banking and mobile branches: LAC-7 versus rest of LAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 5.12 Index of credit information and legal rights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 6.1 Offshore versus onshore credit to the private sector, LAC-4 . . . . . . . . . . . . . . . . . . . . 97 6.2 Evolution of private credit by type of credit across LAC-7 countries . . . . . . . . . . . . . . 98 6.3 Real lending rate, real deposit rate, and EMBI: differentials between LAC-6 and the United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 6.4 Net interest margins: Contribution of different factors in explaining differences between Latin America’s average and that for developing countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 6.5 Real credit to the private sector and the compounded real deposit rate index, LAC-6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 7.1 Percentage of LACs that implemented capital market reforms, various years . . . . . . 108 7.2 Onshore and offshore equity markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 7.3 Average of residuals of domestic and foreign turnover, 2005–09. . . . . . . . . . . . . . . . 110 7.4 Trading activity in domestic equity markets during the 2000s . . . . . . . . . . . . . . . . . 112 7.5 Domestic turnover and concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 7.6 Domestic turnover and institutional investors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 7.7 Domestic turnover and corporate governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 10.1 BCP assessments, LAC-6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 10.2 BCP assessments, Caribbean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 10.3 BCP assessments, rest of LAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 10.4 Compliance to BCP principles by LAC countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 10.5 Financial regulation and supervision progress in nine LAC countries . . . . . . . . . . . . 165 10.6 Complexity index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 10.7 Selected soundness indicators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 11.1 Unconditional probability of booms and crises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 11.2 Synchronization between real output and � nancial cycles . . . . . . . . . . . . . . . . . . . . . 187 11.3 Behavior of credit during downturns in real economic activity: Are � nancial cycles leading real cycles? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 11.4 Unconditional and conditional probability of lending booms . . . . . . . . . . . . . . . . . . 189 11.5 Real exchange rate behavior during downturns in real credit . . . . . . . . . . . . . . . . . . 190 11.6 Foreign capital and credit downturns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 11.7 Main features of the credit cycle over time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 11.8 Reserve requirements and reference rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 12.1 Average number of banks with more than 10 percent of total banking assets, 2006–09 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 12.A.1 CL Financial annual report, 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 Tables 2.1 A simple typology of paradigms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.1 Countries analyzed, by region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 x CONTENTS 3.2 Benchmark model for LAC’s � nancial development indicators . . . . . . . . . . . . . . . . . . 34 5.1 Regressions of indicators of � nancial inclusion on income, population density, and country-group dummies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 5.2 Regressions for deposit and loan fees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 6.1 LAC-7 credit gap by type of credit, 1996 and 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 6.2 Real lending rates by type of � rm, 2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 6.3 LAC bank net interest margins, bank overheads, and private credit . . . . . . . . . . . . . 101 6.4 LAC-7 private credit gap: A decomposition by source . . . . . . . . . . . . . . . . . . . . . . . . 102 6.5 Number of crises by type, 1970–2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 6.6 LAC banks’ interest margins, � nancial soundness, enabling environment indicators, and credit history—growth and crashes . . . . . . . . . . . . . . . . . . . . . . . . . . 105 6.7 LAC private credit, � nancial dollarization, and inflation, 2005–08 . . . . . . . . . . . . . 105 7.1 Benchmarking model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 7.2 Domestic equity turnover and enabling environment indicators . . . . . . . . . . . . . . . . 117 8.1 Mutual fund fees in selected Latin American countries . . . . . . . . . . . . . . . . . . . . . . . 127 8.2 Share of deposits in portfolios of Chilean mutual funds . . . . . . . . . . . . . . . . . . . . . . 128 9.1 Development banks in Latin America and the Caribbean: Operative modality of development banks at the beginning of 2009 . . . . . . . . . . . . . . . . . . . . . 150 10.1 Econometric analysis of BCP ratings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 10.2 Capital adequacy requirements for selected LAC-6 countries . . . . . . . . . . . . . . . . . . 166 10.A.1 BCP consolidation into groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 10.B.1 World Bank survey of supervisory practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 10.D.1 Market share of largest companies and funds in selected LAC countries . . . . . . . . . . 176 10.D.2 Ownership concentration in selected Latin American countries . . . . . . . . . . . . . . . . 177 11.1 Main features of real and � nancial cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 11.2 Size of � nancial booms and the probability of crisis: Probit analysis . . . . . . . . . . . . . 192 Foreword D uring the 1980s and 1990s, �nancial interest rates. Also, equity markets across the sectors were the Achilles’ heel of eco- region continue to suffer from a chronic lack nomic development in Latin America of liquidity. These financial indicators are and the Caribbean (LAC). Since then, these precisely those most closely connected to eco- sectors have grown and deepened, becoming nomic growth. more integrated and competitive, with new In addition, the global �nancial crisis has actors, markets, and instruments spring- brought to the surface a new set of issues that ing up and � nancial inclusion broadening. LAC must address, along with every other To crown these achievements, the region’s region around the globe. The crisis has shown � nancial systems were left largely unscathed that �nancial development and �nancial sta- by the global � nancial crisis of 2008–09. bility interact intensely—something that the Now that the successes of LAC’s macro- international �nancial architecture continues �nancial stability are widely recognized and to ignore by almost single-mindedly focusing tested, it is high time for an in-depth stock- on the latter. taking of what remains to be done. The crisis has also highlighted the need For starters, despite this stability, the for the state to keep the process of finan- dividends in terms of output growth and cial development on a safe track, while also greater social equity still remain to be fully accepting that, once a major crisis hits, the realized. It is broadly accepted that the effec- state is the risk absorber of last resort, a role tive operation of � nancial systems is key to that requires additional thinking and institu- reaching these goals. But as evidenced in tion building. In LAC, the important coun- this flagship report, the region continues to tercyclical role played by public banks in face some core � nancial development gaps the crisis also reopened the debate over the and challenges. development role of the state in �nance, and In particular, banks in the region generally this report sets out useful pointers to help lend less to the private sector compared with further the discussion. banks in comparable countries around the These are hard policy questions that will world, and they do so under generally higher keep the world and LAC occupied for many xi xii FOREWORD years. Fortunately, LAC bene�ts from its rich oversight policies that can adapt to a rapidly historical experience in managing systemic changing environment will be essential to risks. But there is little time to spare. The securing its future. policy decisions of today will shape LAC’s Pamela Cox financial world of tomorrow. As global Regional Vice President turbulences continue to impact the region, Latin America and the Caribbean quality � nancial development and � nancial The World Bank Acknowledgments T his report was produced by a core The flagship report was also guided by team led by Augusto de la Torre, an advisory group of Latin America and the Alain Ize, and Sergio L. Schmukler. Caribbean (LAC) countries’ � nancial policy Other members of the core team include: makers, many of whom provided substantive Deniz Anginer, Luis Martín Auqui, César comments at an authors’ workshop in Decem- Calderón, Francisco Ceballos, Mariano Cor- ber 2010. The members of the advisory group tés, Tatiana Didier, Katia d’Hulster, Miquel are as follows: Dijkman, Erik Feyen, Eva Gutiérrez, Socorro Heysen, Eduardo Levy-Yeyati, María Sole- • From Brazil: Alexandre Tombini (Gov- dad Martínez Pería, Claudio Raddatz, Steve ernor, Central Bank) and Luis Pereira da Seelig, and Luis Servén. Silva (Deputy Governor for International The report benefited from very detailed Affairs, Central Bank) and substantive comments on the entire • From Chile: Luis Céspedes (former Head document by Martin Cihak and Roberto of Research, Central Bank) Rocha (both of the World Bank), as well as • From Colombia: Ana Fernanda Maiguashca from most helpful insights and comments (Regulation Director, Ministry of Finance) at various times and on various parts of the • From Costa Rica: Francisco de Paula study by other peer reviewers and discus- Gutiérrez (former Governor, Central sants, including Stijn Claessens (Interna- Bank) tional Monetary Fund [IMF]), Tito Cordella • From Jamaica: Brian Wynter (Governor, (World Bank), Luis Cortavarria (IMF), Asli Bank of Jamaica) and Brian Langrin Demirgüç-Kunt (World Bank), Eduardo (Chief Economist, Financial Stability Fernández-Arias (Inter-American Develop- Unit, Bank of Jamaica) ment Bank), Marialisa Motta (World Bank), • From Mexico: Guillermo Babatz Torres Aditya Narain (IMF), Andrew Powell (Inter- (President, National Banking and Securi- American Development Bank [IDB]), Robert ties Commission) and Carlos Serrano (Vice Rennhack (IMF), Liliana Rojas Suarez (Cen- President of Regulatory Policies, National ter for Global Development), Pablo Sangui- Banking and Securities Commission) netti (CAF), Sophie Sirtaine (World Bank), • From Peru: Javier Poggi (Chief Economist, and Rodrigo Valdés (IMF). Superintendency of Banking, Insurance, xiii xiv ACKNOWLEDGMENTS and Private Pension Funds) and Manuel Julian Genoud, Laura Malatini, Lucas Luy (Head of the Economic Research Nuñez, Paula Pedro, Virginia Poggio, Juliana Department, Superintendency of Banking, Portella de Aguiar Vieira, Gustavo Saguier, Insurance, and Private Pension Funds) Mauricio Tejada, Patricio Valenzuela, Luis • From Uruguay: Mario Bergara (Governor, Fernando Vieira, Tomas Williams, and Central Bank) Gabriel Zelpo. For competent administrative assistance, we thank Erika Bazan Lavanda The report also benefited from written and Ruth Delgado. input by Olivier Hassler, as well as from The main output of the 2011 Regional conversations with experts in the �eld, includ- Flagship Study consists of this flagship ing Aquiles Almansi, Timothy Brennan, report and a companion edited volume Anderson Caputo Silva, Jorge Chan-Lau, with specialized chapters drafted by dif- Loic Chiquier, Joaquín Gutierrez, Tamuna ferent members of the team. While both Loladze, Heinz Rudolph, Ilias Skamnelos, the flagship report and the papers in the and Craig Thorburn. edited volume result from a team effort, the We thank Francisco Ceballos and Tatiana responsibility for their content ultimately Didier for invaluable help in coordinating and remains with their drafting authors. Thus, putting together the documents that consti- the views expressed in this fl agship report tute this 2011 Regional Flagship Study. Out- may at times differ slightly from those standing professional editing was provided expressed in the supporting papers of the by Judith Goff. We bene�ted from excellent edited volume. research assistance at different stages of the The documents of the 2011 Regional Flag- project provided by Matías Antonio, Mari- ship Study are posted on the website of the ana Barrera, Patricia Caraballo, Francisco World Bank’s Of�ce of the Chief Economist Ceballos, Luciano Cohan, Juan Miguel Cuat- for Latin America and the Caribbean (http:// tromo, Federico Filippini, Ana Gazmuri, www.worldbank.org/laceconomist). Abbreviations ALIDE Latin American Association of Development Finance Institutions ALM asset-liability manager ASBA Association of Supervisors of Banks of the Americas ATM automated teller machine BCP Basel Core Principles for Effective Banking Supervision CAF Corporación Andina de Fomento CEBS Committee of European Banking Supervisors CGD Center for Global Development CPI consumer price index DR depository receipt DTI debt service to income EU European Union FD � nancial development FDI foreign direct investment FDIC Federal Deposit Insurance Corporation FHC � nancial holding company FSA Financial Sector Authority FSAP Financial Sector Assessment Program FSOC Financial Stability Oversight Council GDP gross domestic product IDB Inter-American Development Bank IMF International Monetary Fund LAC Latin America and the Caribbean LCH Letras de Crédito Hipotecario LCR liquidity coverage ratio LLR lender of last resort LTV loan-to-value ratio MOU memorandum of understanding MBS mortgage-backed securities NGO nongovernmental organization xv xvi ABBRE VIATIONS NSFR net stable funding ratio OTC over-the-counter PCG partial credit guarantee PEVC private equity and venture capital PFA pension fund administrator PPP public-private partnerships ROA return on investment ROSC report on the observance of standards and codes SAP Sociedades de Ahorro y Préstamo SBPE Sistema Brasileiro de Poupança e Empréstimo (a housing � nance system) SHF Sociedad Hipotecaria Federal (Federal Mortgage Society) SIFI systemically important � nancial entities SME small and medium enterprises TBTF too-big-to-fail UDI Unidad de Inversión UF Unidad de Fomento Overview 1 T he financial systems of the Latin improvements in the enabling environment, America and the Caribbean region lower macroeconomic volatility, more inde- (LAC) are at a crucial juncture. After pendent and better-anchored currencies, a history of recurrent instability and crisis increased financial liberalization, lower (a trademark of the region), they now seem currency mismatches and foreign debt expo- well poised for rapid expansion. Since the last sures, enhanced effectiveness of regulation wave of financial crises that swept through and supervision, and notable improvements the region in the late 1990s and early 2000s, in the underlying market infrastructures (for financial systems in LAC have continued to example, trading, payments, custody, clear- gain in soundness, depth, and diversity. The ing, and settlement).2 size of banking systems has increased, albeit For all the gains, however, many chal- from a low base; local currency bond markets lenges remain. There is still a nagging con- have greatly developed, both in volumes and trast between the intensity of financial sec- in reach over the yield curve; stock markets tor reforms that LAC implemented over the have expanded; and derivative markets— past 20 years (including aggressive financial particularly currency derivatives—have liberalization and vigorous efforts to adopt grown and multiplied. Institutional inves- internationally recognized regulatory and tors have become more important relative supervisory standards) and the actual size to banks, making the financial system more and depth of the region’s financial systems. complex and diversified. Importantly, much In many respects—notably bank credit to progress has been made in financial inclusion, the private sector and domestic equity mar- particularly through the expansion of pay- ket liquidity—the region’s financial services ments, savings, and credit services to lower- industry is underdeveloped by international income households and microenterprises.1 As comparison. The expansion of bank credit has evidence of their new soundness and resil- been biased in favor of financing consumption iency, financial systems in the region, except rather than production. The provision of long- in some Caribbean countries, weathered the term finance—whether to households, firms, recent global financial crisis remarkably well. or infrastructure—remains problematic. The progress in financial develop- Thus, as LAC enters a new phase of ment in LAC no doubt reflects substantial financial development, encouraged by more 1 2 OVERVIEW supportive enabling environments and a more The conventional view on stable macroeconomy, the time is ripe for �nancial development and an in-depth stocktaking of LAC’s financial its limitations systems and a forward-looking assessment of the region’s main financial development Two major LAC-specific historical experi- issues and policies. Such an assessment is ences stand behind what became over the particularly important considering that the past decade or so the conventional wisdom challenges of ensuring sustained develop- on financial development in the region. ment in a more globalized and possibly more The first one is the state dirigisme over the turbulent world are likely to mutate, giving financial sector that dominated the conti- rise to complexities that may have little to do nent during the epoch of import-substitution with the traditional challenges of the past. industrialization, especially in the 1960s and Latin American and Caribbean policy mak- 1970s. That experience ended up leaving a ers will face in the years to come a range of bad taste in the region. It resulted in atro- novel issues for which they need to prepare. phied financial systems and large fiscal costs This regional flagship report aims at associated with mismanaged public banks providing such a stocktaking and forward- that paved the road for a major swing in favor looking assessment of the region’s financial of market-based financial development. The development. Rather than going into detail second experience is the painful recurrent about sector-specific issues, the report focuses and often devastating currency and banking on the main architectural issues, overall per- crises that hit the region, particularly during spectives, and interconnections. The value the 1980s and 1990s. These crises confirmed added of the report thus hinges on its holis- the dangers that poor macroeconomic fun- tic view of the development process, its broad damentals pose for open financial systems. coverage of the financial services industry Crises thwarted and set back financial (not just banking), its emphasis on bench- development by years at a time, with major marking, its systemic perspective, and its adverse implications on growth, employ- explicit effort to incorporate the lessons from ment, and equity. 5 As the 1990s unfolded, the recent global financial crisis. The book moreover, the wave of financial liberaliza- builds on and complements several overview tion that heralded the shift away from state studies on financial development in both LAC interventionism interacted in perverse ways countries and the developing world that have with underlying macrovulnerabilities, exac- been published in the past decade, includ- erbating financial instability. This led the ing studies by the World Bank. 3 The book, reform agenda to put an increasing emphasis moreover, is part of a wave of financial sec- on regulatory frameworks and the institu- tor development studies appearing in 2011 tional enabling environment—an agenda on that the World Bank has undertaken in other which LAC countries embarked with great regions, including the Middle East and North vigor, particularly since the second half of Africa, and Sub-Saharan Africa.4 Finally, this the 1990s.6 study is highly complementary to the recent These experiences—together with a flagship report of the Corporación Andina de worldwide intellectual shift in favor of free Fomento (2011), which focuses on financial market economics—gave rise to a relatively access. strong consensus in the region on a financial This chapter describes the structure and development policy agenda based on four scope of this regional flagship report. First, basic endeavors: getting the macro right, however, it reviews the current state of think- letting financial markets breath, converging ing in LAC regarding financial development toward Basel-inspired standards of pruden- policy, because that will further help to set tial regulation and supervision, and promot- the stage for the scope, structure, and focus ing the broadening of access to financial of the report. services.7 The first, getting the macro right, OVERVIEW 3 reflected the conviction that unlocking the Gates Foundation). It was also boosted by process of financial development had neces- the microfinance revolution, in which LAC sarily to start with macrostability. This aim played a salient role.10 entailed, in particular, the cultivation of the It is clear that the overall thrust of this local currency as a reliable store of value (precrisis) regional consensus on macrosta- that can underpin financial contracts. Over bility and market-friendly policies is endur- the past 20 years or so, therefore, ensuring ing and should not be ignored going forward. stable and low inflation became the first The reforms undertaken under that consensus order of business toward unleashing finan- paid off handsomely during the recent global cial development. Fiscal reform and the crisis.11 Indeed, in a break with history—and development of local currency public bond with the possible exception of the Caribbean, markets were viewed as natural comple- where a few large financial conglomerates ments to monetary reform. failed—no domestic banking system crisis The second endeavor, letting financial was registered in LAC.12 The region avoided markets breath, was initially manifested domestic financial crises even as financial mainly in a rapid process of financial liberal- systems in the G-7 spiraled down into near- ization.8 Subsequently, it incorporated efforts total collapse, averted only by unprecedented, to strengthen the multiple facets (institu- massive bailouts. tional, informational, and contractual) of the It is also evident, however, that the global enabling environment. All of this was also financial crisis invites a serious reassessment, naturally accompanied by efforts to enhance not just of the policy agenda but, perhaps market discipline, which included a sharp more important, of the conception of finan- reduction or elimination of the direct inter- cial development that supported it. The crisis vention of the state in financial activities, raised questions about each of the region’s including the state’s tendency to quickly bail four basic endeavors mentioned above. It illus- out troubled institutions.9 trated that apparent macroeconomic stability The third endeavor, converging toward (for example, the “great moderation� of low Basel-inspired standards of prudential regu- inflation and output volatility, accompanied lation and supervision, focused, prior to the by low interest rates) can contribute to unsus- global financial crisis, on idiosyncratic risks tainable financial development. It showed and ignored systemic risks. It also favored that market discipline can fail miserably, even limiting the perimeter of prudential regulation more so, paradoxically, in the land of well- to deposit-taking institutions. The underlying informed and sophisticated agents (commer- assumptions were that the soundness of indi- cial bank treasurers, investment bankers, fund vidual financial intermediaries implied the managers, stock brokers, derivatives traders, soundness of the financial system and that rating agencies, and so forth). The crisis pro- the well-informed and sophisticated players vided evidence that market discipline may outside the core banking system would disci- actually tighten with financial development, pline each other. instead of being boosted by it. The fourth endeavor, promoting the The crisis also demonstrated that the broadening of access to financial services Basel-inspired oversight program had major for the underserved (that is, small farmers, flaws, including that this program was based microentrepreneurs, small and medium enter- on a great fallacy of composition—as the prises [SMEs], and low-income households) soundness of the parts does not guarantee was added to the policy agenda vigorously the soundness of the whole. It demonstrated but more recently (around the beginning of that the links between financial stability and this millennium). It was spurred by enthusi- financial development are much more com- astic support from multilateral development plex than believed. Finally, the crisis raised banks, nongovernmental organizations, and red flags regarding policies that seek too foundations (such as the Bill and Melinda aggressively to broaden financial access. It 4 OVERVIEW showed that there can be significant tensions financial development and financial stabil- between financial inclusion (for example, ity. The report then takes stock of where the the drive to make a homeowner out of every region’s financial development lies by (a) ana- household) and financial sustainability. lyzing in more detail some of the reasons and Therefore, to chart LAC’s financial devel- policy implications underlying its banking opment going forward requires, on the one depth and equity liquidity gaps and (b) revis- hand, a careful stocktaking of where the iting two themes that are central to its finan- region’s financial systems are now with cial development: long-term finance and the respect to their development dimensions (as role of the state in risk bearing. The last part regards depth, diversity, access, degree of of the report deals with issues of prudential internationalization, and so forth), as well as oversight, first taking stock of progress in to the enabling institutional and regulatory the region and then analyzing the challenges environment. On the other hand, it calls for faced by the region as regards three main fac- a reassessment of the financial development ets of systemic oversight: macroprudential paradigm that has been in effect to date. As policy, microsystemic regulation, and sys- the global crisis has demonstrated, the reas- temic supervision. sessment needs to take into account two The following sections present the main fundamental themes: that the financial devel- threads that bind together all the chapters in opment process itself can lead to financial the report. instability and that, to avoid such instability, the relationship between financial markets and the state needs to be reframed. Revisiting �nancial development Both of these themes are of significant and The flagship report starts in chapter 2 by increasing relevance to LAC. The region’s revisiting the concept and process of finan- financial systems are experiencing strong cial development, something on which expansion pressures, not least owing to surg- surprisingly little has been written. Financial ing capital inflows, and this situation poses development is all about the gradual grind- risks of financial excesses and bubbles—the ing down of two basic classes of frictions— “dark side� of financial development. In turn, agency frictions and collective frictions. this pressure raises the premium on the qual- Agency (information asymmetry and contract ity of financial development policies, thereby enforcement) frictions limit financial con- highlighting the need for a more effective tracting because they induce a misalignment complementation between the role of mar- of incentives between the principal and the kets and that of the state. agent. Collective frictions (action and cogni- T he 2011 Regional Flagship Study tion) hinder financial development because research project aims at meeting this dual they constrain participation—thereby limit- challenge—stocktaking and reassessment ing the positive externalities in immediacy of the financial development paradigm—to and liquidity that derive from large, inter- set on firmer ground the financial develop- connected financial intermediaries and deep ment agenda for LAC going forward. This markets. The constant interplay between summary report synthesizes the results of a these two classes of frictions—as market comprehensive analysis of the status, pros- participants and financial institutions find pects, and challenges of sustainable financial ways to reduce or circumvent them under development in the region. Much of the back- the stimulus of competition and innova- ground research for this report can be found tion—shapes financial structure and makes in a companion book.13 This report starts by it evolve over time. The gradual easing of presenting a general conceptual framework agency frictions helps boost participation of financial development that provides the (that is, financial inclusion). In turn, by key lines of intersection and organization for unleashing positive network and scale exter- the rest of the report, including links between nalities, the benefits of participation become OVERVIEW 5 self-reinforcing. This process provides broad prudential oversight, the state needs to do it pointers about the order in which various not just from an idiosyncratic perspective but financial activities are likely to emerge, and also from a systemic perspective that makes the shape of the paths they are likely to fol- three types of connections: between the parts low once they emerge. and the whole, across time (to integrate the These patterns are broadly verified dynamics of the process), and between the through an econometric analysis of a large forces of the bright side (financial develop- set of financial indicators. Financial activities ment) and those of the dark side (financial for which agency frictions can be resolved instability). more easily develop first. Activities that are the most strongly boosted by participation have the most convex development paths, Where is LAC? rising steeply once economic and financial The above narrative sketches the broad con- development reaches a certain threshold. ceptual background and main threads on Government borrowing emerges before basic which this flagship report is constructed. banking services, which emerge before capi- Chapters 3 to 5 then offer a comprehensive tal markets. Institutional investors appear description of the current state of financial at various stages of the process, reflecting development in LAC. Chapter 3 system- policy factors as well links with other com- atically compares (and benchmarks) LAC-7 ponents of the development process. At the with other regions and with other coun- same time, the activities that are more tightly tries in LAC.14 The chapter concludes that linked with market development—and hence the region’s financial systems have become benefit the most from the positive externali- deeper in ways that are consistent with the ties of interconnectedness (equities, corporate broad patterns described in chapter 2. There bonds, mutual funds)—are the most convex. has been a transition from a mostly bank- However, financial development also has a based model to a more complete and inter- dark side, as the apparently successful easing connected model in which bond and equity of agency frictions triggers lethal collective markets have increased in both absolute and action failures—negative externalities, free relative sizes; institutional investors (mutual riding, or coordination failures—or triggers funds, pension funds, insurance companies, a second round of agency failures, much as etc.) play a more central role; and the num- building more highways exacerbates con- ber and sophistication of participants have gestion by increasing traffic. Alternatively, increased. At the same time, the strengthen- the positive externalities of increased mar- ing of monetary management has allowed ket participation in the good times can turn an important shift in the nature of financ- into creeping negative externalities and other ing, toward the longer term and into the collective action failures in the bad times. In local currency. either case, the problems arise from wedges However, chapter 3 also concludes that between private and social costs and benefits, LAC’s financial systems remain underdevel- which markets are unable to handle on their oped in some key respects. Even after con- own. Thus, market-based financial devel- trolling for a number of possible economic opment can lead endogenously to crippling and structural determinants, including size, financial instability. (Of course, misguided chapter 3 finds that LAC banks lend less and state interventions can also initiate or add to charge more than they should. The lack of perverse dynamics.) domestic bank credit is only partially off- This interplay of frictions calls for a twin set by other types of financing, particularly policy response. In designing good finan- cross-border. Consumer credit has expanded cial development policy, the state needs to at the expense of firm financing and housing be aware of the potentially adverse longer- finance. Moreover, domestic equity markets term stability implications. In shaping up in LAC are illiquid and highly concentrated, 6 OVERVIEW and insurance is relatively underdeveloped. a particular focus include financing SMEs, While institutional investors are sophisti- bringing down the cost of financial services, cated and large, a large share of their portfo- and reforming creditor rights. lios continues to be allocated to government Chapters 6 and 7 then explore the sever- bonds and bank deposits. Furthermore, there ity of, possible reasons for, and potential is a large heterogeneity within the region, policy implications of the banking and with the rest of LAC being substantially less equity gaps, respectively, that were iden- developed than LAC-7. tified in chapter 3. Chapter 6 argues that, Chapter 4 then looks at some aspects of to a large extent, the banking gap simply financial globalization. It finds that although reflects LAC’s turbulent financial history. LAC spearheaded a strong process of finan- The region has not yet fully recovered from cial internationalization during the 1990s, its the repeated credit crashes of its past. This international financial integration tended to argument puts the spotlight squarely on the stabilize over the following decade, in con- dark side and the need for ensuring finan- trast with other emerging economies. In par- cial sustainability through a good mix of ticular, there has been a significant decline in oversight and development-oriented policies. the share of bonds issued by the public sector But limited demand for credit (that is, a lack abroad and an expansion of local govern- of bankable projects), possibly reflecting ment bond markets. Inversely, however, while LAC’s mediocre output growth, also seems equity financing has tended to shift toward to explain a sizable chunk of the gap. Here, home markets, especially in the advanced of course, the possible policy responses go economies, for LAC it has increasingly gone much beyond the financial sector. However, abroad. Also, in contrast with other emerging growth-inducing financial policies, such as economies, LAC’s equity liabilities continue those that facilitate longer maturity loans to be dominated by foreign direct invest- for SMEs or infrastructure projects, also ment (FDI) rather than by portfolio equity, seem to be called for. Overcoming the bank- which is consistent with the shortcomings of ing gap also has to do with addressing the the local equity markets. At the same time, remaining agency frictions. Interestingly, in LAC’s international financial integration has LAC, the main residual bottleneck is con- become safer—as debt liabilities have fallen, tractual (contract enforcement and creditor reserve assets have risen. rights) rather than informational. Although lack of funding is clearly not Chapter 7 then switches focus to LAC’s an overall issue (in fact, many countries gap in domestic equity market liquidity. It have imposed or are considering controls concludes that both agency and collective on capital inflows), there is still progress to action frictions contribute to explaining be made as regards broadening and deepen- the equity liquidity gap. The substitution of ing participation and financial inclusion, a domestic equity by foreign markets under the theme that is surveyed in chapter 5. Most pull of a bigger (more liquid and connected) available indicators of financial inclusion marketplace is a first obvious reason. The suggest that LAC-7 is not obviously under- increased trading of the larger stocks that performing compared to its peers, particu- went abroad has in LAC a particularly strong larly when it comes to payment and saving adverse effect on the liquidity of the domestic services and microfinance. Yet capital mar- stock market. Much as for the banking gap, kets continue to be highly concentrated, the region’s history of macroeconomic and indicating that access is highly unbalanced financial turbulence seems to have something across firms. Although most LAC-7 coun- to do with this atypical outcome, although tries have adopted comprehensive policy for specific reasons that remain to be eluci- agendas to promote financial inclusion, the dated. The preponderance of buy-and-hold attention given to this agenda has been spot- pension funds, relative to more active insti- tier in the rest of the region. Areas that need tutional traders such as mutual funds, and OVERVIEW 7 weaknesses in corporate governance also of liquidity have not fully materialized. For all seem to have played some role. their positive contributions to overall finan- The policy lessons on how to promote cial deepening, defined-contribution pension equity markets (thereby facilitating market funds continue to have their portfolios con- access) are challenging. Clearly one must deal centrated in public sector bonds, short-du- with the participation externalities, as illiquid- ration bank deposits, and liquid equity and ity begets illiquidity. Without liquidity, exit is corporate debt securities. At the same time, difficult and price revelation is weak, thereby with the demise of the monoline insurers, the discouraging entry. However, the usual mix public sector remains the only entity able to of liquidity-enhancing policies (reducing the directly provide, guarantee, or enhance long- fragmentation of issues, enhancing securities term debt finance. All of these factors are clearance and settlement, organizing securi- taking place in an environment in which the ties lending and borrowing facilities, and so region is awash with investible funds, which forth), while always welcome, will probably is all the more puzzling. Clearly, going long is lead to only modest improvements. Efforts to harder than often believed. further improve property rights, information, Chapter 8 analyzes the problem of long- enforcement, and governance can help domes- term finance and the policy options avail- tic equity market development, although they able to LAC using the same twin lenses— may at the same time favor an even greater agency frictions and collective frictions—that migration of stock trading abroad. Given are used throughout the report. As regards the dominant role of size (of markets and of agency frictions, perhaps the most perva- issues) in stock market development, the pros- sive issue throughout the region (one that is pects for small economies are particularly relevant to pension funds as well as mutual challenging. In this context, a realistic assess- funds) is how to better align the incentives of ment of the costs and benefits of global versus asset managers with those of investors. States regional integration of stock market activ- will need to find ways that help promote mar- ity will be needed. Other key questions are ket discipline by facilitating the comparison whether, to facilitate the development of the of asset managers’ performance, but with- market for primary equity issues, the smaller out exacerbating the focus on short-term countries in LAC should go stricter or lighter returns. Using life-cycle benchmarks and as regards corporate governance standards, nudging defined-contribution pension funds and whether they should, as a result, accept into mimicking the investment behavior of a less liquid secondary market. As yet, there defined-benefit pension funds should help. are no clear answers. There is, however, one At the same time, states need to find sensible certainty: the conventional stability-oriented, ways to provide more guidance to investors to international standards–laden policy agenda resolve the problems of consumer protection will not suffice. and bounded rationality that characterize the pension fund industry, but without under- mining monitoring incentives or promoting Promoting the bright side moral hazard. The next two chapters, chapters 8 and 9, As regards collective frictions, a first route narrow the focus on two basic issues at the governments can take is to promote market core of sustainable financial development: participation, hence liquidity, by continu- long-term finance (“going long�) and the role ing to improve the enabling environment of the state in risk bearing. While LAC has or trading infrastructure. However, LAC made much progress in lengthening contracts, policy makers need to be realistic about the notably public debt contracts, considerable feasibility (and desirability) of relying on a progress still can be made. In particular, the U.S.-style solution that uses liquid secondary hopes initially placed on pension funds to markets to align investors’ desire for an exit help lengthen maturities and overcome lack option with borrowers’ long-term funding 8 OVERVIEW needs. As discussed in chapter 7, this solution and collective action (participation) frictions hinges crucially on size. In addition, it can (that limit the scope for spreading the risk raise systemic risk. Wedges between private throughout market participants) coexisting and systemic interests can widen as private with agency frictions (that concentrate risk agents increasingly free ride on liquidity and through skin-in-the-game requirements). relax their monitoring efforts. Alternatively, Hence, it is the state’s natural advantage serious problems of incentive distortions can in resolving collective action (instead of be created if governments overly commit to agency) frictions that justifies public guar- being the liquidity providers of last resort. antees. The state is thus naturally called to Thus, a second key avenue for dealing with play to its strengths to complement markets, collective frictions is, instead of promoting rather than substitute them. participation, to mandate it. Indeed, when Focusing on the role of the state from this combined with a vibrant annuities industry, perspective raises a policy agenda that is as the mandated private pension funds industry broad as it is thorny. A key implication is that provides one of the most promising levers states, before providing guarantees, should for the development of long-term finance in first exhaust efforts to facilitate the spreading the region. However, design problems that in of risk through private guarantees and pri- many countries currently stand in the way of vate risk sharing. The state can promote par- linking pensions and insurance would need ticipation without taking risks itself through to be removed. policies that directly ease the frictions (where, Chapter 9 then proceeds to revisit the role for instance, a development bank acts as of the state in financial risk bearing, a topic coordinator) or through policies that man- that has risen to greater visibility in the after- date or gently coerce participation, as in the math of the global crisis. Again, the chapter case of the mandatory contributions to pri- analyzes this theme from the perspective of vately administered pension funds that were the underlying frictions. It starts by remind- reviewed above. Given the positive externali- ing the reader that over the past half century ties, the state can also use well-targeted subsi- or so, LAC has undergone large paradigm dies as part of this type of intervention. swings, from state dirigisme to market laissez- Chapter 9 then assesses the rationale faire, and eventually to a more eclectic view. and scope for public guarantee programs. It Throughout, agency frictions and social emphasizes the need for these programs to externalities—whether first-tier public refocus their rationale more directly around banks performed a better or worse agency risk aversion and the agency or collective fric- task in integrating the social dimensions of tions with which a program interacts. A basic lending—permeated the debate. A parallel question in this regard is to account for (and debate developed on public banks’ second-tier price in) the hidden risks that may explain role in the provision of loans and guarantees. why the private sector cannot itself provide the Thus, chapter 9 reviews in some depth guarantees. The more the state seeks to push the conceptual justifications for public the risk frontier, the more scope in principle financial risk bearing. It first concludes that the state has to intervene, but also more need risk aversion, a seldom emphasized factor, for caution. Thus the chapter discusses ways is central to guarantees. Without risk aver- in which the risk to taxpayers can be bounded sion, no guarantee program, whether pri- and public governance enhanced. Clearly this vate or public, can be justified. The chap- is a rich area for further exploration, in Latin ter then concludes that, given risk aversion America and the Caribbean and elsewhere. among lenders, externalities alone justify subsidies but not guarantees; and agency frictions alone justify private but not public Dealing with the dark side guarantees. Thus, public guarantees can be While retarding its financial development, justified only in the presence of risk aversion LAC’s turbulent macrofinancial history has OVERVIEW 9 also stimulated efforts to overhaul regula- and insufficient cooperation and coordina- tion and supervision—that is, prudential tion among supervisors combine to make oversight. At a time when many developed- the challenge more difficult. Effective cross- country supervisors were bent on eas- border cooperation also remains a major ing intermediation through more market- challenge, all the more so in LAC in view of friendly regimes and less expensive capital the importance of foreign banking. and liquidity buffers, many LAC countries All in all, LAC now has a much better went in the opposite direction. Nonetheless, foundation on which to build and deal with chapter 10 shows that, as much as progress the new challenges of systemic oversight. was broad, it was also uneven, both across These challenges include connecting the parts regions and across areas. In a comparison and understanding how one may affect the against an econometric analysis of the Basel others, building up a proactive capacity to Core Principle assessments performed over deal with the unstable market dynamics of the past 13 years (an exercise that clearly has the dark side, and thinking about develop- many caveats), the LAC-6 countries (LAC-7 mental and prudential policies as two sides minus Argentina) generally perform better of the same coin. In view of lead times and than other countries, even after controlling longer-term dynamics, now is the time for for different levels of economic develop- thinking ahead. Moreover, as most of LAC ment. The non-LAC-6 countries perform is currently in the midst of a boom resulting somewhat worse. There are also important from the potentially perilous mix of Asian differences across supervision areas, with accelerated growth and the developed world’s some issues understandably more difficult low interest rates, there is not that much time to tackle than others. to ponder anyway. Thankfully, LAC has a leg Two basic issues concerning the legal up. Its prudential buffers are currently high, framework—the independence of bank LAC supervisors have made important strides supervisors and their legal protection— toward improving traditional oversight, and emerge from a variety of sources as still prob- LAC’s numerous past crises have given its lematic in many LAC countries. On regula- supervisors a definite edge. tory issues, there is also some unevenness. Chapter 11 then reviews the potential Many countries are still not fully meeting the benefits and challenges of macroprudential minimum Basel I international standards on policy in LAC. The chapter starts with a capital requirements, and the implementation thorough comparative analysis of financial of Basel II has been thus far limited. How- cycles. It reveals that LAC credit cycles are ever, in many areas, including on the regula- generally more protracted and abrupt than tion of credit risk, very substantial progress those in other emerging and industrial coun- has been made. On basic supervisory issues, tries. Likewise, cyclical fluctuations in bank LAC-6 countries again tend to perform bet- leverage, housing prices, and real exchange ter than the rest, suggesting that effective rates also exhibit greater amplitude in LAC, implementation might be a problem mostly especially in the downturn. The uncondi- in the lower-income countries. In any event, tional probability of a banking crisis and the important progress has been made across the frequency of crash landings following lending region, including as regards the shift to risk- booms are also higher. Finally, bigger booms based supervision. Finally, as regards consoli- are more likely to end badly. These facts echo dated and cross-border supervision, a com- the history of macroeconomic instability of plex issue that to some extent prefigures the the region. They imply that management challenges of systemic oversight, most LAC of financial risks over the cycle represents countries have had a hard time. Here again an even bigger policy concern in LAC than LAC-6 exceeds its benchmark; however, elsewhere. opaque conglomerate structures, high con- In considering policies to manage systemic glomeration, high ownership concentration, risk over the cycle, the main objective should 10 OVERVIEW not be to eliminate the financial cycle but to are building up, and conversely when they make the financial system more resilient and unwind. At the same time, however, a more to tackle the externalities that amplify the active macroprudential management can help cycles and promote an excessive buildup of relieve some of the pressure from monetary risk. A first priority should be placed on set- policy, thereby helping to reconcile inflation ting realistic objectives along various levels of and exchange rate targets under open capital ambitiousness, ranging from removing exist- accounts—an issue that many LAC central ing procyclicality in traditional regulations, bankers have close to their hearts. Counter- to building financial system resilience to cycli- cyclical deployment of fiscal policy would, cal fluctuations, to dampening the cyclical of course, also help in achieving financial fluctuations themselves. Much more research stability. and testing are needed. How to measure Chapter 12 then shifts the focus from the buildup of risk is a particularly difficult connecting the system through time to con- challenge. In emerging regions such as LAC, necting the parts to the whole—that is, from very close monitoring of credit accelerations macroprudential management to micro- is likely to be needed to disentangle hazard- systemic regulation. The chapter starts by ous credit booms from desirable long-term reviewing issues associated with the set- financial deepening. The quest for developing ting of the outer perimeter of regulation. a robust macroprudential policy framework Although regulatory perimeters are already faces a number of other unresolved issues, widely extended in LAC, the issue is nonethe- including the need to find proper balance less relevant because boundary problems— between buffering the financial system and the incentive to migrate intermediation to the dampening the cycle; between institution- less regulated domains—continue to exist specific and system-wide triggers and tar- and because spreading resources too thinly gets; between price-based and quantity-based may end up compromising the effectiveness tools; and between rules and discretion. of supervision, providing an unwarranted On many of these issues, LAC is not sense of comfort, and possibly breeding behind the crowd. In fact, many LAC coun- moral hazard. tries have already introduced countercyclical Thus, to economize on resources, some provisioning or capital requirements. And countries have resorted to auxiliary models of several countries have used reserve require- delegated supervision for smaller credit coop- ments proactively to help manage capital eratives. Another form of delegation could inflows and the credit cycle. Finally, many involve allowing those entities that fund LAC countries have introduced, in the recent themselves only from regulated intermediar- past, regulations to limit the risks associated ies to be exempt from prudential regulation. with foreign currency exposures, which are Still another approach is to grant the supervi- also systemic in nature and similar in spirit sor statutory authority to readily extend the to the systemic regulations currently being perimeter as circumstances warrant (as in the debated to manage credit cycles. Dodd-Frank act). However, exercising such Chapter 11 also makes the point that discretionary powers is a particular chal- reforms in monetary management, as well lenge given the region’s administrative law as in macroprudential management, may be framework. Another daunting task comes called for. In view of recent evidence show- from the fact that systemic risk that builds ing that low policy interest rates promote the up outside the financial system may end up search for yields and encourage banks to push contaminating the system through its impact the risk frontier, timely monetary tightening on the markets in which both financial and might also contribute to maintaining pru- nonfinancial firms participate, or through dent lending standards in the upswing. This common ownership. approach might mean allowing inflation to However, regulatory arbitrage can also undershoot its target as financial imbalances take place within the perimeter of regulation OVERVIEW 11 when different silos are regulated differently. into the supervisory process (in a context Indeed, licenses granted to intermediaries in of appropriate accountability) but without the LAC region tend to have a narrow scope relaxing regulations so much that prudential of permissible activities, typically separating oversight ends up losing its “teeth.� The lat- commercial from investment banking and ter is an even greater challenge in civil law insurance from banking. The current silo countries, such as those in Latin America, approach is hindered, however, by the weak- where supervisors tend to be constrained to nesses in consolidated regulation stressed in taking only actions that are specified in laws chapter 10. One approach to dealing with this and regulations. issue is to pursue a fully uniform, risk-based A second key issue is how best to com- approach in which all entities are similarly bine and intermingle a top-down perspective regulated, ultimately leading to universal with a bottom-up analysis. Indeed, one of the licenses. There are, however, potential draw- weaknesses in the financial stability analysis backs with such an approach. It is technically published thus far by central banks has often challenging, it could potentially lead to a loss been the absence of the supervisors’ perspec- of diversity that could make the system more tive on what is happening at individual insti- fragile, and it could foster the emergence tutions. The coordination necessary for this of systemically important financial entities process to succeed, down to the technical (SIFIs) that are deemed too big to fail. staff level, is certainly not trivial, particularly Indeed, the region has many SIFIs, and in countries where bottom-up supervision is there appears to be some consensus on the outside the central bank. A closely related need to regulate them differentially. Imple- (but conceptually distinct) issue is the relative menting such a differential treatment will be emphasis on offsite versus onsite supervi- quite challenging, however, in view of data sion. One might think that systemic supervi- and analytical requirements. The global crisis sion, because it involves the forest more than has also highlighted the need to resolve unvi- the trees, is more about offsite than onsite, able financial institutions, particularly SIFIs, but that is unlikely to be the case. Instead, in a nondestabilizing fashion. While the crises systemic supervision calls for a review of of the recent past have led to the introduction onsite supervision, stressing its complementa- in many LAC countries of sophisticated bank rities with offsite analysis. failure resolution frameworks, these frame- The global crisis also calls for a review of works remain largely untested. In fact, crisis the role of market-based financial indicators simulation exercises conducted in several and the reliance on market discipline. The countries have shown important shortcom- key question arising from the crisis is not ings in both tools and processes. Moreover, whether market discipline is good or bad, but the development of failure resolution systems instead how supervisors should make better for financial conglomerates (including the use of market signals. When weak market ones that operate across borders) is in its signals constitute a severe limitation, the key infancy. question therefore becomes one of how policy Chapter 13 closes the flagship by discuss- makers can help develop instruments (such as ing systemic supervision, an issue that prob- subordinated debt) that help price the risk, ably has not received sufficient attention thus thereby facilitating risk discovery. Unless far but that is nonetheless central to an effec- supported in some fashion by the state (and tive systemic oversight. The chapter starts by perhaps even subsidized), these instruments looking at the interface between regulation may simply be too expensive to see the light and supervision. The inherent tensions and of day. An important research agenda for the complementarities between regulation and region, therefore, is to help design, introduce, supervision are an essential part of the “rules and support such instruments. versus discretion� debate. Hence, the main A key requirement for proper market dis- challenge is to build sufficient discretion cipline is analysis and information. Because 12 OVERVIEW much of information is a public good, there reforms undertaken by LAC during the 1990s is a good argument that supervisory agencies and early 2000s can be found in de la Torre, should provide more of it, including informa- Gozzi, and Schmukler (2007a). Chapter 10 tion on (and better analysis of) the system as of this report documents the progress in LAC with respect to banking supervision. a whole, how it is wired and interconnected, 3. Recent overview studies of LAC’s �nancial sec- and what the risks ahead are. When risks are tor include the Inter-American Development detected, supervisors not only need to inform Bank’s (2005) regional flagship report titled and guide but also need to take action. Unlocking Credit: The Quest for Deep and Successfully implementing systemic super- Stable Bank Lending, which focuses on vision will require first building up skills, banking; the World Bank’s (2006) regional which involves a quantum jump, not a mar- study titled Emerging Capital Markets ginal improvement. It also will require suit- and Globalization: The Latin American able organizational arrangements. The need Experience, which focuses on securities mar- for better coordination between monetary kets; the book by Stallings and Studart (2006), and prudential management with a systemic Finance for Development: Latin America in Comparative Perspective; and the Inter- perspective naturally suggests that the cen- American Development Bank’s (2007) regional tral bank will have to play a leading role. As flagship report titled Living with Debt: How central banks assume this role, however, care to Limit the Risks of Sovereign Finance. should be taken that their independence is Relevant overview studies of �nancial sec- not compromised. What will matter most is tor development issues with a global (rather to put in place appropriate decision-making than a regional) focus include two policy and interagency coordinating arrangements. research reports by the World Bank: Finance If a systemic oversight or financial stabil- for Growth: Policy Choices in a Volatile ity council is set up, an important element World (World Bank 2001) and Finance for will be to ensure its accountability. Last but All? Policies and Pitfalls in Expanding Access not least, staff must be encouraged to work (World Bank 2007). 4. See “Financial Access and Stability for the together across agencies. In addition to coor- MENA Region: A Roadmap,� a World dinating at home, supervisors will also need Bank Middle East Financial Sector Flagship, to coordinate better across borders. In LAC, “Financing Africa: Through the Crisis and the importance of foreign banks makes this Beyond,� a joint report from the African even more of a priority. Development Bank, the German Federal Ministry for Economic Cooperation and Development, and the World Bank, and the Notes Global Financial Development Report, a 1. LAC was in fact a salient player in the world- World Bank Flagship Report coordinated by wide micro�nance revolution, which decisively the FPD Chief Economistís Of�ce, forthcom- shifted micro�nance from a nongovernmental ing in 2012. organization–based, grant-intensive activity 5. The uncertainty resulting from macroeconomic to a pro�table, commercially viable banking volatility—particularly high and unpredictable business. inflation—was deleterious to �nancial devel- 2. Financial sector reform agendas in LAC were opment, most of all at the longer maturities. It often aided by Financial Sector Assessment corroded the role of money as a store of value, Program (FSAP) documents undertaken jointly leading to a gradual buildup of currency and by the International Monetary Fund and World duration mismatches. The inflexible exchange Bank in several countries in the region since rate regimes, which were adopted in part to 1998, as well as by technical assistance (includ- bring down inflation expectations, ended up ing in the context of loan operations) provided exacerbating interest rate volatility and cur- by these institutions. Comprehensive FSAP rency mismatches and became vulnerable to documentation, including country reports self-ful�lling attacks. This experience com- and reviews of the program, can be found at pounded the region’s proneness to currency http://worldbank.org/fsap. A fairly detailed crashes associated with unsustainable �scal documentation of the capital markets–related positions. Widespread (currency, duration, and OVERVIEW 13 maturity) mismatches, for their part, boosted 11. See, for example, de la Torre et al. (2010b), the fragility of �nancial systems to currency IMF (2010), and Porzecanski (2009). upheavals, interest rate volatility, and runs on 12. According to the Inter-American Development banks. In addition, repeated �nancial crises led Bank (2005), in recent history, the LAC re- to multiple ownership changes. By facilitating gion has had the highest incidence of banking the entry of often un�t, improper, or poorly crisis in the world. In particular, 27 percent capitalized bankers, some reprivatizations of of LAC countries (35 percent excluding the banking systems in turn compounded the brit- Caribbean) experienced recurrent banking cri- tleness of �nancial development. In addition ses during the 1974–2003 period, compared to to their major—and well-known—adverse 13 percent in Sub-Saharan Africa, 11 percent effects on growth and employment, �nancial in Eastern Europe and Central Asia, 8 percent crises have proved to be highly regressive for in East Asia and the Paci�c, and 0 percent in income and wealth distribution (see Halac and high-income Organisation for Economic Co- Schmukler 2004). operation and Development countries. 6. For a detailed documentation of the �nancial 13. The companion book is a forthcoming edited sector reforms undertaken in LAC during the volume that contains 11 specialized chapters 1990s, see de la Torre, Gozzi, and Schmukler and several appendixes, produced by differ- (2007a, 2007d). ent members of the flagship team. The chap- 7. A discussion of the evolution of �nancial ters of the edited volume are referred to later development policy in Latin America and the on, in the sections of this report to which Caribbean, along with relevant references to the they signi�cantly correspond. The entire flag- copious literature on the subject, can be found ship study (this summary report, the Edited in de la Torre, Gozzi, and Schmukler (2006). Volume, and a wider set of graphs and tables) 8. A characterization of the �nancial liberaliza- can be found at the website of the Of�ce of tion sequencing debate, along with the rel- the Chief Economist for the Latin America evant references, can be found in chapter 4 of and the Caribbean Regional Vice Presidency de la Torre and Schmukler (2007). (http://www.worldbank.org/laceconomist). 9. This microeconomic paradigm underpins 14. The LAC-7 group comprises Argentina, World Bank (2001). Brazil, Chile, Colombia, Mexico, Peru, and 10. See, for instance, Armendáriz de Aghion Uruguay. Uruguay is included but República and Morduch (2005); Robinson (2001); Bolivariana de Venezuela is not, mainly be- Sengupta and Aubuchon (2008); and Yunus cause there is much more data for the former (2003). than for the latter. Financial Development: Bright Side, Patterns, 2 Paths, and Dark Side W hat is financial development (FD), with assumptions about the nature of and how predictable is it? Does it information and the extent of rational- follow a single path or multiple ity. Thus, the two agency paradigms are paths? What are the sequences and shapes of costly enforcement and asymmetric infor- the development paths followed by various mation; the two collective paradigms are indicators of FD?1 In tackling these issues, collective action and collective cognition. this chapter introduces the basic threads • Financial structure reflects economic that run through the rest of this report. 2 agents’ efforts to find the path of least It presents a conceptual framework of FD resistance around the frictions; in turn, based on a simple typology of frictions. FD (the evolution of � nancial structure Using this framework, the chapter explores over time) reflects the gradual erosion and explains some of the patterns and paths of frictions, quickened by innovation, of FD and discusses some of the maladies returns to scale, and network effects. of FD that justify and shape the state’s role • It is thus hypothesized that the order of and policy response. In this way, the chapter appearance of � nancial activities should provides the necessary foundations to assess reflect the intensity of the frictions to the Latin America and Caribbean region’s which they are exposed and should gen- FD and identify the policy challenges that erally correlate with scale effects and lie ahead. Main highlights of the chapter are the shape (convexity or concavity) of the as follows: development paths. • These patterns are, indeed, broadly veri- • The frictions hindering � nancial contract- fied empirically when a set of relevant ing, and hence FD, can be classi�ed into �nancial development indicators is exam- agency frictions, which restrict the scope ined. Thus, public debt, banking, and capi- for delegation, and collective frictions, tal markets develop sequentially and along which restrict the scope for participation. increasingly convex paths. There are, how- • Each of these broad classes of frictions ever, some FD indicators whose paths are can be analyzed under two paradigms, notable and quite telling exceptions. which the chapter de� nes depending on • FD indicators do not all follow the the relevant friction and its interaction same path for all countries, reflecting 15 16 F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E country-specific development policies, opening a wide scope for the financial sys- path dependence, leapfrogging, cycles tem to add value to society. and crashes, and endogenous quantum Two basic classes of frictions hinder finan- developmental jumps. cial transactions—agency frictions and col- • The same gradual easing of frictions that lective frictions. Each class can, in turn, be underlies the “bright side� of FD can also subdivided into two categories: one relates help breed the tensions and fault lines to informational frictions (including agents’ that may eventually burst into the open limited capacity to process and understand in the form of � nancial crises (that is, the information as much as their limited capac- “dark side� of � nance). ity to obtain it); another relates to what can • The dark side may manifest itself as a be loosely defined as relational frictions, that result of one (or more) of three processes is, agents’ capacity to agree and to act upon and their interactions: the easing of collectively beneficial financial arrangements agency frictions that triggers lethal col- and to enforce bilateral contracts (table 2.1). lective action failures or breeds second- Thus, this simple dichotomy underpins four generation agency frictions; the easing of paradigms, two of which—asymmetric collective action frictions in good times information and costly enforcement—are that unleashes crippling collective action associated with agency frictions, and two failures in bad times; or the easing of of which— collective action and collective agency or collective frictions that leads to cognition—are associated with collective problems of collective cognition. frictions. The four paradigms can also be • Obvious caveats apply, as the analysis in associated with the different stages a finan- this chapter is, for the most part, an exer- cial contract goes through, from preparing cise in positive economics with limited the contract to negotiating and enforcing it data; hence, it is suggestive rather than (figure 2.1).3 conclusive. Agency frictions hinder FD because they limit the capacity of individuals to delegate The first section discusses the conceptual and contract bilaterally. Asymmetric infor- framework (the bright side of FD). The next mation frictions lead to a misalignment of two sections apply this framework to the incentives between the “principal� and the empirics of FD, considering the patterns and “agent�—the agent (for example, the banker, dynamic paths, respectively. The fourth sec- the asset manager, the debtor) can use his tion explores the dark side of FD. The final informational advantage to act in ways that section concludes. are not in the interest of the principal (for example, the depositor, the mutual fund The bright side investor, the lender).4 This misalignment, in turn, can trigger the commonly known mar- Financial development is all about the ket failures of adverse selection, moral hazard gradual grinding down of the frictions that and shirking, and false reporting. Information hinder financial contracting. If economies operated in a frictionless Arrow-Debreu world of complete markets, risks would be fully and efficiently internalized in the TABLE 2.1 A simple typology of paradigms price system, suppliers of funds or insurance Incomplete would deal directly in the market with the Full information/ information/ users of funds or insurance, and neither of   full rationality bounded rationality them would have a use for financial service Bilateral Costly Asymmetric providers. In the real world, however, fric- focus enforcement information tions make the markets and the ability to Multilateral Collective action Collective cognition contract incomplete and imperfect, thereby focus FIGURE 2.1 Frictions, paradigms, and failures Contract stage Friction Underlying problem Paradigm Failure asymmetric asymmetric confusion getting the uncertainty, information facts (before, information, and during, and after) learning costs symmetric confusion collective cognition collective failures agency failures agreeing on a participation and coordination failures collective action contract bargaining costs enforcing the enforcement limited pledgeability, costly enforcement contract costs commitment, and liability Source: Authors. 17 18 F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E asymmetry frictions limit financial contracts risk, buying insurance and hedges, staying to those in which agents have their own liquid). In turn, the state seeks to facilitate resources at risk (“skin in the game�) and/ these private responses through a set of pro- or where the principal can adequately screen gressively more intrusive public interventions: and monitor the agent. Enforcement frictions (a) the provision of basic contractual and lead to a misalignment of incentives due to informational infrastructure (that is, public imperfect pledgeability—a situation where services that facilitate private contracting); the agent is unable to credibly commit to (b) regulation and taxation (internalization honor the contract.5 Imperfect pledgeability of externalities or consumer protection); (c) thus restricts financial contracts to those that pure coordination (catalytic involvement, can be effectively collateralized. systemic lending of last resort, and manda- Collective frictions, by contrast, hinder FD tory participation); (d) risk absorption and because they constrain participation—that is, risk spreading (government guarantees); and they constrain financial inclusion, broadly (e) the direct provision of financial services defined. Participation can increase along the by the state. intensive margin—that is, the same players Different components of the financial engaging in more financial transactions— system help deal with frictions in differ- and along the extensive margin—that is, the ent ways. Consider, for example, the need inclusion of new players in financial activity. for information. Capital markets provide Much of the gains from financial activity price signals and motivate the supply of relate to the reduction in transaction costs hard, public information by borrowing and the increase in liquidity and risk diver- firms. Banks fill the information gap by sification benefits that result from multi- generating proprietary information. Fund lateral arrangements in which many agents managers do so by monitoring marketable participate. Such arrangements can be either assets. Market facilitators (auditors, rat- markets, where transactions can be con- ing agencies, credit bureaus) sell special- ducted around a simple trading platform, or ized information and analysis. As another a financial institution (a bank, an insurance example, regarding the need for risk reduc- company, a mutual fund) that offers services tion through diversification, capital mar- whose benefits are pooled across a large num- kets allow investors to buy assets with dif- ber of customers. The higher the number of ferent risk profiles, which banks, insurance participants, the higher the benefits of partici- companies, and asset managers accomplish pation. However, participation is hindered by through pooling. Finally, as regards the cognition frictions—one does not participate need for liquidity, capital markets provide in an activity if one does not comprehend it liquidity by allowing participants to unwind well—as well as collective action frictions, assets at limited cost; instead, banks offer which typically condition the setting up and deposits that can be redeemed on demand operation of multilateral arrangements.6 and at par. Market participants who wish to engage Financial structure is thus a snapshot, at in financial contracting must therefore find a given point in time, of the actual constel- the path of least resistance around these lation of financial services aimed at coping frictions and the associated market failures. with frictions.7 Financial development is, Once a decision is made to participate in in turn, the evolution of financial structure financial contracts, the private responses over time. An early resolution of collective to coping with frictions can be divided into action frictions (including through the intro- two subsets: responses aimed at lessening duction of central banks) allows for basic the frictions themselves (acquiring infor- payment and custody services. A steady mation, using collateral, delegating) and march from “relationship-based finance� to responses aimed at lessening the exposure “arm’s-length finance� ensues thereafter.8 At to these frictions (diversifying and pooling lower stages of FD, financial markets resolve F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E 19 agency frictions by relying on nontradable The patterns and immovable collateral and connections, that is, relationship-based transactions. As The gradual easing of agency and collective the informational and contractual environ- frictions provides broad pointers about the ment improves, private information becomes order in which various financial activities public and other types of collateral become are likely to emerge, and the shape of the available and tradable, gradually allowing paths they are likely to follow once they FD to break free from the tyranny of con- emerge. The order of appearance of financial nections. Similarly, as information becomes activities should reflect the intensity of the more abundant and governance arrange- frictions to which they are exposed. In addi- ments improve, screening and monitoring tion, since activities that exhibit the highest costs come down and lenders can increas- returns to scale should generally be the ones ingly rely on third parties (rating agencies, exposed to the highest collective frictions, market analysts, investment advisers, exter- scale effects should correlate with the order nal auditors); statistical methods (scoring in which financial activities appear and the systems, value-at-risk calculations); and shape (convex or concave) of their develop- accounting and disclosure standards. ment path. Of course, deviations from these T he gradual easing of agency fric- basic patterns may occur in response to tions helps boosts participation. In turn, policy choices, market demands, or linkages by unleashing positive network and scale between the development of various activi- externalities, the benefits of participation ties and instruments. (liquidity, efficiency, and so forth) become Although the time period for which data self-reinforcing. Thus, in the more mature are available is rather short, these patterns phase of FD, there is a quantum leap in can be broadly verified through econo- participation as more clients, players, and metric analysis.10 This is what de la Torre, transactions make markets increasingly Feyen, and Ize (2011) do, based on a broad deep, dense, and interconnected. The rising benchmarking exercise applied to a bat- participation also increases the degree to tery of financial development indicators.11 which financial institutions and capital mar- The results on three key patterns—order kets complement each other. The whole pro- of appearance, convexity, and returns to cess is quickened by financial innovation, a scale—are displayed in figure 2.2, where major driver of FD that reflects and chan- activities are ordered by the per capita nels the forces of competition, deregulation, income level at which they appear.12 Overall and regulatory and tax arbitrage, as well as results conform to what one would expect. theoretical or operational breakthroughs.9 Financial activities that are the least prone to Financial inclusion follows a process that frictions emerge and develop first. Activities largely parallels that of FD. Financial exclu- that are subject to strong frictions require sion is caused by a mix of agency frictions— more time. Some activities (such as debt in which the marginal investor or borrower and equity securities markets) are strongly faces stronger problems of information boosted by scale and network effects, which asymmetries or lack of collateral—and col- give rise to convexity in the shape of the lective frictions—which limit participation. paths after some threshold level of friction Should many marginalized investors or bor- reduction is reached. However, there are also rowers decide to participate, it could bring some important outliers—such as domestic the costs of participation down and allow public or private debt, for which the returns participation to increase and financial mar- to scale are not consistent with the order kets to deepen. Thus, financial inclusion will of appearance or the shape of development gradually expand as innovations or a better paths. The main stylized patterns (and pos- enabling environment gradually erodes both sible reasons for deviations from patterns) types of frictions. are discussed in what follows. 20 F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E FIGURE 2.2 Appearance, convexity, and returns to scale of FD indicator paths de lic pe deb ic as fund s ba miu nce nd ale an ba cred te s ion n lif atio et m ium e m sse s fo de ate de ate sit do a und c tu s o bl iva liz rk ub em an fu les es t ns t nk ts wh s ba ing cia nk it ca tock s ei n ut s es ts ig t bt iv riv po pu e ra ta a b re b nk m sti im pr sur np ba se pr pr lf pi m n su o np de io l in cla tic ua t nk tic in n ig re ife pr m fo nl do no �n appearance convexity returns to scale Source: de la Torre, Feyen, and Ize 2011. Note: This �gure represents three important characteristics of selected �nancial development (FD) indicators and their paths, namely, order of appearance, degree of convexity, and returns to scale. “Appearance� estimates the level of GDP per capita at which the FD indicator becomes relevant. Higher values (along the vertical axis) for the appearance line imply that the indicator starts to develop at higher levels of economic development. “Convexity� captures the speed at which the FD indicator grows as per capita income rises. Positive (negative) values imply that the indicator accelerates (decelerates) at higher levels of economic development. “Returns to scale� is the extent to which the FD indicator bene�ts from scale effects as derived, for example, from larger population sizes. Higher values for this variable imply that the FD indicator in question gains more from scale effects. Public debt markets. Domestic public debt follows an S-shaped path, reflecting solvency con- Emerging early in the game is government straints that eventually limit its size in rela- borrowing (figure 2.2), as governments are tion to GDP (figure 2.3). the first to overcome elementary agency frictions. Given the smallness of domestic markets at low levels of economic develop- Banking services ment, which reflects collective frictions, government borrowing takes place initially The next financial activity is banking ser- abroad and in foreign currency.13 As local vices (figure 2.2). Retail funding (bank markets develop and per capita income deposits) emerges before credit. Bank rises, external public debt declines and is deposit services initially respond to an early replaced by (mostly local currency) domes- need for simple custodial and payment ser- tic debt. In fact, domestic public debt devel- vices. Banks have a harder time lending than ops at a relatively earlier stage than would attracting funds. As intermediation evolves be consistent with the large returns to scale. from relationship-based lending to arm’s- This may be because governments are will- length finance, private credit rises along ing to pay a premium to meet their financ- a convex path—it catches up with depos- ing needs or because a government debt its over time and eventually exceeds retail market is a public good that helps conduct funding as wholesale (nondeposit) funding monetary policy or develop other financial makes up for the slack (figure 2.4). Lending F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E 21 to other financial institutions follows private FIGURE 2.3 Paths for government debt: External and domestic credit and is highly convex. These features are all related. As agency frictions ease up, 50 retail investors are increasingly able to shift into higher-yielding market instruments or 40 to have their funds managed by asset man- 30 variable’s path agers or institutional investors rather than 20 banks. At the same time, banks increasingly 10 lend, including to each other.14 The high con- 0 vexity of these activities reflects the exter- –10 nalities-laden reduction of collective action frictions associated with rising participation –20 (more players and the same players engaging –30 5 6 7 8 9 10 11 in more activity) and denser finance. log of GDP per capita outstanding domestic public debt securities/GDP Capital markets outstanding international public debt securities/GDP After banking is capital markets (figure 2.2). Source: de la Torre, Feyen, and Ize 2011. Private debt securities follow equity. The late appearance of capital markets and the strong convexity of their development path are, of game despite involving limited returns to course, clear manifestations of the complex- scale, which suggests that, unlike the case of ity of agency and collective frictions. Nota- public debt whose growth is primarily con- bly, private debt markets emerge late in the strained by critical mass effects, information FIGURE 2.4 Banking indicators’ paths: Retail and wholesale funding and private credit 70 60 50 variable’s path 40 30 20 10 0 4 5 6 7 8 9 10 11 log of GDP per capita private credit DMB domestic deposits nondeposit funding Source: de la Torre, Feyen, and Ize 2011. 22 F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E and enforcement frictions (rather than col- decisions concerning the organization of lective frictions) are the key binding con- the social security system. But a variety of straint.15 The fact that corporate bonds other public policies can affect the path of a develop after stocks is arguably because country’s FD, including the quality of poli- growing firms can initially substitute bank cies that promote the financial sector and, debt for market debt but will need to issue more generally, the quality and constraints stocks at some point no matter what.16 of economic management that, by affect- ing a country’s growth rate, will naturally affect its financial development. In effect, Institutional investors de la Torre, Feyen, and Ize (2011) find that They appear at different stages of the FD the quality of the informational and con- process (figure 2.2). Pension funds emerge tractual enabling environment, as well as early, mutual funds late. Insurance arises average GDP growth over the previous three somewhere in between, but nonlife insur- decades, are important determinants of the ance emerges earlier than life insurance. actual path of private credit and other key That pension funds appear early in the data FD indicators. reflects the key role played by recent pension A second reason for path uniqueness is reforms. Mutual funds appear late despite path dependence, which results from the not facing increasing returns to scale (that fact that output growth is itself a function of is, not being constrained by market size) financial development and that institutional because the availability of marketable assets rules and arrangements are self-reinforcing is a precondition for their development. The (North 1990). Because today’s FD depends later appearance of life over nonlife insur- on today’s output, which in turn depends ance is in part because the latter is influ- on yesterday’s FD, initial conditions matter. enced by policy—for example, mandatory Thus, FD trajectories can vary from country insurance for motor vehicles—while the to country, especially those financial ser- former is dependent on the development of vices that rely heavily on local institutions capital markets and, hence, is more sub- that are not tradable across borders. De la jected to collective action frictions.17 Life Torre, Feyen, and Ize (2011) find that con- insurance takes more time to bloom because tract enforcement institutions account for a of the need to invest in marketable assets significant share of the deviation of individ- and because of the need for life cohorts to ual paths from the cross-section benchmarks, interact intergenerationally; but once it does and that the payoffs from better contractual bloom, it follows a steeply convex path. institutions continue to rise with the level of financial and economic development. A third reason is leapfrogging. It is most The paths likely to result from financial innovations Financial development paths are likely to be that are transferable across borders; there- unique; that is, the lower-income countries fore, it affects those financial services that of today are unlikely to retrace precisely do not rely too heavily on (nontradable) local the path followed yesterday by the higher- institutions.18 For example, the dynamic income countries. There are a number of paths followed in figure 2.5, panel e, by reasons why this might be the case. A first banks’ net interest margins—which, except obvious, but still worth noting, reason is for the high-income countries, cross the country-specific policies. These could be cross-section path at a sharp angle—are policies that affect the financial architec- probably a reflection of such leapfrogging ture, such as public sector indebtedness— at work. Indeed, de la Torre, Feyen, and which mainly reflects fiscal policy—or they Ize (2011) find evidence that informational could be the assets of pension funds and improvements are easy to transfer across annuity providers—which reflect policy borders inasmuch as they rely on importable F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E 23 FIGURE 2.5 Financial depth indicators: Dynamic and cross-section development paths a. Private credit/GDP (IFS) b. Stock market turnover ratio (S&P) 110 160 stock market turnover ratio (filtered) 100 140 private credit/GDP (filtered) 90 120 80 100 70 80 60 60 40 50 20 40 0 30 –20 20 –40 4 5 6 7 8 9 10 11 4 5 6 7 8 9 10 11 log of GDP per capita log of GDP per capita c. Claims on the domestic nonmonetary d. Mutual fund assets/GDP (NBFI) financial sector/GDP (IFS) 35 claims on the domestic nonmonetary 100 mutual fund assets/GDP (filtered) financial sector/GDP (filtered) 30 90 25 80 70 20 60 15 50 10 40 5 30 0 20 4 5 6 7 8 9 10 11 4 5 6 7 8 9 10 11 log of GDP per capita log of GDP per capita e. Net interest margin (BANKSCOPE) f. Insurance premiums (LIFE)/GDP (AXCO) 0 6 insurance premiums/GDP (filtered) net interest margin (filtered) –2 5 –4 4 –6 3 –8 2 –10 1 –12 0 4 5 6 7 8 9 10 11 4 5 6 7 8 9 10 11 log of GDP per capita log of GDP per capita Asia Eastern Europe G7 LAC-7 other advanced economies spline fit Source: de la Torre, Feyen, and Ize 2011. technologies. Instead, contractual frictions 2.5, panel a, this was clearly the case of pri- cannot be solved by technological invest- vate credit in the low- and middle-income ments and imports.19 countries, which went through pronounced A fourth reason for path uniqueness is cycles during the period of observation. 20 financial cycles and crashes. Credit cycles Crashes can have an even stronger adverse can give rise to short-term deviations from impact on financial depth, especially credit trend growth paths. As illustrated in figure depth, from which banking systems in 24 F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E particular may take a long time to recover. (reflecting problems of governance), or De la Torre, Feyen, and Ize (2011) find that somewhere in between (reflecting prob- this is indeed the case. Volatility per se is lem s of i ncent ive s a nd “sk i n i n t he not a problem for FD; rather, the problem is game� at some level of the monitoring pyra- financial crashes. Remarkably also, the prob- mid). Failures to resolve participation fric- ability of crashes increases as private credit tions are, of course, a routine occurrence “outperforms� its cross-country aggregate in the less-developed financial systems benchmark. This is a noteworthy reminder and justify much of the state’s catalytic that financial development and financial sta- and financial infrastructure-building role. bility interact in complex ways, reflecting the However, they may also happen in well- ever-present duality between the bright side developed systems, particularly in the pro- and the dark side (see below). cess of spreading risk.22 A fifth and final reason is endogenous The second type of finance malady does quantum leaps in FD at the higher end of not come from the financial system’s inabil- the income scale. As illustrated in figure 2.5, ity to reduce frictions. Instead, it is the this is evident in the dynamic paths followed apparently successful financial development by a variety of financial depth indicators ensuing from a reduction of agency and par- that follow highly convex paths. For all these ticipation frictions—a process that is typi- indicators, the dynamic paths for rich coun- cally boosted by innovation—that itself may tries cut the aggregate cross-section bench- endogenously lead to problems of instability mark from below and rise along a steep and unsustainability, even in the absence of positive slope. Such an “explosion� of finan- government-induced social moral hazard. cial activity suggests that, once countries Thus, FD may often take place in a fragile or reach some stage of economic and financial even self-destructive mode that is in conflict development, the positive externalities of with financial stability. This endogenous out- increased participation “ignite� a second- come is what is described in this report as the ary chain reaction of financial activity, both dark side of finance. inside and across countries. Whether such The dark side of finance has three basic explosions are stable and sustainable is an modes. In the first mode, the successful eas- open question, given the maladies and dark ing of agency frictions is, paradoxically, at side of finance, as revealed in the following the source of the problem. It can trigger lethal section. 21 collective action failures, such as negative uninternalized externalities, free riding, or catastrophic coordination failures. For exam- The dark side ple, the availability of public information There are two types of finance maladies and the associated reduction in the ability to (that is, failures to complete markets and appropriate the rents from private informa- achieve efficient equilibrium). The first one tion encourage investors to free ride. Instead refers to failures to reduce the basic fric- of staying put and investing in analysis and tions that hinder FD. This may reflect an monitoring, they may rather invest short and inability to resolve agency frictions or col- rely on market liquidity to exit at the first sign lective (participation) frictions. Agency of possible trouble.23 But the easing of agency frictions and failures continue to dominate frictions can also trigger a second round of the literature on FD and account for most agency failures, much as building more high- of the policy debate, but collective fric- ways can exacerbate congestion by increasing tions and failures are at least as important. traffic. For example, as shown in the global Failures to reduce agency frictions may crisis, the rising reliance on third party moni- occur at the level of the investor (reflect- tors to reduce information frictions, coupled ing his inability to monitor or lack of inter- with the ease of exit, can give rise to a com- est in doing so), at the level of the borrower plex and opaque chain of transactions, where F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E 25 agents have little or no “skin in the game,� exuberance, even if the full implications of thereby creating a new generation of agency the innovations are not well understood. The frictions.24 mood swings first accentuate the upswing; In the second mode of the dark side, the once the market sentiment switches and successful easing of collective action frictions euphoria turns into despair, mood swings on account of rising participation is, again worsen the collapse. 26 paradoxically, what triggers the problems. All three types of maladies call for policy The positive externalities of increased market responses, which is, of course, the central participation in good times turn into crip- reason for discussing them in this report. pling negative externalities and other col- The agency frictions that are at the heart of lective action failures in bad times. While governance and asset management issues are market participation is a “win-win� situation covered in chapters 6–8. The state’s role in for all, as it enhances depth and illiquidity, resolving participation frictions and improv- market withdrawal in times of stress may be ing the spreading of risk is covered in chapter individually optimal but socially harmful as 9. The dark side underpins chapters 10–13, it originates fire sale spirals and self-fulfilling the four chapters that conform the prudential liquidity losses.25 oversight section of this report. In either case, problems arise from signifi- cant wedges between private and social costs and benefits, which markets on their own are Can there be too much �nance? simply unable to handle. Instead, the finan- Before closing this chapter, an important cial activities individuals engage in as self- additional feature of development paths protection for the good times no longer work is worth emphasizing. The counterpart of in the bad times. Individuals and individual convex FD indicator paths (for example, institutions do not internalize the adverse that of private credit to GDP) is that their social or systemic implications of their impact on real development (GDP growth) actions. For example, insuring oneself by must necessarily exhibit decreasing returns. selling risk to others (through, say, credit Unless the rate of growth of GDP keeps default swaps) can actually raise systemic accelerating as the level of GDP rises (a fragility through interconnected risk and clear dynamic impossibility), the impact of lead to contagion and accentuated down- finance on growth should necessarily level ward spirals in bad times, when default risk off at some point. This is exactly the conclu- becomes highly correlated and the value of sion that several recent papers reach when collateral collapses. Or else rising participa- regressing output growth against financial tion along the intensive dimension (the same depth indicators. 27 In itself, this conclusion intermediaries engaging in more transac- does not necessarily imply that there might tions and becoming financial giants) can be such a thing as “too much finance.� If boost social moral hazard by vastly increas- one reasonably assumes nonsatiation, more ing the social costs of individual institution finance should always be better. However, failures through the too-big-to-fail or too- unless kept in check by good public policy, interconnected-to-fail syndromes. the forces of the dark side of finance may But the dark side has yet a third mode in gather strength with FD, reflecting the ris- which the successful reduction of agency or ing potential for financial instability result- participation frictions leads to problems of ing from increasing interconnectedness. collective cognition. The bonanza associated If so, the marginal benefits of financial with FD feeds a collective mood of optimism development may at some point become that puts the system on a disequilibrium path. smaller than the marginal costs of maintain- This is typically amplified by the far-reaching ing financial stability. Should this occur, it effects of financial innovations (for exam- would, of course, throw a whole new light ple, securitization) that unleash bouts of on the FD policy debate. 26 F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E Notes composition) and degree of sophistication of markets and products (debt, equity, deriva- 1. The literature (particularly the empirical tives); leveraged intermediaries (commercial literature) on the process of �nancial devel- banks, investment banks, insurance compa- opment is relatively thin. That �nancial nies, hedge funds, dealers); non-leveraged-asset structure is shaped by the efforts of market managers (brokers, mutual funds, pension participants to circumvent and reduce the funds); and facilitators (accounting and audit- frictions that hinder �nancial contracting is ing �rms, rating agencies, investment consul- hardly a new concept (see Merton and Bodie tants, mortgage originators). 2004). However, few have attempted to ana- 8. See Rajan and Zingales (2003). lyze these frictions systematically in terms of 9. Indeed, the history of �nancial develop- how they interact and what this may imply ment is marked by major waves of innova- for the dynamics of FD. Interestingly, more tion. Consider, for instance, the role in the work has been done on measuring the impact exponential ascent of �nance in the west- of FD on economic growth than on exploring ern world, stemming from the invention of how FD is affected by economic growth (see Italian banking (based on trade-related bills Beck and Levine 2005). When discussing the of exchange) by the Medici in the late 14th impact of �nancial structure on growth, the century; or the introduction of payments sys- literature, at least until very recently, gener- tems based on checking accounts, fractional ally concluded that function matters more reserve banking, and central banking during than form (Demirgüç-Kunt and Levine 1999; the 17th century; or the development of the Allen and Gale 2000). However, more recent government bond market, its seeds already vis- papers (such as Demirgüç-Kunt, Feyen, and ible in the late Middle Ages; or the invention of Levine 2011) have come closer to recogniz- the joint-stock, limited liability company in the ing that there might be such a thing as an early 17th century and the associated mush- “optimal� �nancial structure, that is, that rooming of stock exchanges; or the emergence form might also matter. of marine insurance and life insurance in the 2. This chapter draws heavily on the paper second half of the 17th century; or, in the lat- “Financial Development: Structure and ter part of the 20th century, the development Dynamics� by Augusto de la Torre, Erik of securitization and derivative products. For Feyen, and Alain Ize. an insightful and entertaining rendition of 3. For more details on the paradigms, see de la the history of �nance in the Western world, Torre and Ize (2010, 2011) and de la Torre, see Ferguson (2008). For a recently updated Feyen, and Ize (2011). review of the roots and dynamics of �nancial 4. This strand of thought follows the insights innovation, see Lerner and Tufano (2011). of Spence (1973), Akerlof (1970), and, espe- On the role of competition and deregula- cially, Stiglitz and Weiss (1981). tion in FD, see Rajan and Zingales (2003). 5. This strand of thought is in line with the Examples of theoretical and methodological insights of Holmstrom and Tirole (1998) and breakthroughs that have dramatically influ- Geanakoplos (2009), among others. enced FD include double-entry bookkeep- 6. The collective action frictions of the bright ing, probability theory, life expectancy tables side involve uninternalized positive externali- and actuarial science, and the Black-Scholes ties and coordination failures that prevent option theory. agents from moving to a superior equilib- 10. Remarkably, the sequence derived from this rium where everybody would be better off. cross-section analysis broadly matches that The collective action frictions of the dark found in individual countries through histori- side (discussed below) involve uninternalized cal studies of the process of �nancial develop- negative externalities, free riding, or coor- ment. The literature on the history of �nance dination failures that prevent agents from in the western world is vast. See Ferguson avoiding falling into an equilibrium that is (2008) and Rajan and Zingales (2003). worse for the group as a whole, albeit pos- 11. The authors use a worldwide �nancial data- sibly better for some. base put together on the basis of publicly 7. The shape and nature of �nancial structure is available data by the World Bank (FinStats given by the mix (that is, relative weights and 2009). It covers 40 key FD indicators for F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E 27 a large number of countries for the period 14. The high convexity of wholesale funding 1980–2008 (country coverage and qual- and interbank lending can be viewed as the ity varies). The benchmark paths are esti- growth analogue of the rapid rise of whole- mated by controlling not just for per capita sale funding and bank interconnectedness. income, as is traditionally done, but also for See Shin (2010). various factors that are exogenous both to 15. Thus, while market liquidity is crucial to the policy and to the FD process itself, including development of public sector bond markets, population size and density, age dependency it is much less relevant to the development of ratios, and certain country-speci�c charac- corporate bond markets. teristics (such as the condition of being an 16. The fact that equity markets have an unlim- offshore �nancial center or a major fuel ited upside may also help explain their ear- exporter). The paths are estimated using lier appearance, even under high agency quantile regressions so as to lessen the effects frictions. of outliers. To dampen the ampli�cation 17. The abnormally low (negative) returns to effects of �nancial bubbles and downward scale of casualty insurance reflect the pre- spirals, rather than performing a traditional dominance of foreign trade insurance in panel estimate—which would blend varia- the small open economies. It accounts for a tions across countries and across time—the disproportionately high share of total casu- method used takes the medians of a given alty insurance. See Feyen, Lester, and Rocha �nancial development indicator for each (2011). country and over the entire sample period 18. One may consider, for example, the cases and estimates a cross-section regression of credit card services and e-banking. These over those country medians. This becomes services are now found in most developing the benchmark (cross-section) path against countries, and although they cover a smaller which one can compare the paths followed fraction of the adult population, their func- by individual countries or country group- tionality and quality are comparable to those ings over time. Some of these comparisons in developed countries. In both cases, devel- are reported in chapter 3 of this book. oping countries have been able to leapfrog 12. The level of gross domestic product (GDP) because the associated technology is rela- per capita at which �nancial services start tively easy to import and adapt and because to appear is measured by the intercept of the the services do not depend heavily on local cross-section paths with the horizontal axis. contractual institutions. To limit lower tail distortions when a nonlin- 19. De la Torre, Feyen, and Ize (2011) �nd the ear �t is imposed on the data, for these esti- quality of the informational environment mates we use only a linear per capita GDP (credit information) to be concave with term. Convexity is measured by the coef- respect to income but subject to scale effects. �cient of quadratic per capita GDP when In contrast, the quality of the contractual �nancial indicators are regressed against both environment (creditor and property rights) is per capita GDP and its square. Scale effects convex but not subject to scale effects. This are measured by the coef�cient of population suggests that informational frictions are of size in the controlled regressions. a mostly technological nature. Because of 13. This is, of course, the basic premise of the �xed costs, they are easier to implement in “original sin� literature, which focuses on the larger countries. Furthermore, they are solv- inability of emerging economy sovereigns and able with adequate investments and ready- corporates to issue long-term debt denominated made imports from abroad. Moreover, once in their domestic currency. For the relevant the required investments are in place, there analysis and suggestions for “redemption,� are decreasing returns to further informa- see Eichengreen and Hausmann (1999, 2002); tional improvements as developmental lev- Calvo and Reinhart (2002); Eichengreen, els rise. In contrast, contractual frictions Hausmann, and Panizza (2005). The fact that cannot be solved by technological invest- the better foreign institutional framework facil- ments or imports. They are mostly institu- itates enforcement is an important component tional and reflect collective action frictions of the “original sin� story (see de la Torre and that are trickier to resolve, no matter how Schmukler 2004). large the country. And the payoffs from 28 F I N A N C I A L D E V E LO PM E N T: B R I G H T S I D E, PAT T E R N S, PAT H S, A N D DA R K S I D E better institutions continue to rise with the on �nancial crises (see Minsky 1975). More level of economic development. recently, it was popularized by Kindleberger 20. Although �gure 2.5 shows GDP per capita, (1989) and Shiller (2006). rather than time, on the horizontal axis, the 27. Deidda and Fattouh (2002) �nd that FD two variables are highly correlated because has a positive but statistically insigni�cant GDP per capita generally grows over time. impact on growth in countries with low lev- Thus, the steep up and down movements of els of economic development and a positive private credit in Eastern Europe and the LAC and statistically signi�cant impact on growth region that appear in �gure 2.5 can readily be in countries with higher levels of economic interpreted as pronounced dynamic cycles. development. Rioja and Valev (2004) �nd 21. Haldane (2010a) and Turner (2010) have sug- that there is no statistically signi�cant rela- gested that the precrisis explosion of �nance tionship between �nance and growth at low in the high-income countries was driven by levels of FD, that there is a strong positive precisely such dynamics. relationship at intermediate levels of FD, 22. This is the main theme of chapter 9 of this and that there is a weaker but still positive report. effect at higher levels of FD. Arcand, Berkes, 23. Huang and Ratnovski (2011) show that the and Panizza (2011) �nd that �nance actu- dark side of bank wholesale funding domi- ally starts having a negative effect on out- nates when bank assets are more arm’s length put growth when credit to the private sec- and tradable. tor exceeds 110 percent of GDP. This result 24. See Ashcraft and Schuermann (2008) and is congenial to that in de Gregorio and Gorton and Metrick (2010). Guidotti (1995), who found that in high- 25. See Shleifer and Vishny (2011). income countries FD was positively corre- 26. The importance of mood swings for �nancial lated with output growth during 1960–85 bubbles and panics �nds its roots in Keynes’s but that the correlation was negative in the animal spirits and Hyman Minsky’s writings 1970–85 subperiod. Domestic Financial Development: Where 3 Does LAC Stand? T his chapter describes the state of Latin (as measured by interest rate margins) of America and the Caribbean (LAC) banking intermediation, the liquidity region’s domestic financial develop- of the domestic equity market, and the ment.1 It provides systematic evidence on depth of insurance products. how LAC’s financial sector has evolved and • These gaps are of concern because they where it stands relative to the past and to coincide with some of the � nancial indi- other regions. The analysis focuses on finan- cators that have been shown to be the best cial deepening and the provision of financial predictors of future output growth and services. Aspects related to soundness and because, except for bank margins, there is stability—though clearly linked to financial little evidence of convergence toward their deepening, as discussed in chapter 2—are “benchmark� levels, as determined by the covered in chapters 6–10 rather than here. economic development of the region and The state of financial globalization and its basic structural characteristics. financial inclusion, two key complementary • Other areas of concern include the stub- aspects of financial development, are dis- bornly high concentration of issues and cussed in the next two chapters. Main find- trading in the bond and equity mar- ings are as follows: kets, and the limited capacity of insti- tutional investors to expand their port- • LAC financial systems have deepened folios beyond the safest and most liquid substantially over the past two decades, investments. in ways that conform to the sequencing • There is also substantial heterogene- and paths described in chapter 2. ity in � nancial development within the • In particular, bonds and equity markets region, with the smaller, lower-income, have gained ground, institutional inves- non-LAC-7 countries generally lagging tors now play a central role, new markets LAC-7. 2 and instruments have sprung up, maturi- • However, many cases of successful devel- ties have lengthened, and dollarization opment, such as Brazil’s equity market has been reduced. and Chile’s bond market and annuities • However, signi�cant developmental gaps industry, may be edifying examples for remain, including the depth and ef�ciency other countries in the region. 29 30 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? The rest of the chapter is structured as fol- by gross domestic product (GDP) to enhance lows: the first section presents the methodol- comparability. The second is based on a sta- ogy; the second section synthesizes the main tistical benchmarking exercise that controls, findings; the next three sections focus on in addition to GDP, for a number of factors banking, bond markets, and equity markets, that are exogenous to financial development, respectively; the sixth section discusses the thus allowing a systematic identification of new players; and the final section considers developmental gaps.4 To provide a more com- alternative markets and products. plete perspective, the results of both types of analysis are presented side by side whenever possible. Methodology This chapter shows evidence on the bank- The chapter focuses on LAC-7, although, ing sector, nonbank financial institutions, wherever possible, comparisons with group- and capital markets. The analysis of capi- ings of the smaller LAC economies are tal markets is particularly relevant because provided, namely, the Caribbean countries, many of the recent reforms were thought Central America (including the Dominican to increase competition with banks, reduce Republic), offshore centers (The Bahamas, the cost of capital, and increase access to Barbados, and Panama), and South Ameri- finance. Furthermore, the expectation was can countries. 3 Other emerging regions that capital markets would take off, as and countries, such as Asia, China, Eastern banks have been an important player for Europe, and India, are used for peer group a longer time. In particular, the chapter comparisons. The G-7 countries and other documents basic trends in bond (corporate developed economies provide broader points and government) markets and equity mar- of reference (see table 3.1). kets. It also provides some information on To shed light on how funds are interme- the development of other—newer and less diated and how broadly they are used, this core—markets and instruments, such as chapter provides evidence from both sides of derivatives, securitization, private equity, the market for lendable funds and focuses and credit by retail chains. In addition, it on markets as well as financial institutions. documents the evolution of other large It also investigates how financial activity has financial players, such as pension funds, been evolving qualitatively (currency, matu- mutual funds, and insurance companies. rity, borrower composition, and so on). Two Because the developmental efforts aimed types of cross-country comparisons are con- at completing financial systems started in ducted. The first provides comparisons of dif- earnest in the early 1990s, the focus is pri- ferent dimensions of market activity over time marily on the evolution of financial markets and across regions, scaling depth indicators from that period forward. TABLE 3.1 Countries analyzed, by region Asia Indonesia, Republic of Korea, Malaysia, Philippines, and Thailand Eastern Europe Croatia, Czech Republic, Hungary, Lithuania, Poland, Russian Federation, and Turkey G-7 Canada, France, Germany, Italy, Japan, United Kingdom, and United States Other advanced economies Australia, Finland, Israel, New Zealand, Norway, Spain, and Sweden LAC LAC-7 Argentina, Brazil, Chile, Colombia, Mexico, Peru, and Uruguay South America Bolivia, Ecuador, Paraguay, and República Bolivariana de Venezuela Central America Belize, Costa Rica, Dominican Republic, El Salvador, Guatemala, Honduras, and Nicaragua The Caribbean Jamaica, and Trinidad and Tobago Offshore �nancial centers Aruba, Bahamas, Barbados, Bermuda, Cayman Islands, Netherlands Antilles, and Panama Source: Authors’ compilation. Note: This table lists all the countries analyzed for each region. Some of these countries might not be used in certain tables and �gures because of data availability. DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 31 Main �ndings The nature of financing is also improving to some extent—for example, the maturity In many respects and by several standard structure of both private and public bonds measures, LAC financial systems have is longer—but at a slow pace. In addition, developed and deepened over the past two the dollarization of loans and bonds has decades, much along the sequence predicted declined. in chapter 2. Thus, their overall structure Despite this general deepening, LAC’s has become more similar to that of devel- financial systems show significant develop- oped countries, with bond and equity mar- mental gaps over a number of dimensions. kets gaining ground vis-à-vis the banking Even more surprising, they do not demonstrate sector. At the same time, institutions and a convergence toward the indexes of finan- markets are becoming more complete and cial development observed in the more devel- interconnected, and new markets are tak- oped countries, or for that matter, toward the ing off, albeit somewhat timidly. Nonbank benchmark levels one would expect to find institutional investors now play a much in countries with LAC’s characteristics. 5 more central role (particularly in the LAC Figure 3.1 presents a synoptic view of the countries that have introduced Chilean- total size and composition of financial sys- style pension reforms), and the number and tems. LAC’s financial assets, relative to GDP, sophistication of participants are increasing ranked very close to last among countries (even without taking into account the grow- examined for 2000–09.6 The underperfor- ing participation of cross-border investors). mance comes mainly from the very limited FIGURE 3.1 Domestic �nancial systems 350 300 89 250 65 95 200 63 % of GDP 65 104 73 49 150 36 92 59 61 56 13 26 52 100 36 8 48 20 43 35 31 128 30 17 123 33 50 95 97 20 105 83 84 20 80 51 50 60 36 39 44 0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 90 00 90 00 90 00 90 00 90 00 90 00 90 00 19 20 19 20 19 20 19 20 19 20 19 20 19 20 Asia (5) China Eastern G-7 (5) India LAC-7 (6) other Europe (2) advanced economies (4) banks bonds equities Source: Didier and Schmukler 2011a. Note: This �gure shows the average size and composition of domestic �nancial systems during the 1990s and 2000s. The �gure shows total banking claims, outstanding bonds, and equity market capitalization as a percentage of GDP. Numbers in parentheses show the number of countries in each region. 32 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? bank assets, which—at 44 percent of GDP just below 1 percentage point in the 2000s during the 2000–09 period—are below the decade, down from over 3 percentage points levels found in any other region. during the previous decade), they remain sig- The underdevelopment of LAC’s finan- nificantly above their benchmarks (figure 3.2 cial systems cuts across many indicators. As and table 3.2). LAC-7 banks are also becom- regards banking, private credit underperforms ing more concentrated, which contrasts with relative to benchmark and, although less now trends in other regions, and LAC-7 systems than in the past, so do bank margins. Pub- show the highest penetration of foreign lic bond markets have expanded, but not as banks (figure 3.4). On the other hand, there fast as those in the rest of the world. Private has been an important de- dollarization of bond markets have increased in size, and pri- bank loans (figure 3.3, panel b).7 mary markets have somewhat taken off, but On a country-by-country basis, Chile is markets remain relatively small and illiquid. the only LAC-7 country that meets its bench- Equity markets are approximately on track mark as regards private credit to GDP. All as regards capitalization but lag very substan- other LAC-7 countries are widely below tially as regards liquidity; moreover, they con- benchmark (figure 3.5). When one compares tinue to be highly concentrated in a handful subregions within LAC, the Caribbean coun- of large firms. The development of the insur- tries and Central America show trends simi- ance industry continues to lag its benchmark. lar to those in LAC-7, though South America While institutional investors are sophisticated clearly lags behind its neighbors. However, and large, they channel a significant portion offshore centers are clear outliers, showing an of their funds to governments or banks instead impressive, almost twofold growth in bank- of firms. Furthermore, there is a large hetero- ing depth between the 1980s and the 2000s geneity within the region, with LAC-7 coun- (figure 3.6). tries being substantially more developed than the rest. There is also heterogeneity within LAC-7 itself. For example, Brazil and Chile Bond markets show notable progress in particular areas While bond markets in LAC-7 have grown (the equity and bond markets, respectively), significantly, total bond capitalization, as which, though incomplete, look encourag- well as new issues per year, remained the ing and might be used as exemplars for other lowest in 2000–09 among all the peer coun- countries in the region. The rest of this chap- tries (figure 3.7, panels a and b). Indeed, ter reviews these trends in more detail. statistical benchmarking confirms that both the public and private bond markets con- tinue to lag their benchmarks, albeit public Banking bond markets have made up for a substantial Banking systems in LAC lag behind their share of the slack over the last decade (table benchmarks, whether in terms of deposits or 3.2). Moreover, lack of liquidity remains a credit to the private sector, and this is espe- key concern, with bond turnover being the cially so in LAC-7. The lags are very large lowest of all the peer countries (figure 3.8). (about 20 percent of GDP for deposits and This suggests that primary bond markets 22 percent of GDP for credit) and show no have developed substantially more than sec- sign of catching up any time soon (figure 3.2 ondary markets, perhaps in part because and table 3.2). Reflecting the rapid growth institutional investors tend to buy and hold of personal lending over the past decade, the (see chapter 6). lags appear now to be mostly concentrated in These rather negative broad trends hide commercial lending and, even more starkly, important success stories in particular coun- in mortgage lending, which is considerably tries, however. For example, Brazil and Mex- lower than in any other region (figure 3.3, ico, and to a lesser extent Colombia, have panel a). Moreover, while bank intermedia- large and liquid public debt markets with tion margins have narrowed substantially (to reliable yield curves that can be used to price DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 33 FIGURE 3.2 Banking indicators relative to global benchmarks a. Domestic deposits as a percentage of GDP 40 140 130 domestic deposits/GDP (residuals) 30 domestic deposits/GDP (�ltered) 120 20 110 100 10 90 0 80 –10 70 60 –20 50 –30 40 4 5 6 7 8 9 10 11 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 20 20 20 20 20 log of GDP pc b. Private credit as a percentage of GDP 50 140 40 120 private credit/GDP (residuals) private credit/GDP (�ltered) 30 20 100 10 80 0 –10 60 –20 40 –30 –40 20 4 5 6 7 8 9 10 11 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 20 20 20 20 20 log of GDP pc c. Net interest margin 8 0 6 –2 net interest margin (residuals) net interest margin (�ltered) 4 –4 2 –6 0 –8 –2 –10 –4 –12 4 5 6 7 8 9 10 11 91 93 95 97 99 01 03 05 07 19 19 19 19 19 20 20 20 20 log of GDP pc Asia Eastern Europe G-7 Asia Eastern Europe G-7 LAC-7 other advanced economies LAC-7 other advanced economies spline fit Source: de la Torre, Feyen, and Ize 2011. Note: This �gure shows bank domestic deposits as a percentage of GDP, bank private credit as a percentage of GDP, and bank interest margins across country groups relative to their global benchmarks. The left charts show the evolution of annual median residuals by country group. First, a cross-sectional regression model is estimated on country averages, controlling for GDP per capita, population size and density, young and old-age dependency ratios, a �nancial offshore center dummy, a transition country dummy, and a large fuel exporter dummy. Annual country residuals are derived from the cross-sectional estimates, which are used to calculate annual country-group medians. The right charts plot annual country-group medians of the �ltered variables against the logarithm of GDP per capita. The �ltering consists in removing from the actual annual country observations the expected contributions of all factors except GDP per capita, using the cross-sectional estimates from the above model. A four-knot spline approach is used to account for nonlinearities of GDP per capita in the cross-sectional regression. private fixed income issues and value portfo- h). Also, Brazil, Mexico, and Chile all meet lios and to help develop derivatives markets or exceed their benchmarks as regards the (figure 3.7, panel c). Indeed, domestic public capitalization of their private domestic bond debt capitalization exceeds its benchmark in markets (figure 3.5, panel g). Chile has a both Brazil and Colombia (figure 3.5, panel particularly well-established private bond 34 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? TABLE 3.2 Benchmark model for LAC’s financial development indicators Median actual values (%) Workhorse median residuals Rest of LAC LAC-7 Rest of LAC LAC-7 2000–08 2000–08 1990–99 2000–08 1990–99 2000–08 Bank private credit 36.0 24.2 –3.9*** –0.8** –13.6*** –22.5*** Bank claims on domestic �nancial sector 1.1 2.6 –1.3*** –0.4*** –1.1*** –0.2 Bank credit to government 3.7 10.0 –5.2*** –5.3*** –4.3*** –1.1 Bank foreign claims 8.6 2.5 –0.5 2.0 –4.4*** –5.7*** Bank domestic deposits 37.4 25.4 –10.9*** –4.0*** –13.6*** –20.8*** Bank nondeposit funding 18.4 24.3 –3.1 –1.4 –5.1* –6.5** Net interest margin 4.9 4.8 0.0 1.0*** 3.3*** 0.9*** Noninterest income/total income 25.6 33.5 –9.4*** –4.2*** 3.9 1.9** Total bank �nancial assets/GDP (excluding reserves) 55.5 65.5 –17.6*** –0.7 –16.7*** –19.5*** Life insurance premiums 0.3 0.7 –0.3*** –0.3*** –0.7*** –0.4*** Nonlife insurance premiums 1.3 1.1 –0.1 0.1*** –0.2*** –0.3*** Pension fund assets 7.5 11.7 –4.3*** –0.7* Mutual fund assets 1.1 5.9 –9.8** –5.9 Insurance company assets 2.1 4.0 –2.9*** –7.7*** Stock market turnover 2.3 12.6 –3.7 –10.6*** –18.4*** –28.8*** Stock market capitalization 15.7 33.6 –8.5 –7.7*** –7.7 –0.6 Domestic private debt securities 0.6 9.0 –10.1 –12.3 –1.6*** –3.0* Domestic public debt securities 28.5 19.7 –17.6** 0.7 –21.3*** –12.2*** Bank foreign claims 17.2 24.5 –2.5** –3.4 –4.0 0.0* Foreign private debt securities 1.8 4.0 –1.6* –2.4** –1.9** –2.5*** Foreign public debt securities 9.6 8.9 –3.0 1.3** 0.9 2.6*** Gross portfolio equity assets 0.2 3.1 –4.9*** –0.6*** –0.1 2.4 Gross portfolio debt assets 3.7 1.3 –1.0 1.2 0.1 0.0** Gross portfolio equity liabilities 0.4 5.4 –2.3* –2.2*** 1.2** 1.8 Gross portfolio debt liabilities 9.2 12.2 16.1*** 2.7 –4.9 3.2** Capital/total assets 10.1 11.5 –0.2** 0.5** 0.6*** 2.3*** Liquid assets/total assets 3.6 12.5 –4.8*** –5.9*** 2.4*** 5.0*** Regulatory capital/RWA (risk weighted assets) 14.1 14.4 0.1 2.0*** Bank capital/assets 10.0 9.8 0.7 2.2*** ROA (return on assets) 1.5 1.3 –0.4 0.1 0.4*** 0.5** Source: de la Torre, Feyen, and Ize 2011. Note: The table shows the results of a benchmark model for LAC’s �nancial development indicators. It presents the 2000–08 median for all LAC countries and the median LAC residual for the 1990–99 and 2000–08 periods, respectively, derived from the workhorse median regression model of the �nancial indicator of interest on GDP per capita (squared), population size and density, fuel exporter dummy, age dependency ratio, offshore �nancial center dummy, transition country dummy, and year �xed effects. The asterisks correspond to the level of signi�cance of Wilcoxon rank sum tests for distributional differences of the residuals between LAC and rest of the world. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent market that can be used as an important Also worth noting on the positive side, the point of reference for other countries in the maturity and currency profiles of bonds in region (see box 3.1). Together with macrosta- the LAC-7 countries have improved consider- bility and the withdrawal of the public sector ably.9 The average maturity of private sector from debt markets, this illustrates the major bonds increased by about a year between the role institutional investors can play in pro- 1990s and the 2000s (figure 3.9, panel a); that moting market development. The more recent of public sector bonds by 35 months between but quite rapid development of the corporate 2000–03 and 2008–09 (figure 3.9, panel b). bond market in Brazil is another interesting At the same time, the share of local bonds regional experience worth looking at.8 (corporate and government) denominated in DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 35 FIGURE 3.3 Bank credit a. Composition of private bank credit 100 11 10 10 9 9 8 19 17 17 18 18 14 90 22 24 26 80 19 70 14 43 43 42 34 43 47 47 14 % of total credit 60 37 40 51 49 58 50 40 66 62 30 60 49 45 48 48 50 20 40 41 42 43 31 34 10 25 0 3 7 9 3 7 9 3 7 9 3 7 9 3 7 9 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 00 04 08 00 04 08 00 04 08 00 04 08 00 04 08 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 China Eastern Europe (2) G-7 (2) LAC-7 (6) other advanced economies (3) commercial mortgage personal b. Loan dollarization 60 54 50 47 45 45 40 37 % of total loans 33 30 26 23 20 19 20 18 18 9 8 9 10 5 5 5 0 ) a 4) 3) 5) ies ced (1 in e( 7( 7( Ch ) ia om an (2 op G- C- As on dv LA ur ec er a r nE h te ot s Ea 2000–03 2004–07 2008–09 Source: Didier and Schmukler 2011a. Note: This �gure shows in panel a the average share of commercial, mortgage, and personal credit as a share of total banking credit. Panel b shows the extent of dollarization of banking loans. Foreign-currency-denominated loans are shown as a share of total loans averaged between 2000 and 2009. Numbers in parentheses show the number of countries in each region. (or linked to) foreign currency has declined Equity markets significantly, from 33 percent during the 1990s to about 25 percent in the 2000s in Equity market capitalization in LAC-7, as in the case of corporate bonds (figure 3.10). most emerging economies, has grown rapidly 36 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.4 Foreign ownership and concentration of banking systems a. Number of foreign-owned banks as a percentage of the total number of banks 60 49 50 46 43 % of total number of banks 40 30 30 25 20 22 20 20 17 15 16 12 10 0 ) pe rn 4) 7) ies ced a (4 in 7( 7( ro ste ) ) ia Ch om an (7 (4 G- C- As Eu Ea on dv LA ec er a h ot 1999–99 2000–08 b. Loans by the top five banks as a percentage of total loans 90 82 80 76 74 69 70 % of total loans 64 65 66 63 60 55 54 50 46 44 45 42 40 30 ) pe rn 7) a 6) ies ced a (4 di in 7( 7( ro ste In ) ) ia Ch om an (6 (4 G- C- As Eu Ea on dv LA ec er a h ot 1999–03 2004–07 Sources: Claessens and van Horen 2009; Didier and Schmukler 2011a. Note: This �gure shows bank foreign ownership and concentration. Panel a shows the average number of foreign-owned banks as a share of the total num- ber of banks between 1995 and 2008. Panel b shows the annual average of total credit granted by the top �ve banks as a share of total credit between 1999 and 2007. DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 37 FIGURE 3.5 Total banking assets as a percentage of GDP, within LAC 100 90 87 80 70 65 60 % of GDP 50 48 40 39 40 39 40 37 38 34 34 28 29 28 30 21 20 10 0 lic the ) 6) ) ) (2 (3 (4 7( 7) an C ica ub es C- LA )( be er ep lud LA in m ib n R inc hA rs r Ca te ica ( ut in ica en So m er ec Do l Am or sh ra off nt Ce 1980–89 1990–99 2000–09 Source: Didier and Schmukler 2011a. Note: This �gure shows average total banking claims as a percentage of GDP between 1980 and 2009 across LAC. Numbers in parentheses show the number of countries in each region. during the past decade (figure 3.11, panel a). worsens further when one looks at domes- The statistical benchmarking exercise (table tic stock market turnover, where the LAC-7 3.2) confirms that the LAC-7 group is on countries lag dramatically behind the global track regarding equity market capitaliza- benchmarks (by about 29 percentage points), tion (instead, the rest of LAC lags its bench- with the lag growing over time (figure 3.13 mark). However, when adjusting for changes and table 3.2). Overall, this suggests, some- in equity prices, one observes a much more what disturbingly, that the more the LAC-7 modest expansion (figure 3.11, panel b). equity markets have developed, the more Consistent with the fact that the increase in their liquidity has dried up.10 While the lag market capitalization is mostly explained in turnover is somewhat less severe in the rest by valuation adjustments, there has been a of LAC than in LAC-7, the same story applies decline in capital-raising activity that suggests for the rest of LAC. a declining role for equity financing, with Access to the equity market also remains equity issuance actually declining by about limited and concentrated in a few firms. The 70 percent between the 1990 and the 2000 number of listed firms in LAC-7 markets is decades (figure 3.12, panel a). The picture small and has declined over the past decade 38 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.6 LAC-7 �nancial indicators against benchmark a. Private credit b. Life insurance premiums 2.5 120 100 2.0 80 1.5 % of GDP % of GDP 60 1.0 40 0.5 20 0 0.0 a zil ile a ico ru y a zil ile a ico ru y ua ua in bi in bi Pe Pe a a Ch Ch ex ex nt nt m m Br Br ug ug lo lo ge ge M M Ur Ur Co Co Ar Ar c. Pension fund assets d. Mutual fund assets 70 45 60 40 35 50 30 % of GDP 40 % of GDP 25 30 20 15 20 10 10 5 0 0 a il ile a ico ru y a il ile a ico ru y ua ua az az in bi in bi Pe Pe Ch Ch ex ex nt nt m m Br Br ug ug lo lo ge ge M M Ur Ur Co Co Ar Ar e. Stock market capitalization f. Stock market turnover ratio 120 90 80 100 70 80 60 % of GDP % of GDP 50 60 40 40 30 20 20 10 0 0 a il le a o ru y a il ile a o ru ay ua az az tin bi in bi ic ic i Pe Pe u Ch Ch ex ex nt m m Br Br ug ug n lo lo ge ge M M Ur Ur Co Co Ar Ar g. Outstanding domestic private debt securities h. Outstanding domestic public debt securities 25 50 45 20 40 35 15 30 % of GDP % of GDP 25 10 20 15 5 10 5 0 0 a il ile a ico u y a il ile a ico ru y ua ua az az tin bi in bi r Pe Pe Ch Ch ex ex nt m m Br Br ug ug n lo lo ge ge M M Ur Ur Co Co Ar Ar Source: de la Torre, Feyen, and Ize 2011. Note: This �gure shows �nancial indicators for individual LAC-7 countries against their respective benchmarks (shown as bars). DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 39 (figure 3.12, panel b). The number of firms FIGURE 3.7 Primary bond markets using equity finance on a regular basis is also comparatively small, with only eight firms issu- a. Composition of bond markets ing equity in any given year during the 2000s 120 (figure 3.12, panel c).11 Moreover, not only do 100 few firms access equity markets on a regular 65 80 basis, but also the bulk of equity financing is % of GDP 52 concentrated in few issues. The share of the 60 24 31 top five issues has actually increased from 40 32 78 percent to 92 percent between the 1990s 20 20 23 40 41 48 23 40 33 30 and the 2000s (figure 3.14, panel a). Lastly, 16 24 4 13 14 19 15 10 4 4 5 trading in equity markets is also highly con- 0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 centrated, with the top five firms capturing 90 00 90 00 90 00 90 00 90 00 90 00 90 00 19 20 19 20 19 20 19 20 19 20 19 20 19 20 almost 60 percent of trading (figure 3.14, Asia (5) China Eastern G-7 (7) India LAC-7 (6) other Europe (2) advanced panel b). Again, there is also wide hetero- economies(5) private bonds public bonds geneity across the region, with non-LAC-7 b. Total amount of new issues per year as a percentage of GDP in the private sector countries having generally tiny and illiquid 5.0 4.7 markets—with fewer than 50 listed firms on 4.5 average and turnover rates below 5 percent 4.0 of GDP (figure 3.15). In sum, equity markets 3.5 3.4 across the region remain, by and large, small, 3.0 % of GDP illiquid, and highly concentrated. 2.5 That being said, not all is gloom and 2.0 1.6 1.7 doom. Within LAC, the Caribbean and off- 1.5 1.2 1.1 1.1 1.0 0.8 shore centers show considerably larger and 0.5 more rapidly increasing market capitaliza- 0.0 tions (figure 3.16). Within LAC-7, in terms ) 7) 7) ies ced (5 7( 7( ) ia om an (7 G- C- of equity market capitalization, Chile clearly As on dv LA ec er a stands out, followed by Peru and Brazil (fig- h ot 1991–99 2000–08 ure 3.5, panel e). In terms of domestic equity market turnover, except for Uruguay (which c. Total amount of new issues per year as a percentage of GDP in the public sector 16 just meets its benchmark), all LAC-7 coun- 14.8 14 tries fall short of their benchmarks. How- 13.2 12.9 11.8 ever, from that group, Brazil is the country 12 10.8 that gets closer to meeting it. This relative 10 % of GDP 8.5 success is worth noting (see box 3.2). In 8 addition to better governance (boosted by 6 the Novo Mercado), it might reflect the 3.7 4.4 3.9 4 large size and diversity of the private corpo- 2.0 2.9 1.8 2 1.4 rate sector and the large size of the mutual 0.5 0.7 0.4 0.6 0.5 0.6 0.1 fund industry.12 0 a il ile a ico ru y ua az in bi Pe Ch ex nt m Br ug lo ge M Ur Co Ar The new players 2000–03 2004–07 2008–09 LAC’s financial systems have also become Source: Didier and Schmukler 2011a. more complex from the saver’s perspective. Note: Panel a shows the average size of private and public bonds outstanding in domestic markets as a percentage of GDP between 1990 and 2009. Domestic bonds are de�ned as those issued by While in the past banks captured most of residents in domestic currency and targeted at resident investors. Panels b and c show the average the intermediation between borrowers and yearly amount raised in domestic bond markets as a percentage of GDP. Panel b shows data for the private sector between 1991 and 2008, while panel c shows data for the public sector between lenders, there is now a greater diversity of 2000 and 2009 for LAC-7 countries. Numbers in parentheses show the number of countries in each players, with a broader set of institutions region. 40 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.8 Domestic bond market turnover 200 178 180 160 % total bond market capitalization 140 120 110 100 80 84 80 80 56 58 59 57 60 35 39 39 40 30 31 24 23 21 18 15 20 12 12 0 Asia (3) China Eastern Europe (3) G-7 (4) India LAC-7 (4) other advanced economies (4) 2000–03 2004–07 2008–09 Source: Didier and Schmukler 2011a. Note: This �gure shows the average bond market trading as share of total bond market capitalization between 2000 and 2009. Trading data include domestic private, domestic public, and foreign bonds traded in local stock exchanges. Numbers in parentheses show the number of countries in each region. intermediating savings, providing economy- to 8 percentage points of GDP), and so do wide credit, and offering an ample variety life and nonlife insurance company premi- of products. In fact, in some emerging coun- ums (table 3.2). In LAC, the lags appear to tries, institutional investors have become be more severe, with both pension funds more important than banks. Given the steady and mutual funds also lagging. There is, injections they receive from underlying inves- however, considerable heterogeneity across tors, pension funds, mutual funds, and insur- LAC-7 (and LAC more generally), partly ance companies provide a stable demand for reflecting cross-country differences in the domestic financial assets and thus have an institutional and regulatory environments. important potential role to play in the deep- Chile has the most developed pension fund ening of local capital markets. Nevertheless, system (reaching 70 percent of GDP), while as argued below, LAC still has a long way to Brazil has the largest mutual fund industry go in terms of the sophistication of its insti- (reaching 42 percent of GDP). When one tutional investors, as most of the savings are looks at performance against benchmark, still channeled to bank deposits and govern- Chile is clearly the star performer as regards ment bonds. pension funds and insurance companies, Pension fund assets represent 20 percent while Brazil excels as regards mutual funds of GDP in LAC-7 countries, while mutual (figure 3.5). In fact, Brazil is the only LAC funds and insurance companies represent 10 country where the size of the mutual funds percent and 6 percent of GDP, respectively industry is very large and exceeds its bench- (figure 3.17). Compared to their statisti- mark by an ample margin. cal benchmarks, mutual funds and private Surprisingly however, institutional inves- pension fund assets in the LAC-7 coun- tors concentrate a significant fraction of their tries are broadly on track; however, insur- asset holdings in fixed-income instruments, ance company assets are lagging (by close such as government bonds and deposits, DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 41 BOX 3.1 The bond market in Chile The capitalization of the Chilean corporate bond they are such big players, their investment behavior market grew from 13 percent of GDP during the is tightly linked to developments in the corporate 1990s to 21 percent in the 2000s. Moreover, bond market. For example, their large size usu- reflecting fiscal consolidation, the private sector ally induces them to invest large amounts, which accounts for an increasing share of total outstand- limits the potential demand for smaller issues. b ing bonds—more than 65 percent in the 2000s. The Moreover, Chilean institutional investors typically primary markets for private bonds are highly active, pursue buy-and-hold strategies, keeping bonds in with new bond issues amounting to 3.4 percent of their portfolios until maturity, as shown in Opazo, GDP. Moreover, the use of private bonds is grow- Raddatz, and Schmukler (2009) and Raddatz and ing. In the 1990s, on average, 15 � rms issued bonds Schmukler (2008), which contributes to explaining in local markets in a given year; during the 2000s, the low liquidity of secondary private bond markets. this number increased to 63. Concentration is also In addition, current regulatory restrictions on pen- less of a concern than in other LAC countries. The sion fund investments limit their exposure to nonin- amount raised by the top five issues during the vestment-grade issues, which helps explain the low 2000s was about 30 percent, which is comparable fraction of outstanding high-yield corporate bonds. to the G-7 countries (33 percent). Furthermore, The long maturity of corporate bonds can also be � rms have been better able to tap into long-term associated with the maturity structure of the liabili- � nancing, as the average maturity at issuance was ties of insurance companies, which allows them to 15.2 years, signi�cantly longer than that for other make long-term investments. The nature of their LAC-7 countries or developed countries (with an liabilities, mostly indexed to inflation, also implies a average maturity of 6.8 years and about 10 years, signi�cant demand for inflation-linked bonds. respectively). a These long maturities are generally The regulatory changes that took place in the linked to indexed, high-grade bonds. However, early 2000s also contributed to these developments. liquidity in secondary markets remains limited. In particular, capital market reforms gave pension These developments must be understood in the funds and insurance companies more fl exibility in context of the strong development of institutional their investments. The consolidation of the macro- investors, mostly pension funds and insurance economic and � nancial frameworks also probably companies, and, to a lesser extent, mutual funds. played an important role. Yet, signi�cant challenges These investors, particularly pension funds, pro- remain, including the need to expand access to vide a stable demand for corporate bonds, given smaller � rms. their sheer size (about 65 percent of GDP for pen- sion funds and 20 percent for insurance companies a. Bonds with less than one year in maturity (commercial paper mostly) are excluded from these statistics due to data availability. in 2010). Pension funds, for instance, held about b. Although large investors do not necessarily make large investments, the data 50 percent of the stock of bonds in 2010, and insur- suggest that they typically do so, perhaps reflecting investment practices by ance companies held another 32 percent. Because institutional investors (Didier, Rigobon, and Schmukler 2011). thereby limiting their role in the development in deposits and government bonds has been of corporate bond and equity markets. Thus, slowly declining. There is also significant het- pension funds in LAC-7 countries invested erogeneity across LAC-7, with pension funds on average 32 percent of their portfolios in in some countries heavily invested in govern- government bonds in 2009, compared with ment securities (for example, Argentina, Mex- 40 percent in Eastern Europe, 37 percent in ico, and Uruguay), while in others, pension emerging Asia, and only 16 percent in the G-7 funds have a greater share of deposits in their countries (figure 3.18, panel a). Nevertheless, portfolios (for example, Chile and Peru) (fig- the concentration of pension fund portfolios ure 3.18, panel c). At the same time, the shares 42 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.9 Average maturity of bonds at issuance in local markets a. Private sector 12 10.3 10 9.4 9.2 8.0 8 number of years 6.8 5.9 6 5.3 4.9 4 2 126 532 12,417 12,155 595 464 340 547 0 ) 7) 6) ies ced (5 7( 7( ) ia om an (6 G- C- As on dv LA ec er a h ot 1991–99 2000–08 b. Public sector 25 20.0 20.0 20 16.8 16.0 15.0 number of years 15 12.8 11.6 10 9.2 8.6 6.3 5.4 5.6 5.5 5.5 4.8 5 4.1 3.0 2.7 2.1 2.0 2.1 1.2 il 2,407 1,147 1,145 a 1,168 1,146 697 645 916 578 ru 120 y 163 37 11 Pe 42 19 36 a 9 2 6 0 ile ico e ua az in bi ag Ch ex nt m Br ug er lo ge M av Ur Co Ar C-7 LA 2000–03 2004–07 2008–09 Source: Didier and Schmukler 2011a. Note: This �gure shows the weighted average maturity of bond issuances per year in domestic markets, expressed in years. Panel a shows data for the pri- vate sector for the period 1991–2008. Panel b shows data for the public sector for 2000–09. Numbers in the base of the bars represent the total number of issues over the period. DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 43 FIGURE 3.10 Currency composition of bonds at issuance in local markets a. Private sector 35 33 30 25 25 25 25 % of total issued bonds 20 15 15 15 10 10 5 1 0 ) 6) 6) ies ced (4 7( 7( ) ia om an (5 G- C- As on dv LA ec er a h ot 1991–99 2000–08 b. Public sector 100 1 3 1 5 4 13 10 11 11 14 12 90 23 80 % of total outstanding bonds 70 70 75 60 85 50 100 97 99 100 100 93 90 89 89 87 86 88 40 72 30 20 30 25 10 15 0 3 7 9 3 7 20 9 20 3 20 7 20 9 20 3 20 7 20 9 20 3 20 7 20 9 20 3 20 7 9 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 00 04 08 00 04 08 00 04 08 00 04 08 00 04 08 00 04 08 20 20 20 20 20 20 Argentina Brazil Chile Colombia Mexico Uruguay local currency bonds inflation-linked bonds foreign currency bonds Source: Didier and Schmukler 2011a. Note: This �gure shows the currency composition of domestic private and public bonds at issuance. Panel a shows the average foreign-currency- denominated bonds as a percentage of total bonds issued by the private sector in domestic markets per year between 1991 and 2008. Panel b shows the composition of domestic public bonds issued on average per year (between local currency, foreign currency, and inflation-linked bonds) over the period 2000 and 2009. Numbers in parentheses show the number of countries in each region. 44 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.11 Equity market size a. Market capitalization as a percentage of GDP 100 91 90 86 80 73 70 67 65 63 61 60 % of GDP 51 50 41 40 31 30 30 26 20 13 13 10 0 ) a (7 e 7) a 7) ies ced (5 in di p 7( 7( ro Ch In ) ) ia om an (7 G- C- Eu As on dv LA rn ec er a ste h Ea ot 1990–99 2000–09 b. Average real annual growth in equity market capitalization 90 80 77 70 average yearly growth 60 50 40 39 30 25 20 18 11 10 8 9 7 4 3 3 3 3 1 0 ) a (2 e 7) ia 7) ies ced (5 in op d 7( 7( Ch In ) ) ia om an (7 r G- C- Eu As on dv LA rn ec er a ste h Ea ot 1990–99 2000–09 Source: Didier and Schmukler 2011a. Note: This �gure shows in panel a the average market capitalization of domestic equity as a share of GDP between 1990 and 2009. Panel b shows the average annual growth rate in market capitalization of domestic equity between 1990 and 2009. The market capitalization is measured in U.S. dollars and is deflated by local stock market indexes to measure the real growth. Numbers in parentheses show the number of countries in each region. DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 45 of equity and foreign securities have been FIGURE 3.12 Equity markets slowly increasing (figure 3.18, panel b).13 Comparable patterns are also observed in a. Amount of new capital-raising issues as a percentage of GDP the portfolios of mutual funds. They invest 1.8 on average a large fraction of their portfo- 1.6 1.6 lios in government bonds and money market 1.4 1.4 1.4 instruments. However, like pension funds, 1.2 1.1 mutual funds have been gradually shifting 1.0 % of GDP 1.0 0.9 their portfolios toward equity investments 0.8 0.8 0.7 0.7 (figure 3.19). In Brazil, for example, the share 0.6 0.7 of public bonds declined from 73 percent to 0.4 0.5 0.3 0.3 48 percent between 2003–04 and 2005–09. 0.2 0.2 In Chile, this fraction declined from 14 per- 0.0 cent to 6 percent. However, bank deposits 6) ) a (7 e 7) a ies ced (5 in di p 7( 7( ro Ch In ) ) ia om an (7 C- G- Eu As continued to account for a substantial share on adv LA rn ste r he Ea (63 percent) of the Chilean mutual fund ot ec 1991–99 2000–08 portfolio. b. Total number of listed �rms As discussed in greater depth in chapter 8, in domestic equity markets these trends suggest that the contribution of 6,000 institutional investors to the development of 5,207 local markets may not have been as powerful 5,000 4,441 as could have been expected. However, the number of listed �rms 4,000 relatively more successful cases of Brazil and Chile deserve particular attention. In par- 3,000 2,322 ticular, more research is needed to measure 2,072 2,000 the effects that Brazilian mutual funds might 1,382 have had on the development of Novo Mer- 1,000 715 828 425 441 385 cado and the Chilean pension funds might 225 167 224 175 0 have had on the development of primary 7) ) a pe 7) a ies ced (5 in di 7( 7( ro Ch In ) ) ia om an bond corporate markets. (7 (7 C- G- Eu As on adv LA rn ste r he Finally, private equity and venture capital Ea ot ec 1990–99 2000–09 (PEVC) funds are also becoming significant players in LAC. These funds are particularly c. Total number of �rms issuing equity per year important for financing small and medium 600 enterprises (SMEs). Due to limited data avail- 532 ability, only a glimpse of their relative roles 500 can be seen. Unsurprisingly, PEVC funds are 400 381 number of �rms still relatively underdeveloped in emerging 315 economies, and particularly so in LAC. Pri- 300 vate equity funds raised on average US$4.9 200 208 billion per year in LAC, as against almost 128 103 97 109 92 US$46.0 billion in Asia.14 Venture capital 100 43 22 funds are even less representative, with a total 0 4 8 8 of US$12.0 billion per year raised on average 6) ) na (7 e ) a ies ced (5 (7 7( di p i ro Ch In ) ) ia 7 C- om an (7 G- Eu As LA outside the United States and Europe. None- on adv rn ste r he Ea theless, although small, PEVC funds have ot ec 1991–99 2000–08 been growing rapidly, particularly in LAC. In the first half of the 2000s, PEVC funds Source: Didier and Schmukler 2011a. raised US$1.2 billion on average in LAC, Panel a shows the average amount of capital-raising equity issues as a percent of GDP between and in the second half of the decade, they 1991 and 2008. Panel b shows the average number of listed �rms between 1990 and 2009. Panel c shows the total number of �rms issuing equity per year between 1990 and 2009. Numbers in paren- raised US$7.7 billion. Continued growth will theses show the number of countries in each region. 46 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.13 Stock market turnover relative to global benchmarks a. Residuals 100 80 stock market turnover (residuals) 60 40 20 0 –20 –40 –60 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 20 20 20 20 20 Asia Eastern Europe G-7 LAC-7 other advanced economies 160 b. Filtered 140 120 stock market turnover (filtered) 100 80 60 40 20 0 –20 –40 4 5 6 7 8 9 10 11 log of GDP per capita Asia Eastern Europe G-7 LAC-7 Source: de la Torre, Feyen, and Ize 2011. Note: This �gure shows stock market turnover across country groups relative to their global benchmarks. Panel a shows the evolution of annual median residuals by country group. First, a cross-sectional regression model is estimated on country averages, controlling for GDP per capita, population size and density, young and old-age dependency ratios, a �nancial offshore center dummy, a transition country dummy, and a large fuel exporter dummy. Annual country residuals are derived from the cross-sectional estimates, which are used to calculate annual country-group medians. Panel b plots annual country- group medians of the �ltered variable against the logarithm of GDP per capita. The �ltering consists of removing from the actual annual country observa- tions the expected contributions of all factors except GDP per capita, using the cross-sectional estimates from the model. A four-knot spline approach is used to account for nonlinearities of GDP per capita in the cross-sectional regression. DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 47 FIGURE 3.14 Concentration in equity markets a. Share of amount raised by the top �ve issues as a percentage of the total amount raised 100 97 93 92 90 80 78 72 70 66 % of total amount raised 61 62 60 55 51 50 47 45 42 40 30 29 20 10 43 128 103 97 4 8 315 381 532 109 22 8 92 208 0 ) a (7 e 7) a 6) ies ced (5 in di p 7( 7( ro Ch In ) ) ia om an (7 G- C- Eu As on dv LA rn ec er a ste h Ea 1991–99 2000–08 ot b. Value traded by the top �ve companies as a percentage of the total value traded 80 73 70 63 60 60 58 57 % of total value traded 50 39 39 40 36 30 21 20 20 10 0 ) a (5 e a 6) (5 in di p 7( ro Ch In ) ia C- Eu As LA rn ste Ea 1990–99 2000–09 Source: Didier and Schmukler 2011a. Note: This �gure shows the concentration in equity market activity. Panel a shows the average amount raised per year by the top �ve issues as a share of total issues between 1991 and 2008. Numbers in the base of the bars represent the average number of issues per year. Panel b shows the average share of value traded by the top �ve companies as share of the total value traded per year between 1990 and 2009. Numbers in parentheses show the number of countries in each region. 48 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.15 Within LAC: Trading activity and number of listed require adequate regulation and rigorous dis- �rms in domestic equity markets closure standards. The latter is perceived as a particular issue in LAC, as accessing accurate a. Turnover ratio and objective information for nonpublic firms 30 is not straightforward. In this context, effec- 25 24.5 tive ex ante due diligence, valuation analysis, and ex post business monitoring, key for this turnover ratio (%) 20 17.4 industry, can be rather difficult. 15 12.3 10 Alternative markets and 5 3.7 3.0 4.9 3.2 3.4 4.3 products 1.6 0 As part of a global trend, LAC countries ) bl ica s 7) C rs ) (2 (3 pu in de LA te have seen the development of less traditional 7( ic) n ) an ica Re Dom nclu in cen (2 C- be er ) LA (2 e (i e m ib or forms of financing. This section briefly th ica hA r Ca sh er ut off m So lA reviews some of the most important ones: ra nt Ce b. Total number of listed �rms derivatives, factoring, cooperatives and in domestic equity markets credit unions, retailers’ financing, exchange 250 224 traded funds, and securitization. These illus- 200 trate that financial activity is becoming more 175 complex, with traditional markets capturing number of listed �rms 150 only part of the financing. Unfortunately, the scope of this analysis is constrained by 100 67 data availability. 50 37 38 51 45 Trading in exchange rate derivatives has 28 17 17 grown in dollar terms, but it has actually 0 remained relatively stable as a percentage of ) bl ica s 7) C rs ) (2 (4 pu in de GDP in most countries since the late 1990s. LA te 7( ic) n ) an ica Re Dom nclu in cen (4 C- be er ) LA (2 e (i e m ib Trading in interest rate contracts, on the other or th rica hA r Ca sh e ut off m So hand, more than doubled as a percentage of lA ra nt Ce c: Total number of listed �rms GDP in the 2000s compared to the 1990s per million inhabitants in most emerging regions, with a growth 160 rate of 213 percent among LAC-7 countries 142 number of �rms per million inhabitants 140 (figure 3.20, panels a and b). Nevertheless, 120 106 derivatives overall are relatively illiquid, with 100 aggregate trading representing only a small 80 fraction of trading in developed countries. 60 Thus, turnover in exchange rate contracts 40 stands at about 1.1 percent of GDP in LAC 20 21 17 countries, compared with 7.3 percent of GDP 20 5 7 5 5 4 in the G-7 countries. Moreover, trading on 0 foreign exchange contracts is nearly exclusively ) bl ica s ) C rs ) (2 (7 (4 pu in de LA te ic) n ) an 7 ica Re Dom nclu in cen (4 (98 percent) in U.S. dollar contracts. C- be er ) LA (2 e (i e m rib or th rica hA Ca sh Factoring is another example of a rap- e ut off m So lA idly expanding industry.15 This expansion ra nt Ce 1990–99 2000–09 has taken place mostly in emerging econo- mies, including in LAC-7 where factor- Source: Didier and Schmukler 2011a. ing amounted to 2.6 percent of GDP in Note: Panel a shows average turnover ratio across LAC, de�ned as the total value traded per year in domestic markets over total market capitalization between 1990 and 2009. Panels b and c show, 2008–09 (figure 3.20, panel c). Chile and respectively, the average number of listed �rms and expressed per million inhabitants across LAC Mexico provide examples of particularly fast between 1990 and 2009. Numbers in parentheses show the number of countries in each region. DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 49 FIGURE 3.16 Equity market capitalization as a percentage of GDP, within LAC 90 82 80 73 70 60 % of GDP 50 41 40 34 31 30 26 20 15 9 7 9 10 0 ) 7) ) ) )( s (2 (3 (4 lic de 7( 2) an C ica ub lu C- LA ep (inc be er LA in m ib n R ca hA rs r Ca te ica ri in me ut en So ec om al A or e D tr sh th Cen off 1990–99 2000–09 Source: Didier and Schmukler 2011a. Note: This �gure shows equity market capitalization as a percentage of GDP between 1990 and 2009 across LAC. Numbers in parentheses show the number of countries in each region. development. In these countries, invoices can Credit by retailers (mainly department actually be traded on organized exchanges stores) is another interesting case of direct or online markets. Factoring in Chile is one competition to banks. It is rising rapidly in of the largest among emerging economies, LAC, with Chile as a particularly notable with an accumulated volume of 12 billion example. The largest department stores euros in 2009 (10.7 percent of GDP) and have become nontrivial providers of house- about 14,000 users.16 Factoring in Mexico hold credit in recent years. Although banks also represents an important market, with are still the main external financing source total industry turnover estimated at almost for Chilean households, representing 68 per- 11 billion euros in 2007 (almost 2 percent cent of the total household financial debt, of GDP). In 2001, the Mexican development household credit by retailers accounts for bank NAFIN (Nacional Financiera) cre- 11 percent of total household financial ated an online market for factoring services debt, 17 percent of total consumer debt, and called Cadenas Productivas (productive 35 percent of nonbank debt (figure 3.21, chains), which is proving to be quite success- panel b).19 The model has been so successful ful.17 This reverse factoring program is rela- that Chilean retailers are exporting it to other tively large, having extended US$11.8 bil- countries in LAC. Financial cooperatives and lion in financing in 2008, which represents credit unions present another alternative to a significant share of the factoring market in bank financing.20 However, loans from these Mexico.18 institutions generally account for only a small 50 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? BOX 3.2 The equity market in Brazil The equity market in Brazil has gone through sig- improvements in governance and investor protec- nificant changes over the past decade, bringing tion might well have paid off. clear improvements in corporate governance. In In spite of a timid beginning, due mostly to a December 2000, the São Paulo Stock Exchange number of external shocks, the Novo Mercado took (Bovespa) created three new corporate governance off by the mid-2000s, and the number of companies listing segments through which issuers could volun- listed in these new corporate governance segments of tarily adopt corporate governance practices beyond Bovespa has risen steadily over time. The implemen- those required by the Brazilian Corporate Law and tation of the Novo Mercado has been well received by capital market regulation more generally. Bove- by foreign investors, as well. During 2004–06, for- spa listing segments now include the traditional eign participation in the corporate governance seg- Bovespa, Level 1, Level 2, and Novo Mercado. ments captured, on average, 70 percent of the stock Each of these market segments requires progres- offerings. Santana (2008) argues that the Novo Mer- sively stricter standards of corporate governance. a cado has allowed Brazilian companies and, particu- The main goal of the creation of these distinct seg- larly, new entrant companies to access foreign capital ments, and in particular of Novo Mercado, was to without having to cross-list on international stock reverse the weakening of the equity markets in Bra- markets. For example, among the 27 initial public zil that had been occurring at the end of the 1990s offerings between 2004 and the � rst half of 2006 in by fostering good corporate governance practices, Bovespa, only two companies were listed simultane- such as disclosure, transparency, and account- ously on the New York Stock Exchange. ability. In fact, equity markets have become more liquid and less concentrated, and more � rms have a. See Didier and Schmukler (2011a) for a more detailed analysis. A description of the rules governing these different segments is available at Bovespa’s been issuing equities. These trends suggest that the website (http://www.bmfbovespa.com.br). fraction of the total. In the LAC-7 countries, Finally, securitization has been on the rise they are even smaller than in other emerging as well. Structured finance grew very fast in economies, accounting for only 0.7 percent of developed countries over the last decade, GDP against 2.0 percent in Asia and 1.7 per- the United States being the leading mar- cent in India (figure 3.21, panel a). ket. 21 However, the global financial crisis Another product currently experiencing in 2008 dealt a blow to structured finance, fast development is exchange-traded funds and worldwide net issuance fell from about (ETFs). ETFs are traded portfolios composed US$2 trillion in 2007 to less than US$400 of not only stocks, but also commodities and billion. 22 Since then, it has been slowly bonds. They facilitate portfolio diversifica- regaining ground, reaching almost US$750 tion and possess stocklike features (such as billion in 2010 (still much below precrisis transparency and frequent pricing as well as levels). Across LAC, primary market secu- ease of trading), and they have low trading ritization activity has risen over the past costs. ETFs have been growing considerably decade. Thus, securitized instruments have in developed and emerging countries alike shown signs of increasing depth on different over the past few years. In LAC, these prod- asset classes, particularly in Brazil and Mex- ucts have been on the rise in countries like ico. Gross issuance for LAC countries as a Mexico. ETFs are also gaining ground in sec- whole rose from US$2.0 billion in 2000 to ondary markets, with an increasing share of US$24.4 billion in 2010, with Mexico and total trading in stock markets (figure 3.21, Brazil the largest issuers. Although some of panel c). these issues were cross-border—typically DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 51 FIGURE 3.17 Pension fund, mutual fund, and insurance company assets a. Pension funds 40 35 34 33 30 30 26 25 20 19 % of GDP 16 15 15 15 10 5 5 6 5 2 0 0 0 ) a (7 e 7) a 7) ies ced (4 in di p 7( 7( ro Ch In ) ) ia om an (6 G- C- Eu As on dv LA rn ec r a ste he Ea ot b. Mutual funds 40 36.7 35 34.2 30 25 23.7 % of GDP 20 18.8 17.1 15 11.8 10.0 10 7.9 7.4 7.0 5 4.3 4.0 1.6 2.4 0 ) a (7 e 7) a 7) ies ced (4 in di p 7( 7( ro Ch In ) ) ia om an (6 G- C- Eu As on dv LA rn ec er a ste h Ea ot c. Insurance companies 60 53 50 46 40 35 35 % of GDP 30 20 15 12 14 12 10 7 9 4 6 5 6 0 ) a (6 e ) a 7) ies ced (5 (7 in di p 7( ro Ch In ) ) ia 7 om an (5 G- C- Eu As on dv LA rn ec er a ste h Ea ot 2000–04 2005–09 Source: Didier and Schmukler 2011a. Note: This �gure shows the total assets of domestic institutional investors—namely pension funds (panel a), mutual funds (panel b), and insurance companies (panel c). Panel a shows average pension fund assets as a percentage of GDP between 2000 and 2009. Panel b shows average mutual fund assets as a percentage of GDP between 2000 and 2009. Panel c shows average insurance companies assets as a percentage of GDP between 2000 and 2009. Numbers in parentheses show the number of countries in each region. 52 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.18 Composition of pension fund portfolios a. Worldwide average 100 4 10 10 10 90 18 7 22 80 5 19 70 16 34 25 20 16 % of total 60 50 10 20 37 19 20 40 6 30 40 11 20 32 26 24 16 10 5 3 7 0 ) (4 e 5) 5) ies ced (2 p 7( 7( ro ) ) ia om an (6 G- C- Eu As on dv LA rn ec er a ste h Ea ot cash and deposits public bonds private bonds loans equity mutual funds others b. Average for LAC-7 countries 100 11 90 14 6 80 9 12 70 % of total portfolio 8 60 21 16 50 40 30 51 45 20 10 0 1999–2004 2005–08 government securities �nancial securities and deposits private bonds foreign securities equities mutual funds other investments c. Individual LAC-7 countries 100 3 3 2 3 3 6 5 6 9 8 2 90 11 12 15 3 12 7 17 11 33 80 6 14 34 36 17 20 5 36 10 % of total portfolio 70 9 9 6 32 60 3 11 6 17 50 9 15 11 32 8 88 40 73 12 30 64 59 29 61 59 29 49 11 47 20 30 10 14 21 13 0 4 8 4 8 4 8 4 8 4 8 4 8 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 –0 99 05 99 05 99 05 99 05 99 05 99 05 19 20 19 20 19 20 19 20 19 20 19 20 Argentina Chile Colombia Mexico Peru Uruguay government securities �nancial securities and deposits private bonds foreign securities equities mutual funds other investments Source: Didier and Schmukler 2011a. Note: This �gure shows in panel a the most recent composition of pension funds portfolio holdings for the latest available information, 2009 mostly. Panel b shows the average composition of pension fund investments as share of the total portfolio between 1999 and 2008 for LAC-7 countries, excluding Brazil. Panel c shows the data for individual countries. DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 53 FIGURE 3.19 Mutual fund portfolio holdings a. Brazil b. Chile 100 6 100 4 5 14 11 90 8 90 17 2 2 80 4 9 80 5 70 17 13 70 15 % of total assets % of total assets 60 4 60 50 50 73 40 40 48 63 63 30 30 20 20 10 10 6 11 0 0 2003–04 2005-09 2000–04 2005–09 deposit certi�cates government bonds deposits private bonds private bonds �xed income securities backed by government debt domestic equity foreign equity equity others public bonds c. Colombia d. Mexico 100 90 15 15 100 12 10 80 90 80 9 9 70 % of total assets 70 % of total assets 60 49 46 60 50 50 40 61 63 40 30 30 20 36 39 20 10 10 15 14 0 0 2004 2005–08 2003–04 2005-09 variable income �xed income real estate others deposits domestic public bonds domestic private bonds foreign public bonds foreign private bonds equity others e. Peru 100 12 11 90 1 6 80 70 % of total assets 60 54 45 50 40 30 20 36 33 10 0 2000–04 2005–09 bank deposits bonds equity foreign equity others Source: Didier and Schmukler 2011a. Note: This �gure shows the composition of the mutual funds portfolios in LAC-5 countries. For Peru, fondos mutuos and fondos de inversiones are considered. Equity includes acciones de capital and acciones de inversión for fondos mutuos. In the case of investment funds, equities are composed of acciones de capital, fondos de inversión, and otras participaciones until 2002, and derechos de participación patrimonial from 2004 onward. In the case of Colombia, fondos vigilados and fondos controlados are reported in different tables for 2002. Period averages are calculated using simple averages. 54 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? FIGURE 3.20 Derivative and factoring markets a. Traditional and OTC foreign exchange derivates turnover 10 8.9 9.2 9 8 7.3 6.6 6.9 7 6.2 6 5.5 % of GDP 5 4.5 4 3 2.8 2.7 1.7 1.9 2 1.7 1.6 1.3 1.1 0.9 1.0 1.1 1.1 1 0.7 0.4 0.3 0.3 0 3 7 3 3 3 3 3 7 7 7 7 7 00 –0 00 00 00 00 00 –0 –0 –0 –0 –0 –2 –2 –2 –2 –2 –2 04 04 04 04 04 04 98 20 98 98 98 98 98 20 20 20 20 20 19 19 19 19 19 19 Asia (5) Eastern G-7 (7) India LAC-7 (6) other Europe (5) advanced all contracts OTC contracts economies (7) b. OTC interest rate derivates turnover 8 7 6.7 6 5 % of GDP 4 3.3 3 2 1.7 1.4 1 0.5 0.0 0.1 0.2 0.0 0.2 0.0 0.2 0 ) (3 e 7) a 2) ies ced (4 di p 7( 7( ro In ) ) ia om an (6 G- C- Eu As on dv LA rn ec er a ste h Ea ot 1998–2003 2004–07 c. Total volume as a percentage of GDP in factoring markets 5.0 4.7 4.5 4.3 4.1 4.0 4.0 3.5 3.3 3.1 3.0 % of GDP 2.5 2.6 2.5 2.0 1.6 1.5 1.0 0.7 0.6 0.5 0.4 0.5 0.4 0.0 ) a (7 e ) a ) ies ced (3 (7 (6 in di op Ch In ) ) ia 7 7 om an ur (7 G- C- As E on dv LA rn ec er a ste h Ea ot 2005–07 2008–09 Source: Didier and Schmukler 2011a. Note: Panel a shows the average annual foreign exchange derivates turnover as a percentage of GDP between 1998 and 2007. Panel b shows the average annual OTC interest rate derivates turnover as a percentage of GDP between 1998 and 2007. Panel c shows the average yearly turnover as a percentage of GDP in factoring markets between 2005 and 2009. Numbers in parentheses show the number of countries in each region. DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 55 FIGURE 3.21 Financial cooperatives, credit unions, and exchange-traded funds a. Credit by not-for-pro�t cooperative institutions 9 8.1 8 7 6.8 6 % of GDP 5 4 3 1.9 2.0 1.8 2 1.7 1.6 1.4 1 0.6 0.7 0.2 0.3 0 ) ) 2) a 5) ies ced (5 (2 di 7( 7( In ) ia pe om an (2 G- C- As ro on adv LA Eu r rn he ste ot ec Ea 2005–07 2008–09 b. Credit by retailers in Chile Household debt Consumer debt 3 3 15 6.8 7.8 15 11.1 53 68.2 17 banks banks retailers retailers insurance company loans family companies, funds, family companies, funds, and cooperatives and cooperatives others car �nancing university loans c: ETFs trading as a percentage of total trading in domestic equity markets 2.5 2.2 2.0 % of total trading 1.5 1.2 1.0 0.8 0.7 0.5 0.3 0.3 0.3 0.1 0.1 0.1 0.1 0.0 0.0 0.0 0.1 0.0 ) 5) a 2) ies ced (1 di 7( 7( In ) ia om an (2 G- C- As on adv LA r he ot ec 2000–03 2004–07 2008–09 Source: Didier and Schmukler 2011a. Note: Panel a shows the average amount of credit provided by not-for-pro�t cooperative institutions as a percentage of GDP between 2005 and 2009. Panel b shows credit by retailers in Chile in December 2008. Panel c shows the average ETF value trading per year as a share of total trading in domestic equity market between 2000 and 2009. The share in each country is also reported on top of the bars. Numbers in parentheses show the number of countries in each region. 56 DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? between US$2 billion and US$4 billion over and Schmukler (2011a) for a more detailed the past five years and mostly on futures— characterization of corporate bond markets domestic issuance accounts for the largest within the LAC region. share of the market. Furthermore, securiti- 9. As with bank debt, LAC countries have made an important effort to reduce currency and zation has also expanded across a broader maturity mismatches so as to limit credit set of asset classes, such as new and used car and rollover risks. See Broner, Lorenzoni, and loans, consumer loans, credit card receiv- Schmukler (2011). ables, equipment leases, and mortgages. 10. See de la Torre and Schmukler (2004). This broadening of asset classes is particu- 11. The fact that only a restricted set of �rms uses larly apparent in Brazil and Mexico. capital markets can be at least in part driven by supply factors. For instance, the restricted investment practice of institutional investors Notes is one possible explanation. As documented 1. This chapter draws heavily on the papers in a number of papers, institutional inves- “Financial Development in Latin America: tors tend to invest in larger and more liquid Stylized Facts and the Road Ahead� by �rms, hence limiting the supply of funds to Tatiana Didier and Sergio Schmukler smaller and less liquid �rms. See, for exam- (2011a) and “Benchmarking LAC’s Financial ple, Dahlquist and Robertsson (2001); Didier Development� by Augusto de la Torre, Erik (2011a); Didier, Rigobon, and Schmukler Feyen, and Alain Ize (2011), both of which are (2011); Edison and Warnock (2004); Kang part of the Edited Volume that accompanies and Stulz (1997); among many others. this LAC flagship report. 12. La Porta et al. (1997) and Glaser, Johnson, 2. Throughout this report, the LAC-7 group and Shleifer (2001) show that protection of comprises Argentina, Brazil, Chile, Colombia, minority shareholders is fundamental to the Mexico, Peru, and Uruguay. We include development of a country’s capital market. In Uruguay but exclude República Bolivariana addition, Klapper and Love (2004) show that de Venezuela, mainly because there is much good governance practices are more important more data for the former than for the latter. in countries with weak investor protection and 3. However, results for other LAC countries inef�cient enforcement. According to Shleifer should be taken with caution because of a and Vishny (1997a) and Bhojraj and Sengupta general lack of data availability. (2003), good governance practices increase 4. The statistical benchmarking analysis is the one con�dence among investors as they tend to presented in chapter 2 and explained in more reduce agency risks. Therefore, companies can detail in de la Torre, Feyen, and Ize (2011). access the capital market at lower costs and 5. In fact, as documented in chapter 2, �nancial better terms, increase the value and liquidity of development in the most advanced economies their shares, and improve their operating per- has generally expanded much faster than in formance and pro�tability. Ashbaugh-Skaife, the rest of the world. Collins, and LaFond (2006), for example, �nd 6. This simple addition of bank assets with bond that better corporate governance practices and equity market capitalization entails some improve corporate credit ratings and reduce double counting; for example, bank assets bond yields. De Carvalho and Pennacchi likely include some bond holdings and per- (2011) argue, for the case of Brazil, that migra- haps even equities. tion from traditional markets to the Novo 7. Such development likely comes from the Mercado brings positive abnormal returns to improvement in monetary management and shareholders and an increase in the trading a shift toward floating rate regimes, which volume of shares. Klapper and Love (2004) induced both banks and borrowers to limit �nd that better corporate governance is asso- their currency mismatches. See Ilyina, Guscina, ciated with higher operating performance and and Kamil (2010) and the note by Levy-Yeyati higher Tobin’s Q. Joh (2003) concludes that (2011a) in the companion Edited Volume. �rms with higher control-ownership disparity 8. Market capitalization of corporate bonds in exhibit lower pro�tability. Gozzi, Levine, and Brazil has grown from 12 percent of GDP in Schmukler (2010), however, argue that the 2004 to almost 29 percent in 2009. See Didier causality might go both ways and that better DOMESTIC FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND? 57 �rms go to better corporate governance envi- are carried out on an electronic platform, ronments, not that better governance neces- which allows NAFIN to capture economies sarily increases �rm value. of scale, since most of the costs of the system 13. The differences between panels a and b are are �xed and electronic access enables a large due to three factors. First, the reporting number of �rms and �nancial institutions to periods are different. Second, the country participate. groupings differ in that the Organisation for 18. The program has been so successful that Economic Co-operation and Development NAFIN has entered into agreements with (OECD) data in panel a exclude Argentina development banks in several Latin American and Uruguay, and country data in panel b countries, including Colombia, El Salvador, exclude Brazil because of a lack of disaggre- and República Bolivariano de Venezuela, gation in its statistics. Third, the sources dif- to implement similar programs, while other fer in their classi�cation of foreign securities, development banks in the region are also which are not reported as a separate category considering replicating this model. in the OECD data. 19. This high penetration of the retail sector in 14. These statistics are from Preqin, the indus- Chile has been related to the introduction try’s leading source of information where of in-house credit cards. These credit cards country-level information is not available. issued by department stores became popular Note that Preqin’s designations of countries in Chile because they offered consumer credit, within regions are different from those used especially to the middle-income segment of in this chapter. the population, when the bank credit market 15. Factoring is a �nancial transaction where to this segment was still incipient. accounts receivables (that is, invoices) are 20. Financial cooperatives and credit unions are sold at a discount to a third party. See Klapper typically �nancial institutions owned and (2006) and de la Torre, Gozzi, and Schmukler controlled by their members and operated (2007b) for a detailed discussion of factoring with the purpose of providing credit and per se as well as case studies. other �nancial services to them. Hence, they 16. Launched in 2005 as an alternative to bank aim mostly at credit provision to households factoring, the exchange Bolsa de Productos as well as micro-, small, and medium enter- has been growing fast and might become an prises, either formal or informal. They vary important source of SME �nancing in the signi�cantly in size, ranging from small coop- near future. This exchange allows some form eratives with few members to others compa- of reverse factoring, where invoices can be rable in size to commercial banks. discounted and where the credit risk borne 21. Structured �nance is, in its simplest form, a by the investor is that from the issuers of the process where assets are pooled and trans- invoice. Moreover, no collateral is needed ferred to a third party, commonly referred from SMEs posting the invoice. Keys to the to as a special purpose vehicle (SPV), which success of this initiative are that discounting in turn issues securities backed by this asset invoices in Bolsa de Productos is cheaper than pool. In other words, structured �nance trans- factoring through banks and that it provides actions can help to convert illiquid assets into investors with a higher yield than they can get tradable securities. Typically, several classes of in money markets. securities (called tranches) with distinct risk- 17. See de la Torre, Gozzi, and Schmukler 2007a. return pro�les are issued. This market “spearheaded� by NAFIN pro- 22. Net issuance includes issues sold into the vides reverse factoring services to SMEs market and excludes issues retained by issu- through the creation of chains between large ing banks, while gross issuance includes those buyers and their suppliers. All transactions retained issues. Financial Globalization: Where Does LAC Stand? 4 A s countries become more integrated more equity liabilities), which has made with one another (a trend that Latin globalization much safer. America and the Caribbean (LAC) • In contrast with other emerging econo- has embraced wholeheartedly), a comprehen- mies, LAC’s equity liabilities continue to sive view of financial development requires be dominated by foreign direct investment looking not only at domestic financial activ- (FDI) rather than by portfolio equity, ity but also at financial globalization.1 This which is consistent with the shortcomings chapter explores two dimensions of financial of the local equity markets. globalization. The first is financial diversi- • T he offshore issuance and trading fication, that is, diversifying portfolios by of LAC-7 firms’ equity has increased including foreign assets (or liabilities). The strongly, relative to onshore, which second is financial offshoring, that is, the heightens concerns of market access for use of more efficient foreign jurisdictions to the smaller � rms. conduct financial transactions. The chap- • Public sector bond � nancing has shifted ter’s main findings are as follows: toward domestic markets and away from foreign markets, reflecting countries’ • While LAC’s financial diversification desire to reduce their dollar exposure and has continued to rise, � nancial offshor- to develop their local currency public debt ing (which increased signi� cantly in the markets. 1990s) has not grown much over the past • While LAC countries have made signi�- decade. cant progress toward issuing local cur- • However, valuation effects, in particu- rency bonds abroad, they still have a long lar for equity investments, have been the way to go. main drivers of the increase in � nancial • Corporate bond financing abroad has diversi�cation. experienced gains in maturity similar to • The growing international financial bond � nancing at home, although only a diversi�cation has been accompanied by few large � rms have access to local bond a changing portfolio composition (more markets, and even fewer have access to reserve assets, less debt liabilities, and foreign bond markets. 59 60 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? The first section briefly describes the two local residents for efficiency rather than for dimensions of financial globalization. The risk-diversification purposes. Thus, instead of second section analyzes the extent and impli- listing a stock locally, a firm might prefer to cations of financial diversification. The third list it on a foreign exchange, because the off- and last section explores financial offshoring, shore market might be deeper and more liq- first of bonds and syndicated debt, then of uid. Or, instead of opening a deposit account equity. in a local bank, an individual might prefer to open it at an offshore bank, because it may provide better service for offshore payments. The two dimensions of �nancial In either case, although the financial services globalization provided offshore may be similar to those The financial diversification dimension of provided locally, they also may be cheaper financial globalization is macroeconomic in or may have specific features that make them essence. It relates to a country’s capital flows preferable for specific transactions. Clearly, and aggregate gross foreign positions in when measuring the extent of local residents’ assets and liabilities. Domestic residents financial activity, one needs to consider the invest (or borrow) abroad, and foreigners activities conducted abroad to grasp the full invest locally, because this allows risk to be extent of market depth. diversified more effectively (and returns to be However, one must also take into account equalized) across borders and instruments. the substitutability or complementarity of In addition to enhancing the efficiency of domestic and foreign activity.6 For instance, resource allocation, the increased partici- when stock trading takes place in interna- pation of foreign investors can also benefit tional markets, domestic activity might actu- local market development. 2 It can enhance ally migrate abroad, with agents substitut- liquidity, boost research, improve the quan- ing foreign markets for domestic ones. This tity and quality of information available, is not innocuous, because not all firms may and, more generally, increase transparency have access to international markets. Thus, and promote the adoption of better corpo- domestic exchanges might become illiquid rate governance practices, thereby reducing after large firms go abroad.7 As the smaller agency problems. 3 This process can also firms remain constrained to local financing have downsides, however. In particular, sources, such a migration can reduce not shocks that affect foreign investors (that is, only the liquidity of remaining firms in local changes in risk appetite that result in capi- markets, but also their ability to raise capital, tal flow volatility) can also affect negatively jeopardizing the sustainability of domestic the local economies through volatility and capital markets. At the same time, domestic amplification effects.4 Moreover, because and offshore markets may complement each foreigners tend to provide financing in for- other because, for example, they offer dif- eign currency, this can lead to currency ferent financing choices; thus, foreign bond mismatches. Similarly, while domestic resi- markets might be typically used for assets dents’ investment abroad can help smooth denominated in foreign currency, whereas their consumption and, hence, home out- domestic markets might be used for both puts, such investments may also reflect domestic and foreign currency bonds. capital flight caused by deteriorating condi- tions at home (risks of devaluation, default, and expropriation) that, other things being Financial diversi�cation equal, can reduce the capital available for The financial diversification dimension of domestic financing.5 financial globalization was measured in this Instead, the offshoring dimension of finan- study based on the widely used de facto mea- cial globalization is mainly microeconomic in sures compiled by Lane and Milesi-Ferretti essence. It relates to the use of offshore (rather (2007), which are based on stocks (the stock than onshore) markets or intermediaries by of foreign assets and liabilities) as well as FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 61 flows (gross capital flows by domestic and While Central and South America exhibit foreign residents).8 The raw stock measures similar levels to LAC-7, the Caribbean suggest increasing diversification, with LAC- countries are somewhat more integrated, 7’s integration comparable to that of other largely reflecting their smaller size (fig- emerging economies (figure 4.1, panel a). ure 4.2, panel a). Controlling for GDP and FIGURE 4.1 Foreign assets and liabilities and gross capital flows a. Foreign assets and liabilities 350 320 300 290 250 200 % of GDP 176 150 135 143 133 132 108 109 100 87 79 85 91 86 76 52 56 58 50 39 18 22 0 Asia (5) China Eastern G-7 (7) India LAC-7 (7) other advanced Europe (3) economies (7) 1980–89 1990–99 2000–07 b. Gross capital flows 30 25 20 14 13 % of GDP 8 15 10 10 7 6 5 4 4 12 13 3 4 3 13 5 1 6 3 3 7 1 7 4 5 1 4 5 5 4 5 4 5 2 4 2 1 2 1 1 1 0 20 99 20 99 20 99 20 99 20 99 20 99 20 99 19 09 19 09 19 09 19 09 19 09 19 09 9 19 89 19 89 19 9 19 89 19 89 19 89 19 89 –0 –8 – – – – – – – – – – – – – – – – – – – 90 90 90 90 90 90 90 00 00 00 00 00 00 00 80 80 80 80 80 80 80 19 Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (3) advanced economies (7) flows by foreign residents flows by domestic residents Sources: Authors’ calculations based on Lane and Milesi-Ferretti 2007 and Didier and Schmukler 2011b. Note: Panel a shows the foreign assets and liabilities as a percentage of GDP between 1980 and 2007. Panel b shows the flow by foreign and domestic residents as a percentage of GDP between 1980 and 2009. Numbers in parentheses show the number of countries in each region. 62 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? FIGURE 4.2 Within LAC: Foreign assets and liabilities and gross capital flows a. Foreign assets and liabilities 200 186 180 158 156 160 152 140 135 125 128 121 120 114 % of GDP 100 100 91 85 80 60 40 20 0 Caribbean (2) Central America LAC-7 (7) South America (4) (includes the Dominican Republic) (7) 1980–89 1990–99 2000–07 b. Gross capital flows 25 90 80 20 70 37 60 15 11 % of GDP 50 % GDP 25 40 10 2 30 3 4 44 3 20 5 2 10 7 5 26 1 5 10 4 3 1 4 4 2 2 1 1 1 0 0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –8 –8 –8 –8 –8 90 00 90 00 90 00 90 00 90 00 80 80 80 80 80 19 20 19 20 19 20 19 20 19 20 19 19 19 19 19 Caribbean (2) Central America LAC-7 (7) South America offshore centers (includes the Dominican (4) in LAC (3)-right axis Republic) (7) flows by foreign residents flows by domestic residents Sources: Authors’ calculations based on Lane and Milesi-Ferretti 2007 and Didier and Schmukler 2011b. Note: Panel a shows the foreign assets and liabilities as a percentage of GDP between 1980 and 2007 across LAC. Panel b shows the flow by foreign and domestic residents as a percentage of GDP between 1980 and 2009 across LAC. Numbers in parentheses show the number of countries in each region. FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 63 country structural characteristics using the local equity markets, which, in turn, have same statistical benchmarking model as the been boosted more by price increases prior one described in previous chapters, LAC-7’s to the crisis than by new (primary) issuance. financial diversification (measured as gross The flow measures confirm that capital equity and debt assets and liabilities) is close flows by domestic and foreign agents in the to benchmark (table 3.2). LAC-7 countries have remained relatively However, as with market capitalization in constant and moderate over the last decade chapter 3, the expansion of financial diversi- (figure 4.1b). However, this pattern has not fication also reflects large valuation effects.9 been uniform across LAC, as the Caribbean When one incorporates these effects by scal- countries and offshore centers have experi- ing the foreign holdings of domestic equity enced a strong increase in both inflows and by market capitalization (instead of by GDP), outflows (figure 4.2b). the change in cross-border holdings between While de facto measures suggest that 1999 and 2007 turns negative for LAC, in financial diversification across LAC-7 coun- contrast with other emerging economies, tries has been relatively stagnant over the where the change is positive (figure 4.3). This past decade, a de jure index that measures suggests that the often-cited increase in LAC the degree of capital account openness shows cross-border equity liabilities, rather than a that financial globalization has been on the proactive relocation of international capital, rise over the past 20 years, although it stabi- has largely mirrored the growing depth of lized at a high level about 2004 (figure 4.4a). FIGURE 4.3 Valuation effects 30 25 23.8 20 15 12.3 11.4 % change 10 8.5 8.1 7.0 5.2 5.7 5 3.8 3.6 0.4 0 –0.9 –2.1 –5 –4.4 –10 Asia China Eastern G-7 India LAC-7 other Europe advanced economies foreign equity liabilities/GDP foreign equity liabilities/market capitalization Source: Authors’ calculations based on Lane and Milesi-Ferretti 2007. 64 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? FIGURE 4.4 De jure and de facto �nancial globalization measures a. De jure �nancial globalization measure 3.0 2.5 2.0 1.5 1.0 0.5 index 0.0 –0.5 –1.0 –1.5 –2.0 –2.5 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 Asia China Eastern Europe G-7 India LAC-7 other advanced economies b. De facto �nancial globalization measure: Change in liabilities holdings 80 60 13 9 14 40 26 23 54 35 % change 38 8 5 20 15 4 42 21 1 22 22 14 8 18 16 8 15 11 12 3 6 6 4 3 5 7 0 –1 –1 -4 –5 –3 –9 –11 –17 –27 –20 –40 7 9 99 9 9 7 9 7 7 07 7 9 9 7 –0 –9 –9 –9 –0 –9 –0 –0 –0 –9 –9 –0 – – 00 90 90 90 90 00 90 00 00 00 00 90 90 00 20 19 19 19 19 20 19 20 20 20 20 19 19 20 Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (7) advanced economies (7) equity FDI debt (continued next page) FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 65 FIGURE 4.4 (continued) c. De facto �nancial globalization measure: Change in assets holdings 80 1 70 60 50 37 54 0 19 % change 40 15 30 15 5 13 11 20 13 31 18 9 4 10 7 6 14 3 5 18 4 15 18 3 11 7 3 5 2 5 9 4 3 3 2 3 3 6 0 1 1 1 1 0 –2 0 –2 –5 –5 –10 9 7 9 7 9 7 9 7 9 7 9 7 9 7 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 90 00 90 00 90 00 90 00 90 00 90 00 90 00 19 20 19 20 19 20 19 20 19 20 19 20 19 20 Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (7) advanced economies (7) equity FDI debt reserves Sources: Authors’ calculations based on Lane and Milesi-Ferretti 2007 and Chinn and Ito 2008. Note: This �gure shows in panel a the evolution of a de jure �nancial globalization measure. Panels b and c show the percentage change of de facto �nancial globalization measures for the 1990–99 and 2000–07 periods. Panel b shows changes in liabilities holdings over GDP. Panel c shows changes in assets hold- ings over GDP. It also shows that, since the late 1990s, LAC have risen further during the past decade. countries are the most integrated emerging This is consistent with LAC-7’s less dynamic region, followed closely by Eastern European equity markets (see chapter 7). economies during most of the 2000s, and sig- On the asset side (measuring the invest- nificantly ahead of Asian economies, includ- ments of domestic agents abroad), bank ing China and India. and equity investments have been the fast- When focusing on the change of exter- est growing components of outflows when nal liabilities (measuring the investments of measured as a share of GDP (figure 4.4c). In foreigners in domestic economies as a share net terms, LAC-7 countries have experienced of GDP), LAC-7 portfolio equity and debt declining debt liabilities and increasing equity positions have been on the rise. While FDI liabilities. In fact, they have become net cred- has declined, it still represents an important itors with respect to the rest of the world as share of the liabilities (figure 4.4b). The fact regards debt contracts and have acquired that financial diversification in LAC-7 is still an increasing net debtor position as regards largely based on FDI, rather than on invest- equity contracts, particularly through FDI ment in portfolio equity, stands in sharp con- (figure 4.5). These trends are matched across trast with other emerging economies (except all subregions in LAC (figure 4.6). for Eastern Europe), where portfolio equity These changes in portfolio composi- inflows have been much higher and often tion—from a net debtor to a net creditor 66 net debtor net creditor net debtor net debtor net creditor 0 5 0 –35 –30 –25 –20 –15 –10 –5 10 –50 –45 –40 –35 –30 –25 –20 –15 –10 –5 –80 –70 –60 –50 –40 –30 –20 –10 0 10 20 19 19 19 9 9 9 19 0 19 0 19 0 9 91 9 19 1 19 19 1 9 9 9 19 2 19 2 19 2 9 9 9 19 3 19 3 19 3 9 9 94 19 4 19 4 19 95 95 9 19 19 19 5 9 9 9 19 6 19 6 19 6 9 9 9 19 7 19 7 19 7 98 98 9 19 19 19 8 9 9 9 year year year a. Asia e. India 20 9 20 9 20 9 0 0 0 net debtor 20 0 20 0 20 0 c. Eastern Europe 0 0 0 20 1 20 1 20 1 0 0 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 19 02 9 20 2 0 20 2 0 20 0 19 0 9 20 3 0 20 3 0 20 3 0 19 1 net equity position 9 20 4 20 4 20 4 19 2 05 0 0 Source: Authors’ calculations based on Lane and Milesi-Ferretti 2007. 9 20 20 5 20 5 19 3 0 06 06 9 20 6 20 20 19 4 07 07 07 9 19 5 9 19 6 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 9 net debtor net creditor net debtor net creditor net debtor net creditor 19 7 9 19 8 –40 –30 –20 –10 0 10 20 30 40 50 60 –35 –30 –25 –20 –15 –10 –5 0 5 10 –20 –15 –10 –5 0 5 10 15 20 9 year 19 19 net debt position 20 9 9 9 19 0 19 0 9 20 0 19 0 19 0 0 91 9 9 20 1 19 19 1 19 1 0 9 92 9 20 2 19 2 19 19 2 g. Other advanced economies 0 9 9 9 20 3 19 3 19 3 19 3 9 9 FIGURE 4.5 Net foreign assets as percentage of GDP: Equity and bonds 04 94 20 19 4 19 4 19 0 9 9 9 20 5 19 5 19 5 19 5 0 96 9 9 20 6 19 19 6 19 6 07 9 9 97 19 7 19 7 19 9 9 9 19 8 19 8 19 8 9 year year year Note: This �gure shows the international net equity and net debt positions as a percentage of GDP between 1990 and 2007. 99 9 d. G-7 f. LAC-7 b. China 20 9 20 20 9 0 0 0 20 0 20 0 20 0 0 0 0 20 1 20 1 20 1 0 0 net debt position (except reserves) 20 2 20 2 0 0 0 20 2 20 3 20 3 0 0 0 20 3 20 4 20 4 04 0 0 20 20 5 20 5 0 20 5 20 06 20 06 0 07 07 20 6 07 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 67 FIGURE 4.6 Within LAC: Net foreign assets as percentage of GDP: Equity and bonds a. Caribbean b. Central America 0 net creditor 20 10 –10 0 –20 –10 –30 net debtor –20 –40 net debtor –30 –50 –40 –50 –60 –60 –70 –70 –80 –80 –90 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 07 19 1 19 2 93 19 4 19 5 96 19 7 98 20 9 20 0 20 1 20 2 03 20 4 20 5 06 07 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 9 9 9 9 9 9 0 0 0 0 0 19 19 19 19 19 20 20 year year c. LAC-7 d. South America net creditor 10 30 net creditor 5 20 0 10 –5 0 –10 –10 net debtor –15 –20 net debtor –20 –30 –25 –40 –30 –50 –35 –60 19 0 19 1 19 2 93 19 4 19 5 19 6 19 7 19 8 20 9 00 20 1 20 2 03 20 4 20 5 20 6 07 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 04 20 5 20 6 07 9 9 9 9 9 9 9 9 9 0 0 0 0 0 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 19 19 20 20 19 20 year year net equity position net debt position net debt position (except reserves) Source: Authors’ calculations based on Lane and Milesi-Ferretti 2007. Note: This �gure shows the international net equity and net debt positions as a percentage of GDP between 1990 and 2007. position—have resulted in a much safer down the appreciation of the domestic cur- form of financial integration, thus avoiding rency during the precrisis expansionary many of the downsides of financial global- period and then served as a self-insurance ization.10 In LAC’s not-too-distant past, the mechanism during the crisis, deterring cur- devaluations that typically accompanied rency and banking panics. In fact, when financial crises increased the burden of for- the global crisis erupted, many countries eign currency debt. On top of this, market held international reserves in excess of shutdowns triggered rollover crises because their stock of short-term foreign liabili- of the high incidence of short-term debt. In ties. This, in practice, eliminated concerns contrast, during the recent global crisis, the about debt rollover difficulties, limiting devaluations have implied an improvement investors’ incentives to attack the domestic (when measured in local currency) in the currencies.11 external positions of LAC countries due to their net creditor stance. Moreover, external liabilities were reduced when equity prices Financial offshoring plummeted, thereby shrinking net debtor equity positions. At the same time, the During the 1990s, the LAC-7 countries large pools of international reserves slowed spearheaded a strong process of financial 68 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? FIGURE 4.7 New capital-raising issues in foreign markets offshoring. However, this expansion in the use of foreign markets, including bonds and a. New syndicated loans syndicated debt, appears to have generally 6 stabilized over the past decade. In terms 5.6 5.3 of flows, new syndicated loans expanded 5 (figure 4.7, panel a). However, private bond 4 3.6 issues declined strongly (figure 4.7, panel b). % of GDP 3 2.9 Similarly, there was a substantial decline in equity issues abroad (figure 4.7, panel c). In 2.0 2 1.7 1.5 1.6 relative terms however, the foreign equity 1.4 1 1.2 issues increased substantially as a propor- 0.7 0.4 0.4 0.3 tion of the total (figure 4.8, panel c). At the 0 same time, while the foreign share of total Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (7) advanced private bonds outstanding has remained economies (7) stable, there has been a significant decline 1991–99 2000–08 in the foreign share of total public bonds b. Bond issues by the private outstanding, suggesting that public financ- and public sectors 9 ing has shifted towards domestic markets 8 (figures 4.8, panel a and 4.8, panel b). Such 7 1.4 a decline for public bonds is consistent with 6 1.1 the significant expansions of the local mar- % of GDP 5 3.3 4 kets for government bonds and, at the same 3 1.1 0.9 5.7 6.3 time, a decline in LAC governments’ foreign 2 1.4 1.6 1.0 3.3 indebtedness. 1 2.1 1.6 2.3 1.5 0 1.0 0.1 0.3 0.1 0.4 0.7 0.3 0.3 1.2 0.4 The positive developments in currency and maturity for domestic bond markets docu- 9 8 9 8 9 8 9 8 9 8 9 8 9 8 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 –9 –0 91 00 91 00 91 00 91 00 91 00 91 00 91 00 mented in chapter 3 were largely matched 19 20 19 20 19 20 19 20 19 20 19 20 19 20 Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (7) advanced offshore. Thus, the maturity of LAC-7 pri- private public economies (7) vate sector foreign bonds reached nearly 12 c. Equity issues years in the past decade (up from about 7 years in the previous decade), whereas that 0.8 0.7 for the public sector increased from 8 to 0.7 11 years (figure 4.9).12 Similar increases in 0.6 maturity were observed across all of LAC 0.5 0.5 (figure 4.10). At the same time, both the % of GDP 0.4 public and private sectors were able to issue 0.3 0.3 offshore bonds in local currencies. Thus, for 0.2 0.2 0.2 0.2 0.2 the LAC-7 countries, 7 percent of private 0.2 0.1 0.1 0.2 0.1 0.1 sector bonds and 9 percent of public sector 0.1 0.0 0.0 bonds issued abroad in the past decade were Asia (5) China Eastern G-7 (7) India LAC-7 (7) other denominated in local currency, against vir- Europe (7) advanced economies (7) tually zero during the 1990s (figure 4.11). 1991–99 2000–08 While these amounts are small, especially when compared to the advanced economies, Source: Didier and Schmukler 2011b. Note: This �gure shows capital-raising activity in foreign markets. Panel a shows new syndicated they signal the start of LAC’s overcoming loans abroad as a percentage of GDP between 1991 and 2008. Panel b shows bond issuance abroad of original sin (that is, the inability to issue by the private and public sectors as a percentage of GDP between 1991 and 2008. Panel c shows capital-raising equity issuance in foreign markets as a percentage of GDP between 1991 and 2008. local currency long-term debt in foreign Numbers in parentheses show the number of countries in each region. markets). Clearly, they also point to a long road toward full redemption. FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 69 The issuance and trading of equity abroad FIGURE 4.8 Relative size of foreign capital markets by LAC-7 countries (as well as many emerg- ing economies) has usually taken the form a. Bonds outstanding in foreign markets by the public sector as a percentage of total outstanding bonds of cross-listings through depository receipts. 45 The offshore trading of LAC-7’s depository 40 40 receipts accounts for nearly two-thirds of 35 32 31 total trade, compared to between 30 percent 30 29 % of total bonds and 40 percent for other emerging econo- 25 22 mies (figure 4.12). Empirical work suggests 20 19 16 that such offshoring of LAC equity markets 15 11 has tended to limit the role of onshore mar- 10 8 6 kets, which, as flagged in chapter 3, have 5 2 4 remained extremely illiquid.13 The fact that 0 Asia (5) China Eastern G-7 (7) LAC-7 (6) other advanced access to the offshore equity market is even Europe (6) economies (6) more limited than onshore—only a very b. Bonds outstanding in foreign markets by the private sector small number of LAC firms have access to as a percentage of total outstanding bonds 70 foreign bond and equity markets and, sur- 61 prisingly, this number has declined further 60 55 58 during the last decade (figure 4.13, panels a 51 % of total private bonds 50 46 44 45 44 and b)—is an additional source of concern. 40 36 This issue is explored in further detail in 36 32 29 chapter 7. 30 22 Another important aspect of finan- 20 cial globalization (a sort of “offshoring in 10 7 reverse�) is the large physical presence of 0 foreign-owned banks in LAC’s domestic Asia (5) China Eastern G-7 (7) India LAC-7 (6) other financial systems. The increased presence Europe (2) advanced economies (5) of foreign banks was largely a response to 1991–99 2000–09 the financial crises of the 1990s and affected c. Amount raised in foreign equity markets as a percentage nearly all countries in LAC. Thus, between of total amount raised through equity issues 1997 and 2001, the share of foreign-owned 60 bank assets increased from 23 percent to 50 49 47 43 percent in the LAC-7 countries and from 44 % of total amount raised 26 percent to 38 percent in other South 40 32 32 32 American countries (figure 4.14). In Cen- 30 28 tral America and the Caribbean, the rise 24 20 19 in foreign bank participation took place a 20 16 16 12 little later (during the early 2000s), with the 10 6 asset share of foreign banks growing from 20 percent to 31 percent and from 6 percent 0 Asia (5) China Eastern G-7 (7) India LAC-7 (6) other to 32 percent, respectively. As argued by Europe (7) advanced Domanski (2005), letting the foreign bank- economies (7) ers in would facilitate the recapitalization, 1991–99 2000–08 consolidation, and strengthening of troubled Source: Didier and Schmukler 2011b. banking systems. These objectives ended Note: This �gure shows the size of foreign markets relative to total (domestic and foreign) markets. up trumping nationalistic concerns as well Panel a shows the ratio of outstanding public bonds in foreign markets to total public bonds between 1990 and 2009. Panel b shows the ratio of outstanding bonds by the private sector in as concerns about the potentially adverse foreign markets to total public bonds between 1990 and 2009. Panel c shows the amount raised in effects of foreign banks on access to finan- equity market abroad relative to total amount raised through equity issues in domestic and foreign markets between 1991 and 2008. Panels a and b are based on data from the Bank for International cial services (see box 4.1). Settlements that de�nes international debt securities as those that have not been issued by resi- dents in domestic currency and targeted at resident investors. 70 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? FIGURE 4.9 Average maturity of bonds at issuance a. Average maturity of private bonds in foreign markets 14 11.82 12 10 8.78 8.57 8.81 8.44 number of years 7.73 7.85 7.59 8 7.50 7.35 7.15 6.11 6 4 2 0 Asia (5) China Eastern G-7 (7) LAC-7 (7) other advanced Europe (7) economies (7) b. Average maturity of public bonds in foreign markets 12 11.0 10.2 10 9.5 8.7 9.0 8.5 8.1 7.9 7.8 7.6 8 7.5 number of years 6.4 6 4 2 0 Asia (5) China Eastern G-7 (7) LAC-7 (6) other advanced Europe (7) economies (7) 1991–99 2000–08 Source: Didier and Schmukler 2011b. Note: This �gure shows the maturity (in years) of bonds issued by the private and public sectors in foreign markets between 1991 and 2008. Numbers in parentheses show the number of countries in each region. FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 71 FIGURE 4.10 Within LAC: Average maturity of public and private bonds in foreign markets at issuance a. Maturity of foreign private bonds 12 10.6 10 9.4 8.7 8 7.2 number of years 6.8 7.1 6 4.8 4 3.1 2 0 Central America LAC-7 (6) offshore centers South America (2) (includes the in LAC (1) Dominican Republic) (3) b. Maturity of foreign public bonds 18 16 15.3 14 12.8 12 11.6 11.0 number of years 10 9.0 8.9 7.9 8.1 8 6.0 6.0 6 4 2 0 Caribbean (1) Central America LAC-7 (6) offshore centers South (includes the in LAC (6) America (1) Dominican Republic) (4) 1991–99 2000–08 Source: Didier and Schmukler 2011b. Note: This �gure shows the maturity (in years) of bonds issued by the private and public sectors in foreign markets between 1991 and 2008 across LAC. Numbers in parentheses show the number of countries in each region. 72 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? FIGURE 4.11 Ratio of foreign currency bonds to total bonds at issuance a. Private sector 100 98 100 100 100 97 94 93 90 89 89 % of total foreign bonds by private sector 80 75 70 67 60 50 45 40 30 20 10 23 19 12 18 3 7 23.6 38.2 24 8 38 106 0 Asia (5) China Eastern Europe G-7 (7) LAC-7 (7) other advanced (6) economies (7) b. Public sector 100 100 100 100 99 100 100 96 91 91 90 85 % of total foreign bonds by public sector 80 75 70 60 58 50 40 30 20 10 5.0 8.4 2.7 2.0 3.9 3.0 43.6 65.3 7.2 5.1 22.8 18.2 0 Asia (5) China Eastern Europe G-7 (7) LAC-7 (6) other advanced (7) economies (7) 1991–99 2000–08 Source: Didier and Schmukler 2011b. Note: This �gure shows the ratio of foreign-currency-denominated bonds to total bonds issued by the private and public sectors in foreign markets between 1991 and 2008. Panel a shows data for the private sector. Panel b shows data for the public sector. Numbers in the base of the bars represent the average number of issues per year. Numbers in parentheses show the number of countries in each region. FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 73 FIGURE 4.12 Equity trading in domestic and foreign markets 70 66 63 61 60 50 % of total value traded 40 38 39 38 35 35 30 31 30 23 24 22 20 18 18 18 14 13 10 10 8 8 2.5 3.3 3.7 12.3 14.0 15.0 1.0 1.3 1.7 9.9 11.2 11.6 7.3 10.0 11.0 9.7 11.7 13.0 2.8 3.0 3.0 0 Asia (3) China Eastern G-7 (5) India LAC-7 (6) other advanced Europe (2) economies (7) 2000–03 2004–07 2008–09 Source: Didier and Schmukler 2011b. Note: This �gure shows the cross-country averages of �rm-level value traded in depository receipts over total value traded (in domestic markets and DRs). Only �rms with DR programs identi�ed in the DR Directory of the Bank of New York and with trading data reported in Bloomberg are considered in this �gure. The average number of �rms for countries within each region is also reported at the bottom of the bars. Numbers in parentheses show the number of countries considered in each region. BOX 4.1 Costs and bene�ts of the participation of foreign banks in domestic �nancial systems Numerous papers have analyzed the costs and ben- access to � nance for a majority of domestic � rms e�ts of the participation of foreign-owned banks in and consumers if they only concentrate on a selected domestic � nancial systems (for example, Claessens segment of the market. In particular, the literature and Van Horen 2008; Cull and Martínez Pería 2010). shows that foreign banks tend to have dif�culties in They generally � nd that foreign-owned banks tend lending to borrowers that lack the hard information to be more ef�cient than domestic banks in emerging to prove their creditworthiness (for example, Berger, economies. Foreign-owned banks may count on more Klapper, and Udell 2001; Mian 2003, 2006). diversi�ed funding bases, including access to external However, the evidence is mixed on the overall liquidity from their parent banks even in times of cri- effect of foreign-owned banks on outreach. For exam- sis, which lowers their deposit costs. Foreign-owned ple, competition from foreign banks could conceivably banks can also better capture scale economies and compel domestic banks to pursue new market niches bring technical skills, product innovation, and more (Jenkins 2000). Empirically, de la Torre, Martínez sophisticated risk management practices. They can Pería, and Schmukler (2010) and Beck, Demirgüç- help boost competition in developing countries. Kunt, and Martínez Pería (2011) show that there is However, critics of foreign-owned bank entry no difference in the extent of foreign-owned banks’ argue that foreign banks can destabilize the local involvement with small and medium enterprises banking sector for a number of reasons. They can relative to large domestic and government-owned “import� shocks from their home countries and/or banks. Clarke, Cull, and Martínez Pería (2006) and spread shocks from other developing countries in Giannetti and Ongena (2009) � nd that foreign bank which they operate. Fierce competition with foreign- presence is associated with an improvement in � rm owned banks can threaten the survival of the local indicators, particularly for young � rms, though the banks. In addition, foreign-owned banks can reduce effects are less pronounced for small �rms. 74 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? FIGURE 4.13 Concentration in foreign bond and equity markets a. Amount raised by top five issues as a percentage of total amount raised in foreign private bond markets 100 88 90 80 74 76 % of total amount raised abroad 73 72 69 70 63 59 57 60 49 48 50 45 40 31 30 24 20 10 32 21 21 22 8 18 25.0 45.3 13 47 45 16 43 11.5 0 Asia (5) China Eastern G-7 (7) India LAC-7 (4) other Europe (4) advanced economies (7) b. Amount raised by top five issues as a percentage of total amount raised in foreign equity markets 98 100 98 100 89 87 89 90 85 85 87 84 82 80 79 % of total amount raised abroad 70 65 60 56 50 40 30 20 10 7 6 21 93 5 11 14 26 18 14 11 6 9 24 0 Asia (2) China Eastern G-7 (7) India LAC-7 (3) other Europe (1) advanced economies (7) 1991–99 2000–08 Source: Didier and Schmukler 2011b. Note: This �gure shows the concentration in bond and equity markets. Panel a shows the amount raised by top �ve bond issues by corporations as a percentage of the total amount raised by the private sector in foreign private bond markets between 1991 and 2008. Only country-years with at least �ve issues were considered in this �gure. Panel b shows the amount raised by top �ve equity issues as a percentage of the total amount raised by the private sector in foreign private equity markets during the same period. Panel c shows the cross-country average of �rm-level value traded in DRs for the top �ve �rms over total DR value traded in foreign equity markets between 2000 and 2008. Only countries with more than �ve �rms with DRs programs were considered in this �gure. All DRs identi�ed in the DR Directory of the Bank of New York with trading data reported in Bloomberg are considered in this �g- ure. Numbers in the base of the bars represent the yearly cross-country average number of issues. Numbers in parentheses show the number of countries considered in each region. FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? 75 FIGURE 4.14 Foreign-owned bank assets as a percentage of total banks assets a. Across LAC countries 70 65 67 60 50 % of total bank assets 46 43 40 38 36 39 32 31 30 26 23 20 20 14 10 10 6 0 ) bl nic es 7) ) C rs (2 (4 LA te 7( pu mi ud ic) an ) an ica in cen (3 C- Re Do incl be er LA ) (6 e m rib e ( or th rica hA Ca sh e ut off m So lA ra nt Ce b. Worldwide 70 63 60 57 50 % of total bank assets 46 43 40 30 25 26 23 20 19 18 19 11 10 7 4 3 4 1 0 0 0 ) a (7 e 7) 7) ies ced (5 in p 7( 7( ro Ch ) ) ia om an (7 G- C- Eu As on dv LA rn ec er a ste h Ea ot 1997 2001 2005 c. Individual LAC countries, 2005 100 90 80 % of total bank assets 70 60 50 40 30 20 10 0 in ca te or an pu la To ic Co om o sta bia Ar Bra a Ho nti l ne du a ela s Ni Ch B ra e Bo ua Pa livia ug a ym Jam uay Is ca El rag s lva y M dor ico rb ru os ge zi zu ra Pa land Sa ua c Ve n n Ur am ca il l g ,R d bl ad Re a Tr ini Gua uad Ba Pe Ri an ai ad g Co ba id n m ex n Ec Ca m Do Source: Authors’ calculations based on Claessens and van Horen (2008). Note: This �gure shows the average assets of foreign-owned banks as share of total bank assets in 1997, 2001, and 2005. Panel a compares the average ratio across regions within LAC, panel b compares the average ratio across LAC-7 against other emerging regions and developed economies, and panel c shows the ratio for every LAC country in 2005. 76 FINANCIAL GLOBALIZATION: WHERE DOES LAC STAND? Notes 5. There is a large and rapidly growing litera- ture analyzing cross-border capital flows and 1. This chapter draws heavily on the papers associated asset and liability positions from a “Benchmarking Financial Globalization in the portfolio perspective. See Broner et al. (2010) Latin America and the Caribbean,� by Tatiana and the references therein. Didier and Sergio Schmukler (2011b), and 6. When domestic and offshore activities are “Financial Globalization in Latin America: complements, the correlation between �nan- Myth, Reality and Policy Matters,� by Eduardo cial development and �nancial globalization Levy-Yeyati (2011b), which are part of the should be positive, and when they are substi- forthcoming Edited Volume that accompanies tutes, the correlation should be negative. this LAC flagship report. 7. See Levine and Schmukler (2006, 2007). 2. Foreign investors mostly invest in LAC mar- 8. See also Lane and Milesi-Ferretti (2001). kets but do not raise capital there. Local issues 9. See Gourinchas and Rey (2007) and of equity or bonds by foreign �rms accounted Gourinchas, Govillot, and Rey (2010). for less than 2 percent of the total capital 10. See de la Torre, Calderon et al. (2010a); de la raised in these markets during the 2000s. Torre, Calderon et al. 2010b; Didier, Hevia, Yet foreign �rms have generally been seeking and Schmukler (2011); Gourinchas, Govillot, more �nancing in emerging economies. and Rey (2010); and Gourinchas, Rey, and 3. See Stulz (1999) and Errunza (2001). Truempler (2011). 4. Levy-Yeyati (2011b) �nds evidence that �nan- 11. See Aizenman (forthoming), Aizenman and cial globalization ampli�ed the asset sell-off Pasricha (2010), Frankel and Saravelos (2010), (particularly through benchmarked global among others. equity funds) that occurred after the collapse 12. However, the long maturities could be associ- of Lehman Brothers. He also concludes that ated with relatively short durations if most of the procyclical nature of portfolio inflows, the debt is at floating rates. The data do not which return to core markets in episodes currently allow us to identify such effects. of flight to quality, can amplify the effect of 13. See Levine and Schmukler (2006, 2007). the global cycle on the emerging world in an Chapter 7 analyzes in more detail the possible undesirable way. reasons for this gap. Financial Inclusion: Where Does LAC Stand? 5 T his chapter considers financial inclu- number of dimensions, including, surpris- sion in Latin America and the Carib- ingly, the use of banking services by small bean (LAC) countries, focusing on the and medium enterprises (SMEs). typically underserved, particularly lower- • This relative success could reflect in part income households and small entrepreneurs. the fact that a majority of governments in Both theoretical and empirical studies have LAC-7 have adopted comprehensive poli- argued that financial inclusion matters.1 cies to promote �nancial inclusion. Theory has shown that financial market fric- • At the same time, lack of demand appears tions that prevent financial inclusion can lead to be an important factor explaining to persistent inequality and poverty traps, the limited use of banking services by and empirical work has confirmed the posi- households. tive welfare effects that result from firms and • However, higher pecuniary fees on deposit individuals gaining access to financial ser- and lending products could also be acting vices. 2 This chapter first measures financial as barriers. inclusion in LAC and compares it to that in • In addition to such barriers, another other regions using supply-side data obtained potentially problematic area that requires from bank regulators. Second, it analyzes governments’ attention is the improve- barriers (pecuniary and nonpecuniary) to ment of creditor rights, where LAC-7 lags the use of banking services from a survey of signi�cantly. banks and compares LAC to other regions. • By and large, non-LAC-7 lags LAC-7 in Third, it studies demand-side data from firm- some indicators of � nancial outreach and level surveys and household-level surveys. 3 in some government policies to promote Finally, it analyzes the role of governments in inclusion; thus, there is scope for non- the region in promoting financial inclusion. LAC-7 governments to play a more active Main findings are as follows: role in fostering � nancial inclusion. • Most available indicators of financial This chapter does not address the issue inclusion suggest that LAC-7 is not obvi- of financial sustainability, partly for lack of ously underperforming compared to its data. Nonetheless, it seems safe to say that peers; in fact, LAC-7 outperforms in a the more governments promote (and succeed 77 78 FINANCIAL INCLUSION: WHERE DOES LAC STAND in enhancing) financial inclusion, the more Lack of access can reflect pecuniary bar- attention they need to pay to the soundness riers to the use of financial services, such as and sustainability of such progress. As dem- fees or minimum balances, as well as non- onstrated by the recent global crisis, systemic pecuniary barriers, such as documentation fault lines are more likely to develop where requirements, the number of locations where borrowers’ capacity to repay is stretched individuals can open accounts or apply for thinner. loans, the number of days to process a loan The rest of this chapter is structured as application, and so on. Thus, the chapter uses follows. The first section presents the meth- data from a survey of financial institutions odology used in the study. The second sec- conducted by the World Bank (see Beck, tion presents the supply-side evidence. The Demirgüç-Kunt, and Martínez Pería 2008) third section reviews the demand-side evi- during 2004–05 to quantify barriers to the dence. The final section discusses the role of use of financial services. Because these data government. are available only for the largest countries within Latin America, it is not possible to compare LAC-7 countries to the other coun- Methodology tries in the region. Financial inclusion is hard to measure in There are a number of caveats and limi- practice. A basic challenge is to distinguish tations to this analysis. First, as mentioned access to financial services from the use above, the analysis centers on banks and of financial services. Access refers primar- largely ignores nonbank institutions. Sec- ily to the supply of services whereas use ond, it focuses on payments, savings, and (the observable outcome) is determined by credit services. Because of a lack of data, it demand as well as supply. Therefore, infer- ignores insurance services. Third, indicators ences about financial inclusion from mea- of financial inclusion, such as the number sures of use of financial services do not of deposits or loans per capita, may over- necessarily imply the existence of market estimate the extent of outreach, since some failures that warrant government interven- individuals and firms might have more than tion. Nonetheless, measuring the use of one account. Fourth, the supply-side data financial services is the first and most read- do not distinguish between individuals and ily available way to assess financial inclu- firms in the use of banking services. Finally, sion. This analysis employs both types of in documenting the role of the government in data—supply-side data and demand-side promoting financial inclusion, this chapter is (survey-based) data—to characterize finan- able to describe only efforts and policies, and cial inclusion. As in previous chapters, it does not conduct a welfare analysis of the whenever there are sufficient data, both the impact of these policies. raw data and the data controlled for gross domestic product (GDP) per capita and population density are presented. Supply-side evidence The analysis concentrates on analyz- Overall, the raw statistics on the num- ing access to (and use of) banking services, ber of branches, ATMs (automated teller because banks dominate the financial sector machines), and deposits suggest that LAC in Latin America, especially for households countries are at par with economies in Asia and SMEs, which are at the core of the finan- but lag behind developed countries and cial inclusion concerns and agenda. More- Eastern Europe. The median numbers of over, data for the banking sector are more branches and ATMs in LAC-7 are broadly readily available across countries, facilitating similar to those in Asia but smaller than in comparisons. However, wherever possible, Eastern Europe, the G-7 and other devel- data and initiatives that include nonbanks are oped countries (figure 5.1, panel a). The presented and discussed. LAC-7 countries also outperform other FINANCIAL INCLUSION: WHERE DOES LAC STAND 79 FIGURE 5.1 Number of branches, ATMs, and deposit and loan accounts a. Median number of bank branches and ATMs per 100,000 adults 140 120 118 100 number per 100,000 adults 80 73 (6) 60 54 40 34 37 (7) (7) 32 27 24 22 20 (4) (7) 18 11 10 (3) 13 (7) (7) (7) 10 7 7 (3) (5) (6) (7) (4) 0 ia ub es pe a 7 7 ica om ced di G- C- As ep lud ro ) er In LA lic ies on n Eu m n R (inc ec dva hA rn ra ica a ste ut in ric he So Ea om me ot eD lA th ntra Ce median branches per 100,000 adults median ATMs per 100,000 adults b. Median number of bank deposit accounts and loan accounts per 1,000 adults 2,500 2,022 2,000 1,700 number per 1,000 adults 1,500 1,000 977 (5) 906 918 825 (6) 747 633 498 500 (4) 439 453 (5) 357 (7) (2) 272 137 (6) (2) (4) (2) (2) (3) 0 ia ub es pe a 7 7 ica om ced di G- C- As ep lud ro ) er In LA lic ies on n Eu m n R (inc ec dva hA rn ra ica a ste ut in ric he So Ea om me ot eD lA th ntra Ce median deposit accounts per 1,000 adults median loans per 1,000 adults Source: Martínez Pería 2011. Note: Panel a shows the median number of bank branches and ATMs per 100,000 adults in LAC-7 versus other regions and other countries in Latin America. Panel b shows the median number of bank deposit accounts and loan accounts per 1,000 adults in LAC-7 versus other regions and other countries in Latin America. The numbers in parentheses refer to the number of countries for which data are available. 80 FINANCIAL INCLUSION: WHERE DOES LAC STAND LAC subregions. The median number of fees against a number of possible determi- deposit accounts per 1,000 adults is also nants, including a LAC-7 dummy, confirms broadly similar in the LAC-7 countries that even after controlling for differences in to that in Asia and the non-G-7 advanced banking sector structure, the institutional economies, but below Eastern Europe and environment, and per capita income across the G-7 countries (figure 5.1, panel b). How- countries, fees charged by banks in Latin ever, after controlling for per capita income America are still considerably higher than the and population density, these differences are benchmark. no longer significant (see table 5.1). In fact, most LAC-7 (and LAC) countries appear to outperform the comparator groups (fig- Demand-side evidence ure 5.2). In the case of the number of loans, This section characterizes demand-side evi- LAC outperforms most other regions, even dence on firms’ use and access to banking in the absence of any control (figure 5.3). 5 services with a number of indicators con- There are several possible pecuniary bar- structed from the World Bank Enterprise riers to access. The median minimum bal- Surveys database. First, it examines the per- ances required by banks in LAC-7 countries centage of firms that have a deposit account. are generally in line with what is observed Second, it examines the use of credit prod- for most developing countries. Similarly, the ucts using an indicator variable constructed minimum balances required for maintaining so that it equals one if the enterprise has an savings and checking accounts are lower than overdraft, loan, line of credit, or any bank those required in most developing countries financing for working capital or fixed asset and are in accordance with practices in devel- purchases. It also looks at the median per- oped economies. However, deposit fees in centage of working capital and, separately, LAC tend to be higher than those observed in fixed assets financed by banks. Finally, it other regions (figure 5.4, panel a). Similarly, evaluates the share of firms that consider fees on consumer and residential (mortgage) access to finance to be a major obstacle to loans in Latin America significantly exceed their operations and growth. Across the those in most other comparator countries analysis, it distinguishes between large (figure 5.4, panel b). By contrast, fees on other firms (those with 100 or more employees) types of loans in LAC-7 countries are quite and SMEs (those that employ between 5 and close to those observed in other regions. 99 workers). Because SMEs tend to be more In terms of nonpecuniary barriers, the opaque and more vulnerable to economic number of documents required to open volatility, they are generally expected to be deposit accounts in Latin America also more constrained when it comes to access- exceeds the requirements in most other ing banking services. countries. However, the number of locations Not too surprisingly, the vast majority of where bank customers can open a deposit firms (including SMEs) in LAC (as in other account or apply for a loan is comparable or regions) have a bank account (figure 5.5, larger in LAC-7 countries, even compared panel a). Across large firms and SMEs, the to G-7 economies. The World Bank survey use of bank credit in LAC is more pervasive reveals that with the exception of residential than in Asia and Eastern Europe (figure 5.5, mortgages, which generally take two weeks panel b). Within LAC, the use of credit is to process, the number of days required to more widespread among firms in LAC-7 rela- process other loans in LAC-7 is in line with tive to firms in South and Central America. what is observed in other developing regions. There is, however, a noticeable difference Thus, the main barriers to the use of ser- (across all regions) between the share of large vices in Latin America appear to be pecuniary and small firms that use credit products. costs or fees. Table 5.2, which shows regres- While SMEs in the LAC-7 countries also sions of deposit and residential mortgage loan appear to make the most extensive use of TABLE 5.1 Regressions of indicators of financial inclusion on income, population density, and country-group dummies Log of Log of Log of Log of number of deposits per number of loans per branches per 100,000 adults ATMs per 100,000 adults 1,000 adults 1,000 Adults (1) (2) (3) (4) (5) (6) (7) (8) Log of GDP per capita, PPP 0.691 0.713 1.187 1.189 0.91 0.931 1.236 1.254 [17.98]*** [15.31]*** [15.95]*** [14.97]*** [14.38]*** [13.91]*** [14.93]*** [15.92]*** Log of population density 0.114 0.115 0.046 0.042 0.333 0.345 0.078 0.082 [2.65]*** [2.61]** [0.81] [0.72] [5.17]*** [5.14]*** [1.19] [1.23] LAC –0.082 0.396 0.226 0.648 [–0.43] [2.27]** [1.52] [4.48]*** LAC-7 0.086 0.285 0.417 0.716 [0.61] [1.21] [1.68]* [2.34]** REST of LAC 0.012 0.486 0.363 0.811 [0.04] [1.73]* [1.42] [2.50]** Other countries 0.164 –0.001 0.199 0.157 [1.08] [–0.01] [0.99] [0.54] Constant –4.229 –4.551 –7.735 –7.731 –3.007 –3.400 –6.1 –6.406 [–10.32]*** [–8.34]*** [–10.44]*** [–9.41]*** [–4.55]*** [–4.42]*** [–7.48]*** [–7.83]*** Observations 132 132 117 117 109 109 77 77 R-squared 0.667 0.67 0.796 0.796 0.765 0.767 0.833 0.834 F-test LAC-7 = LAC_Rest 0.0578 0.405 0.0562 0.239 P-value LAC-7 = LAC_Rest 0.81 0.526 0.813 0.626 F-test LAC-7 = Other 0.477 1.769 1.429 14.10 P-value LAC-7 = Other 0.491 0.186 0.235 0.000*** F-test LAC_Rest = Other 0.245 3.740 0.724 12.89 P-value LAC_Rest = Other 0.621 0.056* 0.397 0.001*** Source: Martínez Pería 2011. Note: PPP = purchasing power parity. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent. Robust t-statistics are shown in brackets. 81 82 FINANCIAL INCLUSION: WHERE DOES LAC STAND FIGURE 5.2 Actual versus predicted number of deposits per bank loans to finance their purchases of 1,000 adults fixed assets, the difference in the extent of use between large firms and SMEs is remark- 3,500 ably large compared to that in Asia or East- 3,000 ern Europe (figure 5.6, panel a). Broadly sim- actual accounts per 1,000 adults ilar results are found for the share of working 2,500 capital financed by banks (figure 5.6, panel b). The final indicator of access to credit 2,000 among firms is firms’ responses to a sur- 1,500 vey question about their views on access to finance.6 Again, there is not much difference 1,000 between LAC-7, Asia, and Eastern Europe as regards the percentage of SMEs that declare 500 lack of access to be a severe obstacle (figure 0 5.7). In sum, SMEs in LAC (and especially in 0 500 1,000 1,500 2,000 2,500 3,000 3,500 LAC-7) do not appear to be lagging in abso- predicted accounts per 1,000 adults lute terms behind firms in other developing LAC-7 rest of LAC other comparators rest countries as regards their access and use of bank credit. That being said, the substantial Source: Martínez Pería 2011. Note: Predicted numbers were obtained regressing the log of ATMs per 100,000 adults on difference in access between the larger and the log of GDP per capita in constant purchasing power parity terms and the log of population smaller firms in LAC is a potential matter for density. concern that deserves further exploration. Demand-side information at the house- hold level is very limited and hard to com- pare across countries, because the survey instruments and the samples vary from country to country. However, comparable FIGURE 5.3 Actual versus predicted number of loans per household surveys were recently conducted 1,000 adults by the Corporación Andina de Fomento (CAF) in Argentina, Brazil, Colombia, Peru, 1,500 and Uruguay, which is referred to below as the LAC-5 group (see www.caf.com). The surveys revealed that 51 percent of actual loans per 1,000 adults 1,200 households in LAC-5 have a bank account. Among households that do not have an 900 account, the main reasons cited include lack of funds (61 percent) or absence of a job 600 (19 percent). Only 11 percent of households cited “not trusting financial institutions� 300 or “not being able to meet the require- ments to open an account,� and 7 percent 0 complained about high fees. The statistics 0 300 600 900 1,200 1,500 regarding the use of bank accounts among predicted loans per 1,000 adults households in countries outside the LAC-5 LAC-7 rest of LAC other comparators rest are very similar. Approximately 52 percent of households have an account and, among Source: Martínez Pería 2011. those that do not, 72 percent mention lack Note: Predicted numbers were obtained regressing the log of loans per 100,000 adults on the log of GDP per capita in constant purchasing power parity terms and the log of population of money as the main reason. Only 5 per- density. cent of households complain about high FINANCIAL INCLUSION: WHERE DOES LAC STAND 83 FIGURE 5.4 Annual fees a. Median checking and savings accounts’ annual fees 1.6 1.4 1.4 1.2 % of GDP per capita 1.0 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.2 0.2 0.2 0.1 0.0 0.0 0.0 0.0 0.0 0.0 Asia (4) China Eastern G-7 (5) India LAC-7 (7) other Europe (6) advanced economies (4) annual fees, checking account annual fees, savings account b. Median loan fees 2.5 2.0 2.0 1.8 % of GDP per capita 1.5 1.4 1.4 1.4 1.3 1.2 1.2 1.2 1.1 1.0 1.0 1.0 0.9 0.8 0.7 0.7 0.6 0.5 0.5 0.0 0.0 Asia (4) China Eastern G-7 (5) India LAC-7 (7) other Europe (6) advanced economies (4) mortgage loan fee consumer loan fee SME loan fee business loan fee Source: Beck, Demirguc-Kunt, and Martínez Pería 2008. Note: Panel a shows the median checking and savings accounts annual fees as percentage of GDP per capita. Panel b shows the median loan fees as per- centage of GDP per capita. Numbers in parentheses show the number of countries in each region. 84 TABLE 5.2 Regressions for deposit and loan fees Annual fees checking account Annual fees savings account Variables (% of GDP per capita) (% of GDP per capita) Fee mortgage loan LAC-7 2.47 3.014 4.222 2.153 1.203 1.132 2.501 2.348 2.852 [3.30]*** [2.11]** [2.16]** [3.18]*** [2.62]** [2.27]** [1.88]* [1.85]* [1.81]* Countries outside comparator group 4.141 0.892 0.158 1.454 0.431 –0.251 4.608 1.311 0.729 [2.96]*** [1.01] [0.13] [1.95]* [1.15] [–0.56] [1.54] [1.64] [0.69] Concentration (% of assets held by 0.019 0.058 –0.009 0.014 –0.021 –0.013 top 5 banks) [0.72] [1.62] [–0.90] [1.62] [–0.80] [–0.36] Legal rights index 0.242 0.307 0.014 –0.007 –0.052 –0.09 [1.08] [1.01] [0.18] [–0.10] [–0.27] [–0.45] Credit information index –0.409 –0.533 0.137 0.203 –0.049 –0.053 [–1.01] [–1.04] [0.99] [1.72]* [–0.16] [–0.20] Cost of enforcing contracts 0.077 0.068 0.013 0.014 0.029 0.048 [3.11]*** [2.56]** [1.46] [2.61]** [0.84] [1.26] Heritage index of �nancial freedom 0.089 0.074 0.023 0.017 0.053 0.064 [1.84]* [1.14] [1.59] [0.93] [2.12]** [1.41] Log of GDP per capita (PPP) –1.51 –2.157 –0.549 –0.967 –0.665 –1.053 [–1.97]* [–2.31]** [–2.05]** [–3.06]*** [–1.36] [–1.92]* Share of bank assets held by government banks –0.027 –0.011 –0.003 [–0.57] [–1.03] [–0.12] Share of bank assets held by foreign banks 0.030 –0.002 0.001 [1.17] [–0.36] [0.07] Constant –0.906 5.809 9.756 –1.588 2.637 5.699 0.607 5.182 7.384 [–1.38] [0.82] [1.14] [–2.41]** [1.21] [2.15]** [0.93] [1.21] [1.55] Observations 69 59 45 69 59 45 66 58 44 Pseudo R-squared 0.0152 0.122 0.142 0.0265 0.166 0.296 0.00291 0.0552 0.0967 Source: Martínez Pería 2011. Note: This table shows tobit estimations for deposit and mortgage loan fees against country dummies, along with a series of variables proxying for bank structure, institutional environment, and income per capita. In particular, LAC-7 is a dummy that equals 1 for Argentina, Brazil, Chile, Colombia, Mexico, Peru, and Uruguay. Countries outside comparator group is a dummy that takes the value of 1 for countries other than those in the comparator group, which includes G-7 countries (Canada, France, Germany, Italy, Japan, United Kingdom, and United States); other developed countries (Australia, Finland, Israel, New Zealand, Norway, Spain, and Sweden); comparable countries in Asia (China, India, Indonesia, Republic of Korea, Malaysia, Philippines, and Thailand); and Eastern Europe (Croatia, Czech Republic, Hungary, Lithuania, Poland, Russian Federation, and Turkey). Robust t-statistics are shown in brackets. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent. FINANCIAL INCLUSION: WHERE DOES LAC STAND 85 FIGURE 5.5 Firms’ use of bank accounts and credit products a. Bank accounts 98.9 99.8 99.7 99.2 100 97.5 95.1 94.1 90 87.3 80 70 percent 60 50 40 30 Central America Eastern LAC-7 (7) South America (4) (includes Europe (7) the Dominican Republic) (4) % of large firms % of SMEs b. Credit products 100 90 91 89 90 80 79 71 73 70 68 65 60 60 percent 50 48 40 30 20 10 0 Asia (5) Central America Eastern LAC-7 (7) South America (4) (includes Europe (7) the Dominican Republic) (4) large enterprises SMEs Source: Martínez Pería 2011. Notes: Panel a shows the �rms’ use of bank accounts. Panel b shows �rms’ use of credit products. Numbers in parentheses show the number of countries in each region. 86 FINANCIAL INCLUSION: WHERE DOES LAC STAND FIGURE 5.6 Fixed assets and working capital �nanced by banks a. Fixed assets 40 38 35 33 30 28 27 25 24 23 23 22 21 percent 20 15 13 10 5 0 Asia (5) Central America Eastern LAC-7 (7) South America (4) (includes Europe (7) the Dominican Republic) (6) b. Working capital 35 30 29 25 25 20 20 19 19 percent 16 15 15 15 12 10 9 5 0 Asia (5) Central America Eastern LAC-7 (7) South America (4) (includes Europe (7) the Dominican Republic) (6) large firms SMEs Source: Martínez Pería 2011. Note: Panel a shows the percentage of �xed assets �nanced by banks. Panel b shows the percentage of working capital �nanced by banks. Numbers in parentheses show the number of countries in each region. FINANCIAL INCLUSION: WHERE DOES LAC STAND 87 FIGURE 5.7 Percentage of �rms that consider access to �nance as a severe obstacle 30 28 28 27 25 23 22 21 20 16 percent 15 15 14 13 10 5 0 Asia Central America Eastern LAC-7 (7) South America (4) (includes the Dominican Europe (7) Republic) (6) large �rms SMEs Source: Martínez Pería 2011. Note: This �gure shows the percentage of �rms that consider access to �nance as a severe obstacle. Numbers in parentheses show the number of countries in each region. fees, and 13 percent mention not meeting insufficient income or collateral. Among the requirements to open an account. Over- other countries in the region, 29 percent of all, the fact that half of households do not households did not apply because of insuf- use an account appears to be driven mainly ficient income or collateral. Hence, across by demand considerations, such as the lack Latin America, households that do not apply of savings or income. for loans appear to opt out of using credit Loan use is even less typical than the use services mainly because they have a strong of bank accounts. Only about 21 percent aversion to being in debt. of households in LAC-5 have a loan, and Overall, the household data reveal that 62 percent have never applied for one. In the the use of banking services is rather limited case of other LAC countries (Bolivia, Ecua- in Latin America. Importantly, households’ dor, Panama, and República Bolivariana de responses to questions about why they do Venezuela), only 16 percent of households not use services suggest that lack of income have a loan, and 66 percent never applied and self-exclusion play a more important role for one. A large proportion (70 percent in than supply-side considerations, like high LAC-5 and 66 percent in other LAC coun- fees and stringent documentation require- tries) said they did not apply for a loan ments. It is important to note, however, that because they considered borrowing too risky these surveys are based on a small sample of and preferred not to be in debt. In LAC-5, households that reside only in urban areas. only 24 percent did not apply because of Nationally representative surveys that include 88 FINANCIAL INCLUSION: WHERE DOES LAC STAND rural areas might provide a different picture profitable. In this context, correspondent of the level of usage and the reasons behind banks or mobile branches can play a signifi- it. Furthermore, because these surveys were cant role in expanding the access to financial done only for Latin America, it is not pos- services.8 Over the past decade, bank regu- sible to compare their results to what might lators in LAC-7 have started to allow banks be observed in other developing countries. to enter into correspondent banking arrange- ments. Nevertheless, correspondent banking is less common in LAC-7 countries than it is The role of government among economies in Asia, Eastern Europe, Analyzing the role of the government in pro- and other advanced non-G-7 countries (fig- moting financial inclusion is difficult since ure 5.10). Mobile branching, on the other it can encompass so many aspects—from hand, is more common among LAC-7 coun- documenting whether the government has tries than among countries in Asia (except an explicit mandate to promote financial China and India), Eastern Europe, and G-7 inclusion, to examining specific government economies. Again, LAC-7 countries are way programs and/or interventions targeted to ahead of their neighbors in the region when it improve financial inclusion, to evaluating comes to correspondent banking and mobile the adequacy of the financial sector infra- branches (figure 5.11). structure and the contractual environment. Aside from adopting policies targeted to Furthermore, assessing the welfare impact promote outreach among specific groups of government policies designed to promote (such as SMEs, the poor, or rural inhabit- financial inclusion is particularly hard, since ants), governments can influence the extent it requires isolating the impact of these poli- to which financial services are provided by cies from other factors that can also affect financial institutions and used by the popu- welfare. A full evaluation of government lation at large by ensuring that the appro- policies is beyond the scope of this chap- priate financial sector infrastructure and ter. Instead, this chapter focuses on docu- regulations are in place. In particular, the menting the efforts and policies in place to supply and the use of credit services will promote financial inclusion in LAC and on be influenced by the degree to which credit comparing them to those enacted by govern- information is widely available to banks ments in other regions.7 and the extent to which creditors feel that As regards regulators’ de jure mandate to their rights are protected. On the basis of increase financial inclusion (including having an index that measures rules and practices a document laying out its strategy), LAC-7 affecting the coverage, scope, and accessi- outperforms Eastern Europe but lags sig- bility of credit information made available nificantly behind Asia (figure 5.8, panel a). through either a public credit registry or a A rather similar picture appears for de facto private credit bureau, LAC-7 countries come commitments (figure 5.8, panel b). Across the out ahead of any other comparator develop- LAC region, LAC-7 outperforms other subre- ing countries (figure 5.12). However, when gions (figure 5.9). it comes to the legal rights index, which Access to financial services in many devel- measures the degree to which collateral and oping countries is hampered by the lack of bankruptcy laws protect the rights of bor- a widespread network of banking outlets, rowers and lenders and thus facilitate lend- particularly in rural areas, which finan- ing, LAC-7 countries underperform most cial intermediaries generally do not find developed and developing countries. Clearly, FINANCIAL INCLUSION: WHERE DOES LAC STAND 89 FIGURE 5.8 Governments’ commitment to �nancial inclusion a. De jure commitment 100 100 100 100 90 86 80 70 60 percent 50 43 43 40 30 29 29 29 20 14 14 14 10 0 0 0 Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (7) advanced economies (7) countries with strategy document countries with a mandate to promote savings countries with a mandate to promote SME access countries with a mandate to promote rural access b. De facto commitment 100 100 100 90 80 80 80 71 70 60 57 57 57 percent 50 40 30 29 29 29 20 20 14 14 14 14 10 0 0 0 0 0 0 0 0 0 0 Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (7) advanced economies (7) countries with a dedicated unit to promote savings countries with a dedicated unit to promote SME access countries with a dedicated unit to promote rural access countries offering low fee accounts countries encouraging use of accounts for government transfers Source: Martínez Pería 2011. Note: Panel a shows governments’ de jure commitment to �nancial inclusion. Panel b shows governments’ de facto commitment. 90 FINANCIAL INCLUSION: WHERE DOES LAC STAND FIGURE 5.9 Governments’ commitment to �nancial inclusion: LAC7 versus other LAC comparators a. De jure commitment 100 90 86 80 70 60 percent 50 50 50 43 43 40 33 29 30 25 25 20 17 10 0 0 0 Central America (includes the LAC-7 South America Dominican Republic) countries with strategy document countries with a mandate to promote savings countries with a mandate to promote SME access countries with a mandate to promote rural access b. De facto commitment 100 90 80 71 70 60 57 percent 50 50 40 33 29 29 29 30 25 25 20 10 0 0 0 0 0 0 0 Central America (includes the LAC-7 South America Dominican Republic) countries with a dedicated unit to promote savings countries with a dedicated unit to promote SME access countries with a dedicated unit to promote rural access countries offering low-fee accounts countries encouraging use of accounts for government transfers Source: Martínez Pería 2011. FINANCIAL INCLUSION: WHERE DOES LAC STAND 91 FIGURE 5.10 The adoption of correspondent banking and mobile branches 100 100 100 100 86 86 86 86 80 80 71 60 60 57 percent 43 40 29 20 0 0 Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (7) advanced economies (7) countries that allow correspondent banking countries that allow mobile branches Source: Martínez Pería 2011. FIGURE 5.11 The adoption of correspondent banking and mobile branches: LAC7 versus rest of LAC 100 90 86 80 71 70 60 percent 50 50 40 33 33 30 25 20 10 0 Central America (includes LAC-7 South America the Dominican Republic) countries that allow correspondent banking countries that allow mobile branches Source: Martínez Pería 2011. 92 FINANCIAL INCLUSION: WHERE DOES LAC STAND FIGURE 5.12 Index of credit information and legal rights 9 8 8 7 7 7 6 6 6 6 6 5 5 5 5 index 4 4 4 4 4 3 2 1 0 Asia (7) China Eastern G-7 (7) India LAC-7 (7) other Europe (7) advanced economies (7) median depth of credit information median strength of legal rights Source: Martínez Pería 2011. legal rights reforms should be a priority for dummies for LAC-7, for other LAC coun- LAC-7 governments. tries, and for other noncomparator countries, controlling for population density and GDP per capita. Table 5.1 shows that LAC-7 coun- tries appear to have more deposits and loans Notes than the group of comparator countries and 1. This chapter draws heavily on the paper are similar in terms of branches and ATMs. “Financial Inclusion in Latin America� by The F-tests at the bottom of the table show María Soledad Martínez Pería (2011), which that, after controlling for differences in popu- is part of the Edited Volume that accompa- lation density and income, there are no sig- nies this LAC flagship report. ni�cant differences in the indicators of �nan- 2. For theoretical studies, see Aghion and Bolton cial inclusion between LAC-7 and other LAC (1997), Banerjee and Newman (1993), and countries. Galor and Zeira (1993). For empirical work, 5. It is important to note, however, that data on see, among others, Banerjee et al. (2009), the number of loans are available for a small Burgess and Pande (2005), Dupas and Robinson number of countries—certainly fewer than (2009), and Karlan and Zinman (2010). the number of countries for which data were 3. This chapter uses �rm-level surveys conducted available on deposits, branches, and ATMs. by the World Bank across developing coun- 6. The �rms’ answers obviously reflect �rms’ tries and household surveys carried out by perceptions, so there could be biases if entre- the Corporación Andina de Fomento (CAF) preneurs from certain regions are inherently in the largest Latin American cities. more pessimistic. Nonetheless, considering 4. Differences in branches, ATMs, deposits, and this measure is useful, since it is the only vari- loans across country groupings are explored able that, to some extent, takes into account by regressing these variables against separate the cost of �nancing. FINANCIAL INCLUSION: WHERE DOES LAC STAND 93 7. Some speci�c government initiatives adopted vate credit bureau (or both): (1) Both positive in the region include, among others, the devel- credit information (for example, outstanding opment of correspondent banking in Brazil, loan amounts and pattern of on-time repay- government bank-sponsored micro�nance ments) and negative information (for example, programs in Brazil and Chile, and enhance- late payments, number and amount of defaults ments in conditional cash transfer programs and bankruptcies) are distributed. (2) Data on in Argentina, Brazil, and Colombia. See both �rms and individuals are distributed. (3) boxes 1 through 5 in Martínez Pería (2011) Data from retailers and utility companies as for further information. well as �nancial institutions are distributed. 8. Correspondent banking arrangements are (4) More than two years of historical data are partnerships between banks and nonbanks distributed. Credit registries and bureaus that with a signi�cant network of outlets, such as erase data on defaults as soon as they are repaid convenience stores, post of�ces, drugstores, obtain a score of 0 for this indicator. (5) Data and supermarkets to distribute �nancial on loan amounts below 1 percent of income per services. Mobile branches are of�ces where capita are distributed. Note that a credit regis- banking business is conducted, and they are try or bureau must have a minimum coverage moved to one or more predetermined loca- of 1 percent of the adult population to score a tions according to a predetermined schedule. 1 on this indicator. (6) By law, borrowers have 9. A score of 1 is assigned for each of the following the right to access their data in the largest credit six features of the public credit registry or pri- registry or bureau in the economy. The Banking Gap 6 C hapter 3 showed that banks in Latin is real enough: LAC banks lend less and America and the Caribbean (LAC) charge more than they should. (and LAC-7 banks even more so) • While LAC’s banking gap affects all lagged their benchmarks as regards the scale forms of credit, it affects mortgages the of intermediation and its cost (net interest most and consumer credit the least. margins). This apparent banking gap should • In the case of commercial credit, while the matter, however, only if it translates into lower gap primarily affects small and medium output growth or limits households’ welfare. enterprises (SMEs), the severity of its While a wealth of studies link private bank impact is not clear and needs to be investi- credit to output growth, poverty, and inequal- gated in more depth. ity, what holds across countries may not nec- • The largest fraction of the gap simply essarily be as relevant for a specific country or reflects LAC’s turbulent macrofinancial even region.1 The lack of domestic bank credit history, which puts the spotlight on mac- could be offset by a surplus of other forms of roeconomic policy and good prudential external funding, whether domestic or inter- oversight. national, or it could reflect a lack of demand • Limited demand for credit, reflecting for lendable funds rather than a lack of supply. LAC’s mediocre output growth in the The policy implications of the banking gap past, explains another substantial chunk clearly differ depending on its causes, such as of the gap; while this link between out- whether it reflects weaknesses in the financial put growth and credit goes in the opposite infrastructure and market environment or direction from the one generally empha- whether it is mainly the enduring legacy of a sized in the recent � nance literature, ulti- troubled past. Based on the available evidence, mately it also puts the spotlight on pro- this chapter presents a brief discussion of the ductivity-enhancing credit policies. underlying issues and policy implications. 2 • A significant share of the gap reflects Main findings are as follows: remaining weaknesses in the enabling environment that need to be addressed, • Although partly offset through alterna- particularly regarding the enforcement of tive channels of debt � nance, particularly contracts and the effectiveness of creditor cross-border channels, LAC’s banking gap rights. 95 96 THE BANKING GAP • While there are some indications that LAC but opened for commercial credit, and it has banking systems may also face ef�ciency opened even more for mortgage credit. issues due to insuf�cient competition, the These gaps could just reflect measure- evidence thus far is not conclusive. ment problems, if the lack of domestic bank finance were largely offset by foreign finance The rest of this chapter is structured as or by a surplus of domestic nonbank finance. follows. The first section breaks down the Unfortunately, because of data limitations, credit gap by type of credit and reviews the it is not straightforward to test this possi- evidence regarding the possible substitution bility.3 To gauge the role of foreign finance, of domestic bank credit by other forms of one can turn to the International Monetary lending. The second section sorts out sup- Fund (IMF) statistics on balance of payment ply from demand effects. The third section accounts, which provide data on gross foreign reviews the legacy of LAC’s turbulent past, debt liabilities that can be used as a point of and the final section concludes. reference, albeit private and public accounts cannot be separated. These data indicate that the fluctuations in domestic credit to the pri- Is the banking gap real? vate sector for four of the LAC-7 countries Table 6.1 presents a decomposition of the were to a good extent matched by opposite credit gap by type of credit for 2007 and (albeit much dampened) changes in gross debt 1996, using as a benchmark the median credit liabilities abroad (figure 6.1). While the corre- structure for a sample of middle-income lation is significantly negative (close to minus countries for which data are available. As 40 percent), the two series are clearly orders a proportion of its expected (benchmark) of magnitude apart. Thus, while there is evi- level, total private bank credit is about half dently some substitution, it is quite limited. of what it should be and has widened sub- At the same time, foreign private debt secu- stantially over the past decade. When dif- rities issued by LAC-7 corporations abroad ferent lending categories are considered, the (that is, nonbank credit to corporations) do gap is now greatest for mortgages and small- not outperform relative to the benchmark. In est (although still large) for consumer loans. fact, as shown in the summary benchmarks As already noted in chapter 3, the gap has table shown in chapter 3, they slightly under- narrowed considerably for consumer credit perform (table 3.2). Therefore, one can safely conclude that, while cross-border credit (from markets or intermediaries) may have substi- TABLE 6.1 LAC-7 credit gap by type of credit, 1996 and 2007 tuted for domestic bank credit at the margin, it clearly did not do so on average. Expected Actual Gap Gap/expected To understand the extent of domestic sub- Year: 1996 stitution of bank credit by domestic nonbank Credit to the private sector as a percentage of GDP intermediaries (such as factoring and leasing Commercial 24.7 19.4 5.2 21.1 companies, finance companies, credit coop- Mortgage 8.4 5.4 3.0 35.7 eratives, microfinance corporations, and sav- Consumer 8.8 3.4 5.4 61.4 ings and loans), one can turn to the evidence Total 41.8 28.2 13.6 32.5 presented in chapter 3 on factoring and lend- Year: 2007 ing by credit cooperatives. It suggests that Credit to the private sector as a percentage of GDP Commercial 22.9 14.5 8.4 36.7 such nonbank credit cannot be an important Mortgage 12.6 3.1 9.5 75.4 offset, both because it does not seem to be Consumer 11.2 6.5 4.7 42.0 unusually large relative to peer regions and Total 46.7 24.2 22.5 48.2 because it is too small, relative to bank credit. Source: de la Torre, Feyen, and Ize 2011. Moreover, the region slightly underperforms Note: This table estimates a breakdown by type of credit of the total credit gap identi�ed in chapter (rather than outperforms) the benchmark in 3 based on a comparison of predicted and actual values for four of the LAC-7 countries for which this information is available (Chile, Colombia, Mexico, and Peru). terms of the issue of domestic corporate paper THE BANKING GAP 97 FIGURE 6.1 Offshore versus onshore credit to the private sector, LAC-4 20 10 percent 0 –10 –20 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 private credit/GDP (yearly increases) debt liability flows/GDP Source: Authors’ calculations based on WDI data. Note: LAC-4 includes Brazil, Chile, Colombia, and Mexico. (table 3.2). Thus, the evidence again suggests TABLE 6.2 Real lending rates by type of firm, 2007 that the average (that is, structural) lack of percentage points domestic commercial credit by banks has not Eastern Other advanced been similarly offset by a surplus of domestic Asia Europe LAC-7 economies nonbank funding. Real interest rates to low-risk customers That being said, one can safely assume Small enterprises 2.0 –2.3 5.5 –1.5 that any remaining gap should affect mostly Medium enterprises 1.3 –2.7 3.9 –2.0 SMEs rather than large corporations, since Large enterprises 2.4 –2.5 0.8 –2.2 the latter are able to switch sources of Real interest rates to high-risk customers finance (whether at home or abroad) rather Small enterprises 6.3 –0.9 7.0 0.5 easily, depending on cost and availability. Medium enterprises 5.1 –0.9 6.0 0.0 Even there, however, the evidence as to Large enterprises 6.9 –1.4 1.9 –0.5 whether LAC’s SMEs have had a particularly Source: Beck 2011. hard time getting financing, as discussed in chapter 5, is not particularly conclusive. A of an ongoing process of catching up. LAC World Bank survey of interest rates charged banks seem to have expanded their credit to by banks to small and large firms does sug- SMEs quite rapidly and aggressively, partic- gest that LAC’s small firms may indeed pay ularly to those with supply links to the large more for their credit. However, the differ- corporations. 5 All in all, the only safe con- ence between LAC and other regions is not clusion at this point is that economists just that large (table 6.2).4 Conversely, one could do not know enough about the size of (and view the rapid growth in SME credit dur- fluctuations in) SME lending across banks ing the past decade as indirect evidence of and nonbanks to provide a clear answer a credit gap, since it suggests the existence about the severity of the impact of the credit 98 THE BANKING GAP gap on the productive sector. Clearly, this is Fovissste in Mexico).6 However, even if one an area for priority research. adds these funds to the total supply of mort- This section concludes with a brief look gages, the gap remains large.7 at the other two types of credit. As regards personal (consumer) credit, one could also view its rapid expansion as an indirect indi- Supply or demand? cation of a possible catch-up on an initial gap Sorting out supply from demand effects (figure 6.2). If so, however, the gap has nar- in the market for lendable funds is tricky, rowed considerably. Indeed, there is a wide- since the equilibrium reflects both sides of spread perception among policy makers in the market simultaneously. Nonetheless, the region that the banking system has been an argument based on differential elastici- concentrating too much on financing con- ties of supply and demand can shed some sumption rather than production. In the case light. This section considers prices first, then of mortgages, some substitution could come quantities. from the specialized institutions for housing If supply were much more elastic than finance that flourished in the past. However, demand, then one would expect real interest these institutions have, for the most part, rates (that is, net of expected devaluation and run into deep trouble and been discontinued expected inflation) to be close to world rates (see box 6.1). There is also, in some cases, an and, hence, rather low. If, instead, demand important supply of mortgages coming from were much more elastic than supply, then public provident funds (such as Infonavit and interest rates would reflect local demand FIGURE 6.2 Evolution of private credit by type of credit across LAC-7 countries 80 72 71 70 69 70 70 70 69 68 65 63 62 62 61 61 60 50 percent 40 30 17 20 22 23 22 21 10 10 9 10 12 14 11 10 20 10 19 20 21 21 20 20 19 17 18 17 16 15 16 17 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 mortgage personal commercial Source: Didier and Schmukler 2011a. Note: This �gure shows the average composition of private credit by type of credit for LAC-7 countries. LAC-7 comprises Chile, Colombia, Mexico, and Peru. THE BANKING GAP 99 BOX 6.1 A brief history of housing �nance in LAC Housing � nance grew fast in some LAC countries tions: savings and loans such as Sistema Nacional de in the 1990s; in others, the growth has come in Ahorro y Préstamo (SINAP) in Chile, Entidades the 2000s. Outstanding housing loans grew from de Ahorro y Préstamo in República Bolivariana 12 percent to nearly 20 percent of GDP in Chile, de Venezuela, Sociedades de Ahorro y Préstamo from 2 percent to over 10 percent of GDP in Mexico, (SAP) in Paraguay, and Corporativas de Ahorro y and from 1.5 percent to nearly 5 percent of GDP in Préstamo in Colombia. In all cases, to help attract Brazil. Still, except for Chile, housing � nance relative long-term capital, these institutions bene�ted from to GDP remains, on average, underdeveloped relative an exclusive right to offer indexed savings products. to where it should be for countries with similar char- For example, in Colombia, the Corporaciones de acteristics. LAC even experienced the rare phenome- Ahorro y Vivienda (CAVs) were created at the same non of regression in its development path: Colombia time as a speci�c index, the UPAC, and, by regula- has still not returned to the level prevailing before tion, CAVs denominated all their operations in this the 1998 crisis, and the Argentine market continues form of currency. But the index became problematic to dwindle, having now reached only a fourth of the as it migrated from being based on the consumer level that prevailed before the 2001–02 crisis. price index (CPI) to being based on a nominal inter- The single most important reason for this under- est rate. Moreover, prosperity based on regulatory development is macroeconomic instability, to which privileges proved once again to be fragile. SINAP long-term � nance is highly sensitive, as evidenced collapsed in the 1970s, 10 years after the creation of by the disruptions of supply following episodes of the system, when the Chilean government opened the deep turmoil. As a result, one feature that is strik- indexation option to all institutions. In Colombia, ingly common to many LAC countries is the per- the CAVs collapsed following the 1998 crisis, which vasiveness of indexation, which is intended to pro- was triggered by a dramatic increase in defaults that tect long-term investors against the consequences resulted from retroactive changes in the indexation of unexpected infl ation (see chapter 8 for more on system (back to the CPI-based index). The SAP sec- indexation). tor also collapsed in Paraguay in the wake of the late Historically, the mobilization of savings for hous- 1990s turmoil, again because of asset-liability mis- ing finance was done through specialized institu- matches linked to indexation. conditions and, hence, probably would in LAC are not particularly high. Moreover, exceed world rates. In fact, real lending rates recent work on emerging sovereign bond in LAC have exceeded U.S. rates by close to rates (Broner, Lorenzoni, and Schmukler 800 basis points on average over the past 2010) shows that in normal times the region decade (figure 6.3). faces a fairly elastic supply of foreign funds. The first question to examine, then, is Except at the longer end of the maturity whether the high borrowing rates mainly range in times of world market turbulence, reflect a banking intermediation problem or a bond rates are basically determined by the high cost of bank funding. Real deposit rates, world appetite for risk, with LAC behaving while also higher in LAC, have exceeded U.S. like other regions. The gradual shrinking of rates by only 100 to 200 basis points over the country premiums and their increased depen- past decade. Moreover, the deposit rate dif- dence on global fluctuations in risk appetite ferential has always been below the country (rather than idiosyncratic factors) tell a simi- risk differential, as measured by the Emerg- lar story. ing Markets Bond Index (EMBI) premiums Hence, the bulk of the differential in fund- (figure 6.3). This suggests that funding rates ing costs to borrowers reflects much higher 100 THE BANKING GAP FIGURE 6.3 Real lending rate, real deposit rate, and EMBI: differentials between LAC6 and the United States 15 10 5–year moving average (%) 5 0 –5 –10 –15 –20 p. 4 n 5 ar 86 c. 87 p. 7 n 8 ar 89 c. 0 p 0 n. 1 ar 2 c 93 p. 93 n 4 ar 95 c 6 p. 6 n 7 ar 8 c. 9 p. 9 n 0 ar 01 c 02 p. 2 n 3 ar 4 c 5 p. 5 n 6 ar 07 c. 8 p. 8 n. 9 10 Se 198 Ju 198 Se 198 Ju 198 De 199 Se 199 Ju . 199 M 199 Ju 199 De . 199 Se . 199 Ju 199 M . 199 De 199 Se 199 Ju 200 Se . 200 Ju 200 M . 200 De 200 Se . 200 Ju 200 De . 200 Se 200 Ju 200 M . 19 De . 19 M . 19 De . 19 Se . 19 M . 19 M . 20 De . 20 M . 20 20 c. . . . De deposit rate, LAC–United States lending rate, LAC–United States EMBI spread Source: Authors’ calculations based on local sources. bank interest margins, which is consistent However, the high overheads could also with the benchmarking evidence presented reflect a problem of insufficient scale, which in chapter 3. The obvious question, therefore, itself may reflect high costs of doing bank- is: what is behind these fatter margins? Gelos ing business due to a weak informational and (2009) breaks down bank intermediation enforcement environment. Given the limited spreads in LAC and concludes that higher intermediation (the bank credit gap relative to overhead accounts for about two-thirds of the benchmark), LAC banks may have a harder explained difference between LAC and other time spreading the fixed costs. Using indi- developing countries, with higher deposit vidual bank data, Gelos (2009) indeed finds rates (which translate into higher spreads) that bank size is a key determinant of bank and higher reserve requirements account- spreads. A similar result is obtained here ing for most of the rest (figure 6.4). Gelos using countrywide data. Including the ratio concludes that the higher overheads must of private credit to GDP (a proxy for scale) reflect inefficiencies due to lack of competi- as an additional control in the workhorse tion. However, Gelos fails to detect support- benchmark regressions presented in chapters ing evidence based on either the H-statistic 2 and 3 explains about two-thirds of the cur- (Panzar and Rosse 1987) or bank concentra- rent excess margin (table 6.3). Hence, the evi- tion ratios. A more recent study (Anzoategui, dence suggests, perhaps not too surprisingly, Martínez Pería, and Rocha 2010) reaches a that the high margins and the limited scale similar conclusion based on the Panzar-Rosse of intermediation are largely mirror images H-statistic as well as the Lerner index.8 In of each other. fact, LAC outperforms (rather than underper- This, in turn, suggests exploring in more forms) as regards either of these two indexes. depth LAC’s underperformance as regards Clearly, this is also an area that requires more the size of banking intermediation. Add- research. ing a basic set of enabling environment THE BANKING GAP 101 FIGURE 6.4 Net interest margins: Contribution of different factors in explaining differences between Latin America’s average and that for developing countries 11 equity and bank assets 10 bank profit taxes 9 ICRG score (country risk) 8 average bank size 7 GDP growth contribution 6 property rights 5 deposit rate 4 reserve requirements 3 overhead costs 2 total explained difference 1 total difference –0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 percentage points Source: Gelos 2009. indicators—contract enforcement costs, TABLE 6.3 LAC bank net interest margins, bank overheads, and creditor rights, property rights, and credit private credit information—to the basic benchmark regres- Net Net sions for private bank credit shows that interest interest some of them (enforcement costs and credi- margins margins Overheads Overheads tor rights) have a significant impact. Since Private credit (% of GDP) –0.0261*** –0.0236*** –0.0247*** –0.0229*** LAC significantly underperforms on both of (–8.280) (–6.314) (–9.833) (–8.390) these indicators, the two variables together Contract enforcement –0.247* –0.072 explain about 2.6 percentage points of GDP index (–2.224) (–0.724) of the credit gap (table 6.4). While this con- Legal rights index 0.0538 0.0653 (–0.976) (–1.395) stitutes only a modest fraction (about 13 per- Credit information index –0.0161 –0.0153 cent) of the total gap, measurement noise is (–0.214) (–0.276) likely to bias downward this result. Its share Property rights index 0.00697 –0.00324 of the total explained component of the (–0.883) (–0.467) gap (nearly one-fourth) probably provides a Constant 4.669 15.90** 5.172 17.92*** more accurate sense of the magnitude of its (1.087) (2.426) (1.519) (2.956) importance. Observations 1,280 459 1,280 459 Considering quantities now instead of R-squared 0.36 0.49 0.35 0.48 prices: the low volume of commercial credit Source: de la Torre, Feyen, and Ize 2011. Note: The contract enforcement index is the principal component of the following indicators from could be reflecting a lack of bankable proj- the World Bank’s Doing Business: contract enforcement costs, number of days to enforce a contract ects and, hence, a low demand for credit. (in logs), and number of procedures to enforce a contract. The legal rights index and the credit information index also are from Doing Business. The property rights index is from the Heritage LAC’s lackluster growth could, in turn, be Foundation. Robust t-statistics are shown in parentheses. a reflection of lack of investment rather than Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent. 102 THE BANKING GAP TABLE 6.4 LAC-7 Private credit gap: A decomposition by source going in the same direction, but admittedly also quite impressionistic, is that the recov- Dependent variable: Bank private credit (% of GDP) ery of credit in LAC has been mostly in con- sumer lending rather than in commercial (1) (2) (3) (4) loans, which may suggest that, as a whole Enforcement contract index –4.047* –5.318** (but possibly with the exception of credit (–1.864) (–2.358) Legal rights index 1.662 1.671 to SMEs, as discussed above), firms are not (–1.547) (–1.385) starved for credit (figure 6.2). Finally, while Credit information index –0.526 LAC banks’ high liquidity could just reflect (–0.379) caution, it could also be consistent with lim- Property rights index 0.235 0.069 ited demand for credit. (–1.301) (–0.335) Annualized average sample GDP 7.450*** 5.59** A more direct test of possible demand growth (2.935) (–2.433) effects is to check the extent to which out- Credit crash dummy (% of period) –86.92*** –77.69** put growth has affected finance. In the (–2.816) (–2.042) wake of the finance and growth literature, Workhorse controls Yes Yes Yes Yes it has been customary to think in terms Explained credit gap based on LAC-7 median values: of current finance affecting future output Contract enforcement index 1.51 1.98 Legal rights index 0.64 0.65 growth. Here, the opposite test is carried Credit information index 0.47 out: has past output growth affected current Property rights index 1.15 0.34 finance? Should this be the case, it could Annualized average sample GDP mean that another causality link—going growth 3.70 2.80 from low output growth to low investment, Credit crash dummy (% of time) 7.02 6.27 Total explained gap 3.77 3.70 7.02 12.04 low demand for lendable funds, and hence Gap 20.9 20.1 18.9 15.7 a low stock of commercial credit—may also Percent of total explained gap 18 18 37 77 be at play. Indeed, this is the case. LAC’s Source: de la Torre, Feyen, and Ize 2011. underperforming output growth over the Note: The contract enforcement index is the principal component of the following indicators from past 30 years appears to explain a sizable Doing Business: contract enforcement costs, number of days to enforce a contract (in logs), and number of procedures to enforce a contract. The legal rights index and the credit information index part of its current banking underperfor- are from Doing Business. The property rights index is from the Heritage Foundation. Robust t-statistics mance (table 6.4). are shown in parentheses. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent. Are these the ghosts of a lack of savings. And lack of investment could, turbulent past? in turn, reflect low productivity rather than a LAC’s macrofinancial history of the past high cost of funds. Indeed, a sizable literature 30 years was turbulent, to say the least. emphasizes LAC’s structural productivity A quick comparison with other regions and growth bottlenecks, which derive from is sufficient to prove the point. LAC was institutional weaknesses as well as overval- the region where crises were both the most ued real exchange rates.9 frequent and the most encompassing, fea- However, directly ascertaining whether turing the full range and mix of currency, investment is driving savings or whether banking, and debt crises (table 6.5). A savings are driving investment is a peril- bird’s-eye view of events can be obtained ous matter, given the obvious endogene- by contrasting the dynamics of real interest ities. Should the supply of foreign savings rates in LAC with those of real bank credit be relatively elastic, this would, of course, since the late 1970s, based on medians for tilt the balance toward investment driving the LAC-6 countries (Argentina, Brazil, savings. As discussed above, and with the Chile, Colombia, Mexico, and Peru) and proper caveats, this seems to be consistent comparing real interest rates in LAC with with the relatively high elasticity of demand those in the United States (figures 6.3 and for LAC bonds. Another piece of evidence 6.5).10 There were three clear credit cycles: THE BANKING GAP 103 one during the early 1980s, one lasting a lasting reflection of the sharp collapses most of the 1990s, with an interruption of credit during the 1980s and 1990s. To in 1995 due to Mexico’s “tequila crisis,� test this notion, one needs to differentiate and one that is ongoing after a brief inter- between collapses due to genuine disin- ruption due to the global financial cri- termediation (that is, banking crises) and sis. The first cycle started with a period of easy money and low real U.S. rates. TABLE 6.5 Number of crises by type, 1970–2007 It ended brutally in 1982 with U.S. inter- est rates rising sharply in the wake of Any External type Federal Reserve Chairman Paul Volcker’s debt Domestic Banking Currency of stabilization efforts and LAC rates going Country or region  crisis debt crisis crisis crisis crisis in the opposite direction, as the region’s Asia (5) 5 0 14 17 27 inflation rates went through the roof. The China 0 0 3 2 5 second cycle started with LAC’s mostly Eastern Europe (7) 6 2 13 15 28 failed exchange-rate-based stabilizations G-7 (7) 0 0 16 1 17 that resulted in high real interest rates, India 0 0 1 1 2 strong currency appreciations, and large Other advanced capital inf lows; that cycle ended with economies (7) 0 0 9 10 18 twin crises in most countries. The third LAC 47 13 53 72 149 cycle started in the early years of the Caribbean (2) 5 0 3 7 14 Central America millennium under the t win impetuses (+ Dominican of domestic macrostabilization and the Republic) (6) 11 2 13 14 33 strongly stimulative world environment LAC-7 (7) 16 7 21 36 63 resulting from China’s accelerated growth Offshore centers and large U.S. deficits. in LAC (3) 2 1 1 0 3 In view of this eventful background, a Other South key question is whether the comparatively America (4) 13 3 15 15 36 low levels of credit in the region today are Source: Broner et al. 2010. FIGURE 6.5 Real credit to the private sector and compounded real deposit rate index, LAC-6 400 350 300 index 100 in Jan. 1978 250 200 150 100 50 0 Se . 19 8 Ju . 19 8 M y 19 9 ay 80 De . 19 1 Oc c.19 2 Au t. 1 82 Ju . 1983 ne 84 Fe r. 19 5 De . 19 6 Oc . 19 7 Au t. 1987 Ju . 1988 Ap e 19 9 Fe . 19 0 b. 91 Oc . 19 2 Au t. 1992 Ju . 1993 Ap 19 4 Fe r. 19 5 De . 19 6 Oc . 19 7 Au t. 1997 Ju . 19 8 ne 99 Fe . 20 0 De . 20 1 Oc . 20 2 Au t. 2 02 Ju . 2003 ne 04 Fe r. 2005 De . 20 6 Oc . 20 7 Au t. 20 7 Ju g. 2008 20 9 10 v 7 p 7 l 7 ar 8 8 8 b 8 c 8 n 8 r 9 c 9 ne 9 9 b 9 c 9 g 9 r 0 b 0 c 0 b 0 c 0 0 ne 0 No . 19 M 19 g 9 Ap 19 De 19 Ap 20 g 0 Ap 20 n g g Ja LAC real credit index rate Source: Author’s calculations based on WDI data. 104 THE BANKING GAP collapses due to debt monetization (that is, partly offset, for the countries that allowed inflationary outbursts resulting in strongly it, by financial dollarization.12 negative real interest rates). While the for- Remarkably, the credit crash variable mer are associated with a clear worsening of also helps explain banks’ high interest mar- both banks’ and borrowers’ balance sheets, gins as well as their comfortable financial the latter are more likely to be associated soundness indicators (profitability, capi- with an improvement, so they could leave tal, and liquidity) (table 6.6). This suggests different imprints. At the same time, how- that banks that underwent crises were able ever, debt monetizations are likely to have to raise their margins (thereby raising their undermined the credibility of local cur- profitability), reflecting a forward reassess- rencies, thereby boosting domestic finan- ment of risks as well as perhaps a need to cial dollarization. Hence, unless countries recoup the losses incurred during the crisis. allowed dollarization to take hold—despite At the same time, they became more prudent its drawbacks—one would also expect a in managing risk, which led to less lending lasting impact on banking systems’ capac- and higher prudential buffers. While this is ity to intermediate. not too surprising, it is rather remarkable To test for these effects, a worldwide that these effects continue to be evident a credit crash variable is constructed that decade or two after the crises.13 reflects mild, strong, and severe annual drops in the ratio of private credit to GDP; the variable is included in the basic bench- Where is LAC now? mark regressions of private bank credit, Altogether, the evidence suggests that the together with a dummy for countries that domestic banking gap, although partly off- underwent episodes of extreme inflation set through alternative channels of debt during the past 30 years. To test for induced finance, particularly cross-border channels, dollarization effects, a deposit dollarization is nonetheless real enough. LAC banks lend variable is added as well as a variable repre- less and charge more than they should. How- senting the interaction of inflation and dol- ever, while there are good reasons to think larization. The credit crash variable is indeed that, in the case of commercial credit, this very significant, explaining as much as a gap affects SMEs more than it does the large third of the current credit gap in LAC (table corporations, the size of the impact—even 6.6). In contrast, the hyperinflation dummy on SMEs—is not sufficiently clear, given the is not significant. However, inflation and its currently available information. Obviously, interaction with financial dollarization are more research is needed; of particular value also jointly significant (table 6.7). would be an analysis of credit information Hence, the evidence appears to lead to that provides more insight into lending to the following joint set of conclusions: (a) the marginal borrowers. More research is also banking crises of the past have taken a very needed to ascertain the possible impact of the significant toll on LAC’s financial interme- lack of credit on firms’ leverage, activity, and diation, and the region is still paying for the investment, based on available enterprise-level sins of its abrupt cycles;11 (b) in contrast, the financial accounts data. Similarly, in the case monetizations induced by hyperinflation of mortgages, the gap also looks real. How- have not left a significant imprint; (c) how- ever, not enough is known about the precise ever, inflation has had a significant negative extent to which other forms of housing finance impact, not because it weakened balance (including from public provident funds) may sheets, but because it made financial con- be offsetting the lack of bank credit. tracting more difficult, particularly at the The largest fraction of the banking gap longer time horizons required for housing simply reflects LAC’s turbulent history. Even finance; and (d) the latter effect was at least though much time has passed, LAC has not THE BANKING GAP 105 yet fully recovered from the repeated credit TABLE 6.6 LAC banks’ interest margins, financial soundness, enabling crashes of its past. The main policy lesson environment indicators, and credit history—growth and crashes here is that financial sustainability is the Liquid name of the game. The long-run costs of Net Capital as assets as financial crashes are too large to be taken interest Return on % of total % of total Regulatory lightly. This puts the spotlight squarely margin investment assets assets capital on macroprudential policy and good sys- Enforcement –0.155 0.0438 0.353 2.366 0.247 temic prudential oversight. These issues are contract index (–0.713) (–0.42) (–0.857) (–0.726) (–1.326) addressed in chapters 10–13. Legal rights 0.156 0.051 –0.192 0.787 0.312*** Historically low demand for credit, as index (–1.318) (–0.931) (–0.903) (–0.496) (–3.131) proxied by LAC’s mediocre output growth Property rights 0.0246 0.00367 0.0659* –0.0129 0.0347** index (–1.192) (–0.387) (–1.769) (–0.0502) (–2.481) of the past, explains another substantial Annualized –0.0971 0.00887 –0.0202 0.156 –1.054*** chunk of the gap. To the extent that output average sample (–0.471) (–0.108) (–0.0642) (–0.0504) (–8.303) growth has to do with other (nonfinancial) GDP growth policies, such as macroeconomic policy or Credit crash 20.83** 5.068** 17.52* 13.39 27.90*** supply-side structural policies to enhance dummy (% of (–4.332) (–2.266) (–1.897) (–0.183) (–7.092) productivity and competitiveness, the pos- period) sible policy responses go beyond the finan- Constant 7.914 9.730** –19.96 187.9 12.26 cial sector. However, one can also argue (–0.958) (–2.504) (–1.170) (–1.408) (–1.611) Observations 88 88 98 98 78 that output growth (and, hence, ultimately Pseudo financial depth) could have been boosted R-squared 0.48 0.32 0.21 0.19 0.27 by financial policies focused on overcom- Source: de la Torre, Feyen, and Ize 2011. ing the limited marginal productivity of Note: The contract enforcement index is the principal component of the following indicators from capital by lowering the cost of finance; Doing Business: contract enforcement costs, number of days to enforce a contract (in logs), and number of procedures to enforce a contract. The legal rights index and the credit information that is, policy might have increased the index are from Doing Business. The property rights index is from the Heritage Foundation. Robust number of bankable projects by increasing t-statistics are shown in parentheses. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent. their profitability. A possible avenue for doing this is to stimulate demand for lon- ger maturity loans by SMEs through pub- lic guarantees, an issue that is addressed in chapters 8 and 9. TABLE 6.7 LAC private credit, financial dollarization, and inflation, A significant share of the banking gap 2005–08 also has to do with remaining weaknesses Dependent variable: Average private credit to GDP in the enabling environment. Much progress (1) (2) (3) (4) (5) has been made in resolving informational Dollarization – (period –17.73* 49.93* frictions. Indeed, LAC is ahead of many mean) (–1.770) (–1.851) emerging markets in the development of Dollarization – (last) 55.77** 62.59* credit bureaus, for example. But LAC still (–2.002) (–1.774) has a long way to go in addressing contrac- Log period inflation – –6.38*** –15.16*** –15.27*** –17.81*** tual frictions, particularly as regards the (period mean) (–2.952) (–3.572) (–3.690) (–2.988) enforcement of contracts and the preserva- Dollarization (mean)* 23.23** tion of creditor rights. While there are some Log Inflation (mean) (–2.249) Dollarization (last)* 24.84** 29.45* indications that LAC banking systems may Log Inflation (mean) (–2.261) (–1.893) also face efficiency issues associated with Constant 176.6** 167.2*** 118.4 103.1 133.6 insufficient competition, the available evi- (–2.321) (–2.653) (–1.557) (–1.347) (–1.234) dence is not conclusive. Should lack of com- Observations 128 162 128 128 86 petitiveness be confirmed through further R-squared 0.68 0.73 0.72 0.72 0.73 research, a policy agenda to address it would Source: de la Torre, Feyen, and Ize 2011. need to be developed. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent. 106 THE BANKING GAP Notes volume of new loans in 2010. They are also present in Brazil, Colombia, and República 1. On links between private credit and growth, see, Bolivariana de Venezuela. for example, Beck, Levine, and Loayza (2000), 7. For Mexico, the supply of mortgages as a pro- and King and Levine (1993). A recent survey of portion of GDP increases from 3.0 percent to the literature can be found in Arizala, Cavallo, 5.3 percent in 2009 when credit by Infonavit, and Galindo (2009), who analyze the impact Fovissste, and Sociedad Hipotecaria Federal of �nancial development on industry-level total are considered, which is still signi�cantly factor productivity. On links between �nan- below the benchmark level of 12.6 percent cial depth, poverty, and equal opportunity, see for 2007 indicated in table 6.1. Beck, Demirgüç-Kunt, and Levine (2007) and 8. The Panzar-Rosse H-statistic contrasts the Rajan and Zingales (2003). elasticity of a �rm’s revenue with that of its 2. This chapter draws heavily on the papers input costs (under perfect competition, an “Benchmarking Financial Development� by increase in input prices should lead to a one- Augusto de la Torre, Erik Feyen, and Alain for-one increase in output prices and, hence, Ize (2011) and “Financial Globalization: revenue). The Lerner index calculates the Some Basic Indicators for Latin America and disparity between prices and marginal costs the Caribbean� by Tatiana Didier and Sergio (that is, it is a measure of the markup). Schmukler (2011b), which are part of the 9. See for example McMillan and Rodrik (2011) forthcoming Edited Volume that will accom- for a discussion emphasizing the low growth pany this LAC flagship report. of output and employment in LAC’s higher- 3. BIS reports the data only on syndicated loans productivity sectors. reported in chapter 4. 10. The compounded real (deposit) interest rate 4. Given the very limited coverage of the survey, is included in the �gure because it provides this evidence is mostly indicative. some indication of “autonomous� changes 5. Beck, Demirgüç-Kunt, and Martínez Pería in credit that are simply driven by the com- (2011) document some of the drivers under- pounding of interest rates. Thus, the differ- lying the rapidly expanding capacity of large ence between the real credit and real interest banks to lend to SMEs based on public infor- rate lines is a measure of the truly exogenous mation rather than relationship lending. De la component of the credit cycle. Torre, Martínez Pería, and Schmukler (2010) 11. Interestingly, when adding a simple credit vol- �nd that the involvement of LAC banks with atility variable (the year-to-year variance of SMEs appears to be increasing, particularly for private to GDP credit) as an additional control the SMEs connected to the large corporations. in the benchmark regressions of credit, it is Based on a background econometric study for not statistically signi�cant. Hence, it is credit the Financial Sector Assessment Program for crashes—but not volatility per se—that leave Chile, Didier (2011b) con�rms that the “con- a substantial and lasting imprint on �nancial nected SMEs� get better treatment in terms of development. access to bank credit. 12. These links between �nancial depth, infla- 6. Provident funds for housing, based on man- tion, and dollarization were �rst explored by datory savings, are particularly notable de Nicolo, Honohan, and Ize (2005). in Mexico. According to the Asociación 13. Evidence showing that LAC banks currently Hipotecaria Mexicana, they accounted for enjoy substantial prudential buffers (capital, almost 85 percent of the number of new hous- liquidity, and pro�tability) is presented in ing loans and more than 70 percent of the chapter 10. The Equity Gap 7 A s argued in chapter 3, domestic equity Contributing factors to the low domestic markets across Latin America and trading of the smaller � rms may include the Caribbean (LAC) remain illiq- the negative spillover effects of the uid and highly concentrated. This relative offshoring of the larger stocks; the domi- underdevelopment stands in contrast to the nance of buy-and-hold pension funds rela- significant number of capital market reforms tive to more active institutional traders introduced over the last two decades and the such as mutual funds; and weaknesses in improved macroeconomic stance (figure 7.1), corporate governance (particularly with which were expected to help LAC equity respect to minority shareholder rights and markets develop toward their benchmark. protections) and the general enabling envi- This chapter attempts to shed light on the ronment (particularly as regards property possible causes for this underperformance. rights). It first reviews some of the most common • For reasons that remain to be fully eluci- explanations put forth by practitioners, dated, the region’s history of macroeco- policy makers and academics, including the nomic and � nancial turbulence may also effects of globalization, free float, market be a potential culprit of the low domestic concentration, institutional investor behav- liquidity. ior, and corporate governance. The chapter • Besides the obvious improvements in mac- then refers to additional possible causes. rostability, stock market infrastructure, Main highlights are as follows: and the general enabling environment (which should all help, but at the margin), • The offshoring of stock market turnover developing a proper policy agenda remains in LAC (the trading of equities abroad) thorny, particularly for the smaller coun- has been particularly large. tries and the smaller � rms, given the deci- • Such offshoring seems to significantly sive importance of scale (size of markets explain the domestic underperformance and of issues) and network effects in stock in the trading activity of the larger � rms; market development. however, it does not directly explain the low trading activity of the smaller � rms, The rest of this chapter is organized as fol- since the latter are not traded abroad. lows. The first five sections review the impact 107 108 THE EQUITY GAP FIGURE 7.1 Percentage of LACs that implemented capital market reforms, various years 100 100 94 91 92 88 80 63 64 62 percent 60 56 40 31 33 25 27 20 15 0 0 at cy s en dy ce nd law of m cre gen em to ng t ro g a ste n ts es s di en io ng us ss t p rin sy tra hm a ra c ry en lea ng er blis iso di em c rv sid a tra in est pe ar su ttl se reform before 1990 by 1995 by 2002 Source: de la Torre and Schmukler 2007. Note: This �gure shows the cumulative percentage of Latin American countries having implemented spec�c capital markets reforms at different points in time. of globalization, free float, market concen- the domestic market—the “liquidity migra- tration, institutional investors, and corporate tion� effect. Second, it can lead to a drop governance, respectively. The sixth section in the trading and liquidity of the stocks of reviews additional factors, and the final sec- the remaining domestic firms. This in turn tion concludes. can happen through two effects. The first effect (“negative spillovers�) is linked with the increase in cost per trade at home due Effects of globalization to fixed costs. The second effect (“domestic Calculation of effects begins by analyzing trade diversion�) follows from the fact that whether the lack of liquidity (that is, the low the internationalization of stock issuance turnovers) in the domestic equity markets and trading induces improvements in repu- can be due to the offshoring of stock mar- tation, disclosure standards, analyst cover- ket activity described in chapter 4 as part age, and the shareholder base, all of which of financial globalization. As the offshoring induce investors to shift their attention of stocks does not engender balance sheet from firms trading onshore to firms trading mismatches, it has no systemic vulnerabil- offshore.1 ity implications, even when the country that Figure 7.2, panel a shows the “total turn- engages in offshoring has a weak currency. over� for the stocks of (large) firms with However, the offshoring of stock issuance depository receipts (DRs) in the New York and trading can have adverse effects on the Stock Exchange—obtained as the sum of liquidity of the domestic stock market. It onshore and offshore trading divided by can do so through various channels (Levine their market capitalization reported for the and Schmukler 2007). First, it can shift the onshore market.2 It also shows the domestic trading of firms that issue abroad out of turnover for these same (large) firms and the THE EQUITY GAP 109 FIGURE 7.2 Onshore and offshore equity markets a. Average turnover ratio during the 2000s 2.0 1.8 1.6 1.6 1.5 1.4 1.3 1.3 1.2 1.2 turnover ratio 1.2 1.2 1.1 1.1 1.0 1.1 1.0 0.8 0.8 0.5 0.5 0.5 0.4 0.4 0.3 0.2 0.0 Asia China Eastern Europe G-7 India LAC-7 other advanced economies domestic turnover, �rms with DR programs total turnover, �rms with DR programs domestic turnover, all �rms b. Domestic and international value traded as a percentage of domestic market capitalization 2.5 2.0 0.3 0.1 0.2 turnover ratio 1.5 0.1 0.3 0.1 0.2 0.2 1.0 1.5 1.6 1.7 0.0 1.2 0.6 1.2 0.5 0.1 0.1 0.0 1.0 1.1 0.4 1.0 0.4 0.4 0.4 0.5 0.3 0.3 0.0 5 0 5 0 5 0 5 0 5 0 5 0 5 0 –0 –1 –0 –1 –0 –1 –0 –1 –0 –1 –0 –1 –0 –1 00 06 00 06 00 06 00 06 00 06 00 06 00 06 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Asia China Eastern Europe G-7 India LAC-7 other advanced domestic turnover international turnover economies c. Amount raised in foreign markets as a percentage of total amount raised through equity issues 60 49 50 47 44 % of total amount raised 40 32 32 32 30 28 24 20 19 20 16 16 12 10 6 0 Asia China Eastern Europe G-7 India LAC-7 other advanced economies 1991–99 2000–08 Source: Didier and Schmukler 2011b. Note: This �gure shows characterizes domestic and foreign equity markets. Panel a shows the average between 2000 and 2010 of the turnover ratios in domestic markets for �rms with DR programs as well as the total turnover ratios, which consider domestic and foreign trading activity. It also shows the aggregate turnover ratio in domestic markets for all listed �rms. Panel b shows, for �rms with DR programs, the average ratio of domestic and foreign value traded as a share of their market capitalization (that is, domestic and foreign turnover ratios between 2000 and 2010). Panel c shows the amount raised in equity market abroad relative to total amount raised through equity issues in domestic and foreign markets between 1991 and 2008. All DRs identi�ed in the DR Directory of the Bank of New York, with trading data reported in Bloomberg, are considered in this �gure. 110 THE EQUITY GAP turnover for all firms in the domestic mar- total domestic equity turnover suggests that ket. The striking result is that, once offshore the liquidity for the smaller firms (that do trading is taken into account, the turnover not trade abroad) is extremely low in LAC, of the large LAC firms nearly triples and is compared to other regions. Moreover, it has even higher than that of its Asian or Eastern remained broadly stable even as the total European peers. Indeed, for the large LAC turnover of the large firms has increased very firms, turnover abroad dominates turnover substantially (figure 7.2, panel b), which is at home, much more so than in other regions. broadly consistent with the negative spillover The effect is so large that for the large LAC view. firms with DR programs there does not seem A further check on the impact of offshor- to be an equity gap. Thus, offshoring does ing on the domestic turnover can be obtained appear to be largely responsible for the atypi- by first controlling the cross-country data on cally low domestic trading for these LAC domestic stock market turnover and foreign firms. Moreover, the increase in total turn- turnover for economic development (gross over for the large LAC firms has occurred domestic product [GDP] per capita) and coun- nearly entirely offshore (figure 7.2, panel b). try size (population) and then plotting the And the new issues of equity by LAC firms residuals of the two series against each other have been mostly done offshore, rather than and running a regression (figure 7.3). Interest- onshore (figure 7.2, panel c). ingly, the regression line is negative, suggest- The story for the smaller LAC firms is ing some substitution between onshore and quite different, as they typically do not have offshore trading. At the same time, however, access to foreign stock markets. The very low except for one country (República Bolivariana FIGURE 7.3 Average of residuals of domestic and foreign turnover, 2005–09 0.8 0.6 foreign turnover (residuals) 0.4 0.2 0 –0.2 y = –0.071x – 0.0043 –0.4 –1.5 –1 –0.5 0 0.5 1 1.5 2 domestic turnover (residuals) LAC-7 others Source: Authors’ calculations based on Didier and Schmukler 2011b. Note: This �gure shows the scatterplot of domestic and foreign equity turnover, de�ned as the average total value traded per year in domestic and foreign markets over domestic market capitalization for those �rms with DR programs. The 2005–09 average residuals of domestic and foreign turnover are reported. Residuals are obtained from ordinary least squares regressions of domestic or foreign turnover on GDP per capita and population. All DRs identi- �ed in the DR Directory of the Bank of New York, with trading data reported in Bloomberg, are considered in this �gure. THE EQUITY GAP 111 de Venezuela), all other LAC countries are by large shareholders (typically rich families), bunched up on the upward left of the scatter which is more likely to be the case in coun- plot, suggesting that LAC countries are atypi- tries with poor investor protection and high cal in that they trade more than expected off- concentration of wealth, secondary market shore, less than expected onshore. This sup- liquidity might dry up.3 In this case, the turn- ports the view that offshoring has displaced over figure is distorted, since trading captures domestic stock trading much more in LAC only the free-floating shares while market than in other regions. capitalization aggregates all the shares. A final check is reported in table 7.1. It This argument is assessed by adjusting the shows the results of a battery of regressions equity turnover ratio (value trade divided by of domestic stock market turnover using the market capitalization), that is, by considering same workhorse benchmarking model used in the denominator only the market capital- in chapters 2 and 3. While the limited data ization of the free-floating shares.4 Figure 7.4 availability is a strongly limiting factor, a shows that adjusting for free float increases remarkable result is that once one introduces LAC’s domestic turnover significantly. How- in the regressions the share of foreign trad- ever, it does the same in the case of other ing, the LAC dummy ceases to be signifi- regions. In fact, in Eastern Europe and cant. This result provides additional support China, the correction is even stronger than it to the view that much of the apparent LAC is in LAC. Hence, LAC’s low free float does equity gap can be explained by the region’s not appear to account for much, if any, of the extraordinary reliance on offshore trading. region’s abnormally low domestic stock mar- ket turnover. Free float Another argument often put forth as a pos- Market concentration sible reason for LAC’s equity gap is the low The evidence presented in chapter 3 also indi- level of freely tradable stocks (free float). That cated that access to equity markets in LAC is, a substantial component of market capital- remained limited and concentrated, with the ization in LAC corresponds to shares that are bulk of equity market activity depending closely held and hence not available for trading on a few issuers and issues. By limiting the in open exchanges. When firms are controlled options available to portfolio managers, the TABLE 7.1 Benchmarking model Dependent variable: Value traded over GDP (1) (2) (3) (4) (5) LAC-7 dummy –50.54*** –72.62* –26.13 –101.4** –75.71** (–2.756) (–1.766) (–0.559) (–2.501) (–2.458) Foreign value traded as % of total value traded –114.9* (–1.827) Foreign market capitalization as % of total market capitalization 411.5** (2.143) Amount raised by top 5 equity issues as % of total amount raised –31.85 (–0.604) Constant –40.89 194.5 54.84 –417.4 407.9 (–0.186) (0.181) (0.0535) (–0.402) (0.557) Workhorse controls Yes Yes Yes Yes Yes Number of observations 86 34 34 34 47 R-squared 0.613 0.435 0.509 0.532 0.444 Source: Authors’ calculations based on Didier and Schmukler 2011b. Note: Robust t-statistics are in shown parentheses. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent. 112 THE EQUITY GAP FIGURE 7.4 Trading activity in domestic equity markets during the 2000s 4.0 3.7 3.5 3.5 3.0 2.5 turnover ratio 2.2 2.0 1.9 2.0 1.8 1.5 1.2 1.3 1.1 1.2 1.1 1.0 1.0 0.5 0.5 0.2 0.0 ia a e 7 a 7 om ced in di p G- C- As ro Ch In ies LA on n Eu ec dva rn ra ste he Ea ot domestic turnover domestic turnover (free float ajusted) Source: Authors’ calculations based on Dahlquist et al. 2003 and Didier and Schmukler 2011a. Note: This �gure shows the trading activity in domestic equity markets. It shows the average turnover ratios, de�ned as the total value traded per year in domestic markets over total market capitalization, between 1990 and 2009. It shows the same data but adjusted by the Dalhquist (2003) free float indicator for each country for 2003. high concentration could limit trading, lead- to LAC’s equity gap, as pension funds do ing to the low turnover. not engage in active trading but instead As a check of whether this is the case, a mostly buy-and-hold. 5 As discussed in more similar exercise to that described above is depth in chapter 8, current regulations tend conducted for domestic and offshore turn- to reinforce the preference for “buy-and- over. That involves first running separate hold� investment strategies, which can be regressions of equity market turnover and detrimental to market liquidity.6 More gen- concentration ratios on GDP per capita and erally, institutional investors tend to invest population. The residuals for the two vari- in larger and more liquid firms, hence lim- ables are then plotted and displayed along the iting the supply of funds to smaller and less fitted regression line (figure 7.5). The results liquid ones.7 are unimpressive. The regression line is flat— To check the impact of the pension- indicating that market concentration has little fund biased institutional investor universe correlation with turnover—and LAC coun- in LAC capital markets on the domestic tries are distributed fairly evenly above and equity market turnover, the exercise again below the fitted line—suggesting that market controls the data for GDP per capita and concentration for the region as a whole is not population. The residuals are again plot- particularly atypical. In fact, market concen- ted, along with the fitter regression line tration appears to be a feature of many coun- (figure 7.6). Remarkably, the regression tries, particularly developing ones. line is flat as regards pension funds but clearly upward sloping as regards mutual funds and insurance companies. This may Institutional investors be viewed as supporting the evidence pre- The predominance of pension funds among sented in chapter 2, whereby the growth institutional investors could also contribute of pension f unds has a strong policy THE EQUITY GAP 113 FIGURE 7.5 Domestic turnover and concentration a. Share of the amount raised by the top �ve issues as a percentage of total amount raised 0.5 0.4 0.3 0.2 % of total (residuals) 0.1 y = –0.0007x + 0.0888 0.0 –0.1 –0.2 –0.3 –0.4 –0.5 0 50 100 150 200 250 domestic turnover (residuals) b. Value traded by top �ve companies as a percentage of total value traded 0.4 0.3 0.2 0.1 % of total (residuals) y = –0.0001x + 0.0192 0.0 –0.1 –0.2 –0.3 –0.4 –0.5 0 20 40 60 80 100 120 140 160 180 200 domestic turnover (residuals) LAC-7 others Source: Authors’ calculations based on Didier and Schmukler 2011b. Note: This �gure shows scatterplots of the 2005–09 average residuals of two measures of concentration in domestic equity markets against the average residuals of the domestic turnover ratio. Panel a shows the average amount raised per year by the top �ve issues as a share of total issues in domestic markets. Panel b shows the average share of value traded by the top �ve companies as share of the total value traded per year in domestic markets. The turnover ratio is de�ned as the total value traded per year in domestic markets over domestic market capitalization. Residuals are obtained from ordinary least squares regressions of the variables on GDP per capita and population. LAC countries are reported in darker colors. 114 THE EQUITY GAP FIGURE 7.6 Domestic turnover and institutional investors that, in contrast with other institutional investors such as mutual funds, pension funds do not contribute much to stock a. Pension funds 1.2 market liquidity because they mostly buy 1.0 and hold. In this interpretation, the fact 0.8 that most LAC countries are bunched up % of GDP (residuals) 0.6 under the regression line as regards mutual 0.4 funds but are more evenly distributed 0.2 y = 0.0002x – 0.0101 around the line as regards pension funds 0.0 would suggest that the low equity turnover –0.2 could have something to do with the pre- –0.4 dominance in the region of buy-and-hold –0.6 0 50 100 150 200 250 pension funds and the relative underdevel- domestic turnover (residuals) opment of mutual funds (presumably more b. Mutual funds active traders). 1.5 1.0 0.5 Corporate governance % of GDP (residuals) 0.0 Weak corporate governance practices –0.5 y = 0.0063x – 1.4292 are also a commonly stated argument –1.0 to explain the low development of stock –1.5 markets (see box 7.1). Figure 7.7, panel a –2.0 compares LAC to other regions as regards –2.5 the anti-self-dealing index and the antidi- 0 20 40 60 80 100 120 140 160 180 200 domestic turnover (residuals) rector rights index, two widely used cor- c. Insurance companies porate governance indicators. 8 To some 1.5 degree, LAC-7 lags behind as regards both 1.0 indicators. However, the region fares sig- 0.5 nificantly worse in the case of the anti- % of GDP (residuals) 0.0 self-dealing index, where the gap between –0.5 LAC-7 and other regions (except Eastern –1.0 y = 0.0063x – 1.4292 Europe) is large (0.39 compared to 0.55– 0.80). The plotting of the controlled residu- –1.5 als of these two governance indicators and –2.0 the domestic turnover (following the same –2.5 0 20 40 60 80 100 120 140 160 180 200 procedure as that described above) leads domestic turnover (residuals) to a similar conclusion (figures 7.7, panel LAC-7 others b and 7.7, panel c). The regression line for Source: Authors’ calculations based on Didier and Schmukler (2011a, 2011b). the anti-self-dealing indicator is clearly Note: This �gure shows scatterplots of the 2005–09 average residuals of institutional investors’ assets upward-shaped, which suggests that it is relative to GDP against the average residuals of the domestic turnover ratio. Panel a shows the assets of pension funds as a percentage of GDP. Panel b shows the assets of mutual funds as a percentage more closely connected with market devel- of GDP. Panel c shows the assets of insurance companies as a percentage of GDP. The turnover ratio is opment. At the same time, most LAC coun- de�ned as the total value traded per year in domestic markets over domestic market capitalization. Residuals are obtained from ordinary least squares regressions of the variables on GDP per capita tries are bunched up under the regression and population. LAC countries are reported in darker colors. line, which confirms LAC’s strong under- performance as regards this indicator and component, whereas that of other institu- suggests that LAC’s equity gap might have tional investors tends to go hand in hand something to do with it.9 Given the difficul- with economic and financial development. ties in measuring corporate governance and However, if one interprets the causality in the multiple dimensionality of this concept, the other direction, it could also suggest some caution is warranted, however.10 THE EQUITY GAP 115 BOX 7.1 Corporate governance and equity market development There is a large literature relating corporate gov- associated with higher operating performance and ernance practices and stock market development. higher Tobin’s Q. Joh (2003) concludes that � rms According to Shleifer and Vishny (1997a) and with higher control-ownership disparity exhibit Bhojraj and Sengupta (2003), good governance prac- lower pro�tability. tices increase con� dence among investors as they The positive links between good corporate gov- tend to reduce agency (information and enforce- ernance and stock market development are docu- ment) risks. Therefore, companies are likely to have mented in many papers. For example, La Porta et al. access to capital at lower costs and better condi- (1997) and Glaser, Johnson, and Shleifer (2001) tions, to increase the value and liquidity of their show that protection of minority shareholders is fun- shares, and to improve their operating performance damental to the development of a country’s capital and pro�tability. market. In addition, Klapper and Love (2004) show Ashbaugh-Skaife, Collins, and LaFond (2006), that good governance practices are more important for example, � nd that better corporate governance in countries with weak investor protection and inef- practices improve corporate credit ratings and reduce �cient enforcement. bond yields. De Carvalho and Pennacchi (2011) However, using evidence from the stock market argue, for the case of Brazil, that migration from internationalization process, Gozzi, Levine, and traditional markets to the Novo Mercado brings Schmukler (2010) argue that the causality might go positive abnormal returns to shareholders and an both ways and that better � rms go offshore, to better increase in the trading volume of shares. Klapper and corporate governance environments, not that better Love (2004) � nd that better corporate governance is governance necessarily increases � rm value. Other factors low domestic stock market turnover when introduced separately. Interestingly, however, Finally, LAC’s low growth, turbulent mac- they lose their significance when the LAC-7 rofinancial history, and remaining weak- dummy is added. This could suggest that, nesses in its enabling environment might while financial crashes and low growth affect also contribute to explaining its low domes- many other countries outside the LAC, they tic equity market turnover. Checking for have had special consequences in the case such effects is done by adding to the econo- of LAC, so much so that they have become metric workhorse model presented in chap- tightly interwoven with LAC specificities (the ters 2 and 3 proxy measures of economic LAC-7 dummy). Also, our econometric test prospects (average GDP growth for the past for enabling environment indicators suggests three decades) and macrofinancial turbu- that contract enforcement costs, property lence (credit crashes, as defined in previous rights, and information are also part of the chapters). The calculation also adds vari- story of the low turnover in LAC’s domes- ous indicators of the quality of the enabling tic stock markets (these variables retain sta- environment (contract enforcement, prop- tistical significance even when introduced erty rights, credit information). together with the—also significant—finan- The results (table 7.2) are tentative as they cial crashes variable). However, like the do not fully survive robustness tests, but they financial crashes and growth variables, the do hint in some specific directions while enabling environment indicators also lose sig- underscoring the need for more research. nificance once the LAC-7 dummy is added. In particular, financial crashes and low Again, this might suggest that the effects of a growth are significantly associated with the low growth, financial crashes, and enabling 116 THE EQUITY GAP FIGURE 7.7 Domestic turnover and corporate governance environment weaknesses are wrapped up tightly in the region’s history and are crucial a. Anti-self-dealing and antidirector indexes in making LAC what it is now as regards 1.0 1.00 financial development. 0.8 0.78 0.78 0.79 0.64 0.67 0.64 Conclusions average index 0.6 0.55 0.55 0.58 0.54 A first clear conclusion in this chapter is that 0.4 0.34 0.39 offshoring appears to account for much of 0.20 LAC’s domestic equity market turnover gap. 0.2 This gap probably does not matter for the 0.0 larger firms: whether their stock is traded Asia (5) China Eastern G-7 (7) India LAC-7 (7) other Europe (6) advanced in Mexico City or in New York is largely economies (7) immaterial and the fact that it is traded anti-self-dealing antidirector mainly in New York may actually produce b. Domestic turnover and ex-ante private control of self-dealing index additional benefits for these firms. But the 0.8 gap does matter for the smaller firms to the 0.6 extent that they cannot rely on international 0.4 markets and are thus constrained by the lack index (residuals) 0.2 of access to equity finance at home. Even if 0.0 these firms’ access to debt finance at home was adequate (which does not seem to be the –0.2 case, as discussed in chapter 6), that would –0.4 y = –0.0003x – 0.0051 not substitute for the lack of access to equity –0.6 finance, as the latter plays a unique role in –150 –100 –50 0 50 100 150 200 domestic turnover (residuals) long-term business expansion. c. Domestic turnover and antidirector rights index Two interrelated but clearly distinct ques- 3 tions in this regard are first: Why is the off- 2 shoring of equity turnover so large in the case of LAC? And second: Why has offshor- 1 ing seemingly had such a depressing impact index (residuals) 0 on domestic equity market liquidity? Levine and Schmukler (2007) shed light on the sec- –1 ond question, providing evidence that illus- –2 trates the channels through with the adverse –3 effect appears to work, as noted above. Yet y = 0.0035x – 0.0204 there are no solid answers as to why LAC has –4 an abnormally large amount of stock traded –150 –100 –50 0 50 100 150 200 domestic turnover (residuals) abroad relative to what is traded at home in LAC-7 others the first place, and why the equity markets for the remaining firms have lagged so much Source: Authors’ calculations based on Djankov et al. 2008. Note: This �gure characterizes the relation between corporate governance and liquidity in behind. The history of low economic growth domestic equity market. Panel a shows the anti-self-dealing and antidirector indexes. The anti- (to the extent it is associated with uninspiring self-dealing index intends to capture the strength of minority shareholder protection against practices where management or controlling shareholders use their power to divert corporate expected returns to investment) and, perhaps wealth to themselves. The antidirector index intends to capture the stance of corporate law more importantly, LAC’s history of financial toward shareholder protection. Higher levels of the indexes imply stronger shareholder protection. Panels b and c shows scatterplots of the residuals of two indexes of corporate governance against crashes may have something to do with it. the residuals of the domestic turnover ratio for 2003. Panel b shows the ex-ante private control of However, the empirical tests conducted above the self-dealing index. Panel c shows the antidirector rights index. The turnover ratio is de�ned as the total value traded per year in domestic markets over domestic market capitalization. Residuals do not directly address the reasons underlying are obtained from ordinary least squares regressions of the variables on GDP per capita and offshoring. Hence, the results speak more to population. the depressing effect of such factors on equity THE EQUITY GAP 117 TABLE 7.2 Domestic equity turnover and enabling environment indicators Dependent variable: Stock market turnover     (1) (2) (3) (4) (5) (6) LAC-7 dummy –35.83*** –27.02** –32.12** (–4.837) (–2.505) (–2.635) Credit crash dummy (% of period) –134.00* –109.20 –125.70*** –72.19 (–1.902) (–1.524) (–3.049) (–1.042) Annualized average sample GDP growth 6.315** 3.164 0.440 (2.424) (1.385) (0.225) Contract enforcement index –4.298* –1.477 (–1.970) (–0.495) Credit information index 4.452*** 2.091 (2.738) (0.976) Property rights index 0.592*** 0.510** (3.211) (2.153) Constant 306.9*** 359.3** 353.9*** 296.0*** 445.7*** 319.3** (4.010) (2.390) (2.786) (2.730) (4.343) (2.348) Workhorse controls Yes Yes Yes Yes Yes Yes Observations 107 107 86 86 103 84 Pseudo R-squared 0.46 0.44 0.49 0.54 0.47 0.55 Source: de la Torre, Feyen, and Ize 2011. Signi�cance level: * = 10 percent, ** = 5 percent, *** = 1 percent. market development overall than to the off- seem to differ that much from that of pension shoring per se.11 funds. A second, more tentative conclusion is As regards the possible impact of LAC’s that the preponderance of pension funds over weaknesses in corporate governance on the other institutional investors and the remain- equity market, one might take the view— ing weaknesses in corporate governance, con- considering the experience of Brazil’s Novo tract enforcement, and property rights have Mercado, whose development appears to also contributed to LAC’s domestic equity have been spurred by the tightening of gov- market turnover gap. As regards the role ernance norms—that such weaknesses are of played by institutional investors, to the extent first-order importance. Yet, much of the suc- that the (policy-driven) rapid development of cess of the Novo Mercado may have more to pension funds has retarded the growth of do with Brazil’s comparative size advantage other institutional investors such as mutual than with governance reforms. Indeed, an funds, one could conclude that such policy alternative for the smaller countries might choices might have had a cost in terms of be to follow a “lighter governance� path that equity market development. But much cau- is more suited to the smaller firms, while tion is needed in interpreting the evidence. accepting the trade-off of having a reduced First, the impact of pension funds on the scope for minority shareholders (who would development of the mutual fund industry is be less reluctant to own stock under lighter not clear. While pension funds may substitute governance arrangements). Such a light ver- mutual funds by giving investors an alterna- sion might be characterized by more benign tive savings channel, pension funds may also accounting and public disclosure standards, help mutual funds develop by investing part more private equity placements and over-the- of their portfolios in them. Moreover, as counter activity, less reliance on centralized shown in chapter 8, the asset management local exchanges, and concentrated (rather behavior of LAC’s mutual funds does not than atomized) stock ownership. However, 118 THE EQUITY GAP given the structural illiquidity and, hence, improvements in stock market infrastructure limited price revelation capacity of these mar- that can also help, including those aimed at kets, their ultimate economic benefits are reducing fragmentation in issuance and trad- unclear. ing, enhancing securities clearance and set- To overcome the constraints imposed by tlement arrangements, organizing securities small size of markets, many have recom- lending and borrowing facilities, improving mended the cross-border integration of LAC valuation methods, promoting contract stan- stock markets. Indeed, the past decades have dardization, and upgrading financial report- seen several attempts toward regional integra- ing. However, what is also evident is that tion of stock exchanges around the world.12 such reforms would at best only correct for Recently, Chile, Colombia, and Peru reached a modest proportion of the domestic equity an agreement of this sort, which focuses on turnover gap and may, rather, continue to integrating such functions as listing, order deepen the trend in favor of offshoring. routing, and execution. The agreement also Thus, the larger questions remain. What fosters efforts aimed at regulatory conver- is the proper level of governance standards gence, but it so far does not envisage common one should shoot for? Should the smaller clearing and settlement systems. Despite the countries simply “throw in the towel,� for- potential benefits of securities market inte- get about developing a local stock market, gration in terms of scale and network effects, and accept that equity funding is available these attempts have tended to fail (Lee 1999). mainly for their large resident corporations Many reasons have been given for this lack and mainly through the use of international of success, including legal and regulatory dif- stock markets? Or should they persevere for ferences across countries, adverse effects of the sake of their smaller firms? Is there any different national currencies in the absence significant advantage in pursuing regional of sufficiently developed currency deriva- stock market integration compared to simply tives markets, informational barriers across promoting global integration? The only thing markets (including differences in accounting one can take for certain is that LAC will need and disclosure standards), and larger than to look beyond the simplest conventional expected difficulties in integrating market wisdom: macrostability and compliance with infrastructures. However, as discussed in de international standards might help, but they la Torre, Gozzi, and Schmukler (2007a), even won’t suffice. if these obstacles were appropriately over- come, there remain some fundamental rea- sons that cast doubt on the proposition that Notes regional integration of stock exchanges would 1. Levine and Schmukler (2007), for example, be superior to the alternative of a deeper and �nd empirical evidence of a signi�cant nega- better integration with the developed stock tive effect of offshoring on domestic stock markets, for example, the New York Stock market liquidity, and their results are consis- Exchange, which are unrivalled in terms of tent with both views expressed above. Some the depth of their liquidity pools and quality theories, however, argue instead that offshor- of their contractual environment.13 ing may enhance integration and thereby In any event, ascertaining the policy path stimulate domestic trading and boost the for stock market development in LAC, espe- liquidity of domestic �rms. See, for example, Alexander, Eun, and Janakiramanan (1987); cially for the smaller countries, is fiendishly Domowitz, Glen, and Madhavan (1998); and difficult, much more than is commonly rec- Hargis (2000). ognized. Of course, there are enabling envi- 2. The domestic market capitalization of these ronment reforms (for example, in property �rms includes all the stocks issued at home rights and corporate governance frameworks) even if they are completely traded abroad (via that everyone agrees should help. There are in DRs). addition, even for the small countries, many 3. See Dahlquist et al. (2003). THE EQUITY GAP 119 4. Free float data used in this assessment are by choosing where to list. While this might provided by Dahlquist et al. (2003). While it have boosted the value of the antidirector is the only one available for a large sample of index, which measures the extent of legal countries, it is unfortunately a bit outdated (it protection, it might not have had an immedi- refers to 1997). ate effect on actual self-dealing practices. 5. Raddatz and Schmukler (2011) show that 10. In unreported results on two more corporate Chilean pension fund administrators (PFAs) governance indicators (ex ante private con- trade infrequently. On average, a PFA trades trol of self-dealing and public enforcement only 13 percent of its assets, and the monthly index), LAC-7 seems to be overperforming, changes in asset positions correspond to to the point of being even slightly ahead of just 4 percent of the initial total value of the G-7 countries. the assets. This result contrasts sharply with 11. Caution in interpreting the increase in off- the 88 percent mean turnover ratio found in shoring as caused by macroeconomic vola- Kacperczyk, Sialm, and Zheng (2008) for a tility is further warranted considering that sample of 2,543 actively managed U.S. equity Claessens, Klingebiel, and Schmukler (2006) mutual funds between 1984 and 2003. �nd that improvements in macroeconomic (as 6. See Gill, Packard, and Yermo (2004). well as institutional) fundamentals are associ- 7. See, for example, Kang and Stulz (1997); ated with a proportionally greater expansion Dahlquist and Robertsson (2001); Edison of turnover offshore than onshore. and Warnock (2004); Didier, Rigobon, and 12. This process has been particularly strong Schmukler (2010); and Didier (2011a), among among European countries (Claessens, Lee, many others. and Zechner 2003; Licht 1998; McAndrews 8. See Djankov et al. (2008). The anti-self- and Stefanadis 2002). dealing index intends to capture the strength 13. While it is true that regional �nancial inte- of minority shareholder protection against gration may reduce trading and issuance practices where management or controlling costs because of economies of scale, it seems shareholders use their power to divert cor- doubtful that such cost reductions would be porate wealth to themselves. The antidirector greater than those that could be achieved index, conversely, tries to capture the stance by global integration. Similarly, while it is of corporate law toward shareholder protec- true that neighboring investors may have tion. Higher levels of the index imply stronger informational advantages on regional �rms shareholder protection. Djankov et al. (2008) compared to more remote foreign investors, �nds the anti-self-dealing index to be a bet- it is not clear that such advantages would ter predictor of a variety of measures of stock be better exercised by trading in a regional market development across countries. market than in a global one. Likewise, the 9. In terms of individual countries, Argentina, conjecture that regional stock exchanges Mexico, and Uruguay have the weakest would facilitate access for medium-size corporate governance indicators. Brazil is enterprises needs to be reexamined, for these an interesting case, as its antidirector index �rms are segmented out of the international takes the maximum possible value while its and local stock markets mainly because of anti-self-dealing index is one of the lowest in the small size of their potential issues, not the region. This might be a result of the recent because of the size of the markets. The solu- developments in the Brazilian stock market tion, therefore, is arguably not with bigger (discussed in chapter 3) whereby �rms can markets, regional or global, but with bigger adhere to stricter corporate governance rules issue sizes. Going Long 8 I n the previous chapters, the Latin America in long-term � nancial contracts; in addi- and the Caribbean (LAC) region’s � nan- tion to sound macroeconomic policies, this cial development and main developmen- will require persistence in promoting the tal gaps were discussed from a broad cross- use of consumer price index (CPI)-indexed country perspective. This chapter narrows bonds for the long-maturity segments of the focus to explore an issue that is at the public debt. core of effective, sustainable �nancial devel- • LAC countries will also need to strengthen opment: the capability of “going long,� that contract rights and their enforcement, two is, of sustainably generating suf�cient long- areas where LAC lags. term �nancial contracts—long maturity and • As regards market discipline, LAC will long duration—for private agents and proj- need to promote standardization and ects.1 Chapter 6 identi�ed the lack of mort- benchmarking of funds (especially pen- gages as a central component of the banking sion fund management) in terms of asset gap in LAC. Chapter 7 discussed the possible valuations and returns, in a way that reasons underlying the liquidity gap in the focuses investors on long-term rather than equity market. This chapter digs deeper into short-term returns. the credit gap by reviewing why LAC has had • At the same time, governments will need such a hard time providing long-term �nance, to � nd the proper balance between regu- particularly mortgages. The equity gap is also lating the way investors invest (effectively examined again, but from a different and playing “big brother�), on the one hand, complementary angle. In particular, the ques- and letting them assume the risk, on the tions asked are: Why are investors so hesitant other. to become locked into less liquid equity secu- • LAC will also need to � nd the best mix of rities or longer-term debt contracts? To what channels and instruments to help spread extent is this problem speci�c to LAC, and risk and reduce participation costs, includ- what can be done about it? Main policy con- ing through securitization, covered bonds, clusions are as follows: mutualized second-tier � nance facilities, and a general deepening of markets for • LAC countries will need to consolidate the insurance and hedging products aimed at gains made in the use of the local currency reducing default risks over longer horizons 121 122 GOING LONG (such as rainfall-index-based insurance for exposed to rollover or price risk. On the sav- farmers or home-equity insurance). ings side, investing long term is central to • LAC policy makers will need to be realis- resolving the problems of retirement, edu- tic about the feasibility (and desirability) cation needs, health shocks, and premature of relying on liquid secondary markets to death—key facets of life-cycle consumption promote participation by aligning inves- smoothing. The capacity of investors to stay tors’ desire for an exit option with bor- put—that is, to hold risk over long horizons rowers’ long-term funding needs. Where and capture the risk premium—makes a domestic markets are small, reforms huge difference in accumulated returns over aimed at promoting secondary market a lifetime.2 Thus, a � nancial system’s capac- liquidity are likely to lead to only modest ity to spread long risk effectively is crucial improvements; and where market arrange- to a viable private pension, education, or ments succeed in deepening domestic health system. liquidity, wedges between private and sys- While the financial architecture in LAC temic interest can widen as private agents has evolved in a variety of ways to facili- increasingly free ride on liquidity and tate the spreading of long risk, barriers and relax their monitoring efforts. pitfalls remain. Overcoming them is key to • Hence, an alternative avenue for dealing a harmonious, sustainable financial devel- with collective action frictions is to man- opment. However, this is a challenge that date participation, as in private pension even the most developed financial systems funds, for example; when combined with have struggled to meet, as evidenced by the policies to promote the development of recent U.S. mortgage crisis. In fact, the sus- a vibrant annuities industry, as in Chile, tainable production and use of long-term this approach has been one of the bright- instruments—say, a 30-year � xed-rate mort- est and most promising in the region. gage or a lifelong annuity—are clear signs of Nonetheless, the design problems that a mature and socially useful �nancial system. stand in the way of developing the annui- In this sense, going long is a key point of ties market in many countries will need to intersection between � nancial development be removed. and �nancial stability. The two sides of the market for going long The rest of this chapter is organized as typically have a hard time meeting, because follows. The � rst section reviews the main long-term �nancing commitments face large issues faced in going long. The second section aggregate risk (inflation, growth, political discusses the options. The third section takes risk) as well as idiosyncratic risk (project stock of LAC’s current situation. The � nal risk on the one hand and agency risk— section concludes by briefly reviewing the asymmetric information and/or enforcement main policy challenges going forward. costs—on the other). The longer one com- mits, the more exposed one becomes to these risks. Where will prices be 20 years from What are the issues? now? Where will society be? Will the project The importance of long-term contracts for fail? What will the borrower do, and will he economic and social well-being is uncon- still be there? Hence, while borrowers would tested. On the borrowing side, limiting the like to limit the risk of long-term investments cost of issuing long-term debt is central to by � xing the terms of � nance, lenders may viable housing, infrastructure, and long- not be willing to assume the risk associated term corporate �nance. If long-term borrow- with going long and may thus prefer to lend at ing is too expensive, borrowers must limit short durations. The problem is complicated their funding—thereby limiting investment further by the fact that making fully rational and, in the end, economic growth—or they long-term investment decisions is beyond the must keep it short term—thereby becoming capacity of the average investor.3 The price at GOING LONG 123 which lenders are willing to assume the long company, which may then engage in risk may exceed the price that borrowers are maturity transformation and absorb the ready to pay. risks while providing downside risk insur- ance to investors (the asset-liability man- agement option). What are the options? • Invest directly in long-dated assets, but As discussed in chapter 2, one of the main limit the risk by relying on secondary mar- functions of � nancial intermediation is pre- ket liquidity and/or by purchasing market cisely to assist in closing the two key related insurance (the market insurance option).4 mismatches that arise in the process of going • Invest directly in long-dated assets, but long. There is first a mismatch between limit the risk through public guarantees households’ need to invest long term (for (the public insurance option). example, for old age or in education) and their natural preference for staying short so Clearly, these options are not mutually as to have ready access to their funds. And exclusive. In fact, the asset management then there is a mismatch between lenders’ option and the market option work largely preference for staying short and borrowers’ together, as asset managers invest mostly in need for going long. Thus, depending on the the market. However, these options bring degree of risk and uncertainty aversion, up conceptually distinct issues and provide, the perceived risk-return trade-offs, and therefore, a useful way to organize the dis- the constellation of relative risk premiums, cussion. All of them face problems. investors can In the asset management option, inves- • refuse to “bite the bullet� and remain in tors can choose between a variety of funds, short-term positions; from the most liquid—money market • “bite the bullet� and go long even if this funds—to the least liquid—hedge funds, means becoming illiquid; venture capital funds, private equity funds— • go long, but only where they perceive that with defined-contribution pension funds, they can exit early if needed (that is, if other mutual funds, and personal brokers they bene�t from the liquidity provided lying somewhere in between. 5 In all cases, by the market or an institution); or investors are fully exposed to the risk of the • go long, but only if they can limit risk by fund investments and face a basic agency purchasing insurance from other market problem: Will the asset managers adequately participants or from the state. ful�ll their clients’ interests?6 If investors are proactive and switch funds in response to Frictions (both agency and collective) will fund fees and short-term returns (as typi- dictate which (or which combination) of these cally occurs in mature stages of financial options is likely to dominate at a given stage development), asset managers may take too of �nancial development. Thus, depending on much risk and trade too much in order to how �nancial services shape up to circumvent attract customers by boosting short-term or limit the frictions, investors in practice face returns.7 In that case, though they may the four following options: appear to be “alpha traders� who beat the • Invest indirectly in long-dated assets (via market by having superior skills, they actu- a fund), thereby remaining fully exposed ally are buying assets with substantial tail to the risks, but rely on an asset manager risk. 8 By contrast, if principals (investors) for the day-to-day management of the risk are passive (as is generally the case in the and an early exit option when needed and less developed systems), asset managers have if allowed by the fund (the asset manage- little incentive to screen and monitor coun- ment option). terparty risk, may not optimize in terms of • Invest in a leveraged intermediary, such the risk-return trade-off, spend too much in as a commercial bank or an insurance marketing, trade too little, and still collect 124 GOING LONG juicy fees. Except for the most illiquid and In the market insurance option, within specialized funds, where investors know the suf�ciently developed systems, investors may manager suf�ciently to entrust their funds buy directly from a broad menu of debt and for extended periods of time, liquidity (the equity securities. Debt securities (bonds) ability to withdraw on short notice) is the range from those that mainly bear market name of the game. This can often lead asset risk (covered bonds10) to those that bear both managers to promise more liquidity than market and default risk (asset-backed securi- they can deliver. ties), with structured products (for example, In the asset-liability management option, with senior tranches or built-in guarantees) investors again face a range of options, from somewhere in between. To cope with the the shortest and most liquid (commercial risks of such long-dated exposures, investors banks’ checking account deposits) to the may rely on market liquidity (that is, count longest and least liquid (life insurance and on being able to dispose of the security at annuities), with repurchase (repo) agree- little cost by selling it in the secondary mar- ments and investment banks lying some- ket). Where available, investors may also where in between. Asset-liability managers use market insurance, hedges, or deriva- (ALMs) fully benefit from the upside of tives, such as credit default swaps, to hedge their investments and provide insurance and against credit risk, or swaps, futures, and liquidity to investors by interposing capital, options to hedge against market risk (inter- provisions, or liquidity buffers between their est rate risk, currency risk, and equity price assets and liabilities. ALMs, however, do not risk). As illustrated by the recent global cri- bear the full downside, reflecting limited lia- sis, none of these instruments is problem free. bility as well as public policy (deposit insur- For instance, asset-backed securities put all ance or bailouts). Hence, unless checked the risk squarely on investors (or the guaran- by market discipline or regulation, ALMs tors) but may expose them to originate-and- have an incentive to screen and monitor distribute incentive problems. Covered bonds but may still take too much risk with some- may shift the credit risk to other bank claim- one else’s money. ALMs will also naturally ants, including depositors, who become junior tend to engage in maturity transformation, claimants; structured products create risks of acquiring interest rate risk but pocketing excessive complexity and opacity; derivatives the maturity premium if things go well, and pose counterparty risk. In addition, insurance taking capital losses or going bust if things markets are quite incomplete, even in mature go badly. In both cases, ALMs raise the �nancial systems.11 stakes through leveraging. ALMs focused In the public insurance option, govern- on life insurance have their funding secured ments can provide direct long-term � nance (for example, through the sale of long-term through first-tier public banks or facilitate insurance policies) and are thus not exposed the acquisition of long-term finance from to the risk of runs. Such freedom from runs the private sector (in the form of commercial and the need to match assets to long-dated loans, infrastructure bonds, housing bonds, liabilities make life insurance companies the annuities, and even equity stakes). In the sec- ideal dedicated long-term investor. ALMs ond case, the state can provide explicit guar- focused on banking, by contrast, offer antees aimed at limiting the risk of long-term deposit redemption at par and on demand instruments or implicit guarantees aimed and, hence, can be subject to liquidity runs at facilitating the early exit from long-term that induce them to engage in socially inef�- finance commitments by the private inves- cient � re sales and deleveraging.9 This limits tors (liquidity lending of last resort or risk the maturity transformation that banks can absorption of last resort). In all cases, the perform. Through runs or insolvency, both main problem is the inadequate pricing of types of ALMs can ultimately fail, however, risk that may undermine monitoring incen- prompting public bailouts. tives and promote social moral hazard (that GOING LONG 125 is, the incentive to take on excessive risk with In sum, for financial development to be the expectation of keeping the gains but pass- well anchored and sustainable, a solution ing on the losses to the state). needs to be found where investors invest with Thus, depending on whether and which a long-term horizon (including in illiquid (agency or collective) frictions are resolved securities) but with only limited risk absorp- through the above market arrangements or tion by the state, or if with risk absorption by policy enhancements, there are four possible the state, with adequate pricing of the public outcomes: guarantees. 1. All frictions and incentive problems are satisfactorily resolved. Where is LAC? 2. Agency frictions are not resolved. LAC has made considerable progress over the 3. Agency frictions are resolved, but collec- past decade in addressing many of the issues tive (participation) frictions are not. raised above. In particular, it has achieved 4. Whether agency frictions are resolved or a remarkable lengthening of maturities in not, the state assumes a dominant role in its public sector debt and has by and large resolving collective frictions. (It can do so restored the primacy of its local currencies through guaranteeing private contracts with respect to the dollar, as documented in or becoming a financial intermediary chapter 3. Yet challenges lie ahead. Among that issues long-term bonds and provides the many wonders expected from defined- long-term � nance directly through � rst- contribution pension funds (several of which tier operations.) have indeed been realized, at least partially), Outcome 1 is, of course, the happy one, many hoped that the funds would also where full � nancial lengthening (investing in contribute to lengthening the maturity of long and possibly illiquid instruments) takes investments and overcome the lack of liquid- place. Outcome 2 is the opposite case, where ity in capital markets. Yet, the reality in these financial lengthening and deepening just particular aspects has been somewhat sober- does not happen (short-termism dominates). ing, for, as shown in chapter 3, in much of Outcome 3 is where investors invest long the region these funds continue to concen- term but only if they can exit by resorting to trate their portfolios in public sector bonds or secondary market liquidity. However, under relatively short-term bank deposits or in pri- systemic events, unless the state is there vate debt securities. In many LAC countries, to provide liquidity of last resort (thereby the public sector remains the main or even effectively converting a type 3 outcome into the only entity able to provide, guarantee, or a type 4 outcome), this raises an insoluble enhance long-term debt � nance. And this is collective action problem. Moreover, as dis- despite the fact that the region is nowadays cussed in chapter 2, this can encourage free awash with investible funds, which makes riding on market liquidity and weaken mon- the situation all the more puzzling. Clearly, itoring incentives, thereby also exacerbating going long is harder than often believed. agency problems. Hence, absent regulation that internalizes systemic risk, the � nancial Pension funds system will develop, but in a brittle, crisis- prone way. Similarly, in outcome 4, the De� ned-contribution pension fund admin- absorption of risk by the state may help istrators (PFAs) have become the dominant deepen the availability of long-term con- asset managers in several � nancial systems tracts, but this is likely to be accompanied throughout the region.12 Because of imper- by social moral hazard (that is, the transfer fections on the demand side (that is, the con- of risk to the state), giving rise to fault lines tributing workers) and complex industrial that will ultimately rupture with devastating organization issues,13 they suffer from three repercussions. interrelated problems: they are expensive for 126 GOING LONG what they do; they invest for the most part as pure asset managers (rather than ALMs). in relatively safe, liquid, plain vanilla assets; Their incentives are tilted toward the short and they tend to spend too much on mar- term and against risk taking. 20 Thus, initial keting and collection. Although the share hopes for a direct impact of pension funds of government bonds and bank deposits has on the development of long-term � nance in been declining, it still accounts, on average, a broad and sustainable manner have so far for about 60 percent of PFAs’ portfolios.14 been ful�lled only to a limited extent.21 And where PFAs invest in market instru- ments, they mostly buy and hold, limiting Mutual funds their investments to the highest rated and most liquid instruments. The preference for In contrast with pension funds, much less is investing short term and safe, rather than known about LAC’s mutual funds. In prin- long term and risky, largely reflects the fact ciple, the high interest rate margins of com- that de� ned-contribution pension funds are mercial banks open much space for mutual pure asset managers that do not have a for- funds to compete and to offer better returns mal liability, let alone a long-term liability. compared to bank deposits. However, Pensioners, not PFAs, are the ones absorbing LAC’s mutual funds appear to bask in the all the investment risks.15 Finally, and most glow of these high margins more than com- importantly, to take advantage of both the pete. Reflecting the importance of � nancial higher liquidity and the risk diversi�cation conglomeration and name recognition, they bene�ts, PFAs increasingly invest offshore are heavily concentrated and tightly related rather than onshore (�gure 3.21b). to banking groups. Their fees are substan- Management fees have been coming down tially above the 1 percent to 2 percent range but are still relatively high, resulting in gener- that prevails in most developed countries ally very high rates of pro�tability for PFAs.16 (table 8.1). 22 At the same time, the proceeds To the extent that fees are proportional to from the fees are mostly spent on advertis- inflows and that fees collected for past contri- ing and distribution rather than on asset butions do not move with workers when they management. switch PFAs, a PFA that captures relatively In fact, mutual fund managers in LAC younger workers (with fewer assets) will have appear to have few internal or external incen- higher earnings than one that serves mainly tives to engage in high-quality asset manage- the oldest workers (with more assets).17 Since ment. Performance bonuses are rarely used; net inflows to PFAs do not strongly respond to when they are used, they are paid on gross investment performance or management fees returns, with no distinction between alpha but tend to respond instead to the number of and beta risk. 23 Moreover, the size of the salespeople, PFAs have a built-in incentive to bonus is small, managerial turnover is high, attract the youngest customers by spending and portfolio management is usually an heavily in marketing.18 The available evidence entry-level position.24 Nor are there substan- also suggests that PFA managers face strong tial performance incentives coming from the direct incentives toward conservatism. Com- market. The systematic response of mutual pensation generally increases in relation to funds’ customers to short-run returns is lim- the ranking among peers in monthly or quar- ited in the region, and investor behavior is terly gross returns, which raises incentives rather volatile.25 In addition to a possible lack to take risk. However, there are also tight of investor sophistication, the latter is likely controls and penalties for deviating from the to reflect a lack of transparent and standard- industry benchmark or hitting the regulatory ized information, which prevents meaningful band, which limits risk taking and promotes and timely benchmarking between peers. herding.19 On balance, the evidence sug- Altogether, the incentives faced by mutual gests that, despite their presumed long-term fund managers overwhelmingly lead them to investment mission, PFAs behave very much invest conservatively by buying short-term GOING LONG 127 TABLE 8.1 Mutual fund fees in selected Latin American countries percent Annual Funds with Entry fees Exit fees Office Funds with Sample fixed performance (% amount Exit (% funds with expenses minimum of   fee fee deposited) fees sliding scale) fees investment funds Balanced funds Argentina 2.6 0 0.5 0.50 0 0.0 50 2 Brazil 2.0 70 0.0 0.50 0 0.0 100 10 Chile 4.8 0 0.0 2.38 60 0.0 80 5 Colombia 1.9 0 0.0 0.00 20 0.0 100 5 Mexico 4.2 0 0.0 0.00 0 0.8 50 4 Peru 3.0 0 0.0 1.17 0 0.0 100 3 Average 3.1 12 0.1 0.76 13 0.1 80 Bond funds Argentina 2.6 0 0.2 3.00 20 0.0 80 5 Brazil 2.4 0 0.0 0.00 0 0.0 100 10 Chile 1.5 0 0.0 0.00 40 0.0 100 5 Colombia 1.6 0 0.0 0.00 0 0.0 100 4 Mexico 3.2 0 0.0 0.00 0 0.0 60 5 Peru 2.5 0 0.0 1.92 0 0.0 100 3 Average 2.3 0 0.0 0.82 10 0.0 90 Equity funds Argentina 3.9 0 0.2 0.40 20 0.0 80 5 Brazil 2.3 40 0.0 0.50 0 0.0 100 10 Chile 4.4 10 0.0 0.00 70 0.4 70 10 Colombia 0.6 0 0.0 0.00 0 0.0 100 1 Mexico 3.7 0 0.0 0.00 0 0.2 20 5 Peru 3.2 0 0.0 1.83 0 0.0 100 3 Average 3.0 8 0.0 0.46 15 0.1 78 Money Market Funds Argentina 1.2 0 0.2 3.00 20 0.0 80 5 Brazil 2.1 0 0.0 0.00 0 0.0 100 10 Chile 1.3 0 0.0 0.00 0 0.0 80 5 Colombia 1.2 0 0.0 0.00 0 0.0 100 4 Mexico 4.0 0 0.0 0.00 0 0.0 80 5 Peru 2.2 0 0.0 0.08 0 0.0 100 3 Average 2.0 0 0.0 0.51 3 0.0 90 Source: Raddatz 2011. or liquid assets or buying and holding only groups and frequent mark-to-market weak- the most highly rated assets, including a high nesses, reinforces mutual fund incentives to proportion of government bonds and bank invest mainly in highly liquid assets while liabilities (table 8.2). 26 The relative impor- creating systemic vulnerabilities to runs. tance of money market mutual funds that invest heavily in short-term, low-risk instru- Personal brokers ments leaves even less scope for searching for yield. At the same time, the strong empha- Mutual funds’ weaknesses—together with sis on allowing investors to have immediate the high wealth concentrations that are access to their money exposes mutual funds typical in the region—appear to open sub- to signi�cant redemption risk. This, together stantial room for personal brokerage and with mutual funds’ strong links with banking wealth management services for the more 128 GOING LONG TABLE 8.2 Share of deposits in portfolios of Chilean mutual funds Type of mutual funda Mean Minimum Maximum Cumulative share of total assets Money market funds 78.0 8.2 98.0 40.0 Medium- and long-term funds 30.7 0.0 88.9 75.1 Short-term maturity 65.5 21.9 93.1 86.8 Capital market funds 0.7 0.0 2.8 92.3 Source: Raddatz 2011. a. Types according to circular number 1578, Values and Securities Supervisor (SVS). http://www.svs.cl/normativa/cir_1578_2002.pdf. sophisticated and wealthier investors. How- extent. While, initially, regulation severely ever, the industry is mostly unregulated, little constrained the participation of institutional is known about its operations (except that, as investors in such funds, ongoing reforms are with mutual funds, a �xed fee is charged on increasingly opening the door for investment assets under management), and there are very in these instruments by pension funds and few directly available statistics on its impor- mutual funds. How much they invest in long- tance.27 The preference for personal brokers term assets is, however, dif�cult to ascertain, reflects a gradual, step-by-step switch from given the dearth of data. “related lending� to “related investing.� The increasing availability of public information Commercial banks gradually reduces the importance of tradi- tional banking intermediation. Yet, because Commercial banks for the most part stay of the lack of suf�cient public information on short, largely matching the duration of their asset managers, investors continue to rely on assets to those of their liabilities. This, in private information to pick and choose them. turn, reflects their own risk-management However, brokers also seem to be increas- preferences—LAC banks are already highly ingly attracting smaller and less sophisticated pro�table without taking too much risk, as investors, much as they do in U.S. investment shown in chapter 10—as well as pruden- banks. As a result, given the lack of regula- tial requirements, which generally penalize tion and transparency, some supervisory maturity and duration mismatches. Com- authorities in the region are starting to voice mercial banks limit the risks of going long by concerns about issues of investor protection. using floating-rate mortgages or by match- While the portfolios managed by personal- ing the long duration of their assets to that ized wealth management services are more of their liabilities through covered bonds (as likely to be invested with long horizons in in Chile) or long-term public � nance (as in mind, it appears that a large share is invested Brazil or Mexico, where commercial banks in international markets. channel second-tier long-term loans from development banks). Until the global crisis hit, commercial banks in the region were Hedge funds, venture capital funds, increasingly following the route taken in and private equity funds advanced economies, of passing their long- High-risk, low-liquidity funds aim at the next term exposures on to institutional investors stage of wealth and sophistication. They are and the capital markets while increasingly starting to appear in the deepest markets of relying on fee income (rather than on inter- the region, particularly Brazil, with an often mediation margins) to boost pro�ts. dominant participation of offshore funds. 28 Although in LAC, these types of funds are Investment banks mostly unregulated and seem to have the capacity to leverage themselves, they do not In many LAC countries, commercial banks seem to have done so yet to a substantial are not allowed to engage in capital market GOING LONG 129 activities directly. Instead, � nancial groups While competition can lead the life insur- can engage in capital market activities ance companies to translate the higher through an investment banking subsidiary returns from investing long into attractively with a separate license. In other countries, priced policies, those companies become, in brokers and dealers effectively conduct the the process, exposed to nontrivial solvency business of investment banking. Thus far, risks. At the same time, the illiquidity of LAC investment banks (and broker-dealers) their liabilities, the complexity of their busi- have mainly conducted brokerage and ness activities, and the structure of their underwriting activities. While they have contracts—which are concerned with limit- scope for engaging in large-scale maturity ing moral hazard on the side of the insured transformation, as in the United States, thus rather than that of the insurance company— far they have not taken the opportunity, severely limit the scope for direct market in large part because capital markets have discipline from policy holders. 32 Nonethe- not yet reached the stage where � nancing less, reinsurers (who understand the insur- long-term securities with overnight repos ance business well) are likely to exert more might be viewed as a sustainable and pro�t- effective market discipline on those insur- able activity. In addition, LAC investment ance companies that choose to reinsure their banks, where they formally exist, tend to be residual exposures. In any event, insurance subject to the same capital and other pru- companies in general, and life insurance dential regulations as commercial banks. companies in particular, need to be heav- Hence, the type of leverage in investment ily regulated and effectively supervised to banks that we saw in the United States is ensure that their assets (including unpaid not feasible in LAC. premiums) can cover their expected liabili- ties and that short-run pro�ts are not over- stated (or, equivalently, that expected losses Life insurance companies are adequately and promptly provisioned). With the notable exception of Chile, life Substantial progress appears to have been insurance products in LAC are underdevel- made in this area in many countries where oped, in part reflecting the remnants of the life insurance companies are signi�cant, but sour memories left by inflationary erosions much remains to be done. of the not-too-distant past. In Chile, their An important related issue is that, while success has hinged essentially on three fac- annuities-oriented life insurance is already tors: the establishment of the UF (Unidad a maturing industry in Chile and poised to de Fomento), a CPI-indexed unit of account thrive in LAC countries where Chile-style that eases long-term investing by allowing pension reforms were implemented, there is an investors to hedge against inflation at virtu- insuf�ciently smooth connection between the ally no cost; the mandated nature of de� ned- accumulation phase (the period during which contribution pension funds, which creates a workers contribute to their pension funds) captive demand for annuities upon retire- and the decumulation phase (the period after ment; 29 and the government guarantees on retirement, when workers withdraw their annuities, which limit the risks of going long. funds or receive a monthly pension from an Thus, in sharp contrast with mutual funds annuity until death). Pensioners are subject and defined-contribution pension funds, to important risks at the juncture of the two life insurance companies, especially those phases (the so-called annuitization risks) that provide annuity products, invest a large that need to be addressed. In other countries fraction of their portfolio at long horizons.30 where the systems of de�ned-bene�t, manda- They have to search for long-term assets to tory, privately administered pension funds are match their long-term liabilities and are not maturing, much policy attention will need to exposed to runs—that is, lack of liquidity is be paid to ensure an adequate formation of a not a pressing issue.31 robust annuities industry.33 130 GOING LONG Housing �nance in the possession of commercial banks, which therefore capture the tax breaks. Ini- The limited development of mortgages in tially, commercial banks acquired the MBS most of LAC (documented in chapter 6) is to recompose their balance sheets in the testimony to the difficulties of going long wake of the 1998–99 housing � nance crisis (box 8.1). In Chile, which has the deepest involving savings and mortgage corpora- mortgage market of the region, mortgages tions (Corporaciones de Ahorro y Vivienda). h ave d e velop e d mo s t ly a rou nd U F - As in Chile, Mexico, and Uruguay, plans are denominated covered bonds issued by banks under way to prepare new legislation that at terms that match the mortgage loans, would facilitate the issue of covered bonds with life insurance companies and pension under a modern contractual and legal frame- funds being the largest buyers and holders work (see box 8.1). of such bonds. In Brazil, the mortgage market remains I n Mexico, the private market for largely under the direct control of large state mortgage-backed securities (MBS) that was institutions with tailor-made arrangements, starting to mushroom under the nurturing including a government-owned commercial care of state-related institutions—mainly, the bank (Caixa Economica Federal) that is the Federal Mortgage Society (SHF—Sociedad main player of the housing � nance system, Hipotecaria Federal), the National Housing SBPE.35 The latter is underpinned by below- Fund for Workers (Infonavit), and the Hous- market cost of funding, mobilized through ing Fund of the Social Security and Services passbook savings accounts ( Caderneta Institute for State Employees (Fovissste)— de Poupança) at semiadministered rates became mostly frozen in their tracks in the (taxa referencial), and a taxlike worker wake of the global � nancial crisis. Investors contribution (Fundo de Garantia do Tempo were spooked by the apparent similarities de Serviço, or FGTS). This system, which between the SHF-led mortgage development churns out long-term mortgages from short- strategy and that of Fannie Mae and Fred- term savings deposits and taxlike funds, die Mac in the United States. Contagion results in a segmentation that hampers the was exacerbated by the � nancial dif�culties development of market-based finance and encountered by some newly unregulated but hence the overall provision of housing loans, poorly monitored Sofoles, the special-purpose while generating some liquidity issues for the finance companies largely funded by SHF, SBPE sector itself.36 and by the market’s dif�culty in sorting out the good Sofoles from the not so good. The Bonds insuf�cient standardization of securitization issues did not help. Thus, the market for MBS Progress in macroeconomic management, in Mexico has become mostly circumscribed along with strong demand from institu- to issues by the large public entities in charge tional as well as foreign investors over the of developing low-income housing (Infonavit last decade, has contributed to considerable and Fovissste). However, these public issues improvement in LAC in lengthening the have tended to crowd out the private issues maturity of both public and private bonds because they issue paper on terms that pri- and in moving away from the perils of dol- vate issuers cannot match. A legal framework larization. As documented in chapter 3, yield to introduce covered bonds is now being pre- curves have been extended and flattened, and pared as part of a reform of bankruptcy rules the local currency has made substantial gains applicable to banks.34 relative to the U.S. dollar. The development In Colombia, better standardization, of longer-maturity, �xed-rate, local currency large tax breaks, and important regulatory corporate bonds is particularly remarkable, advantages have kept the market for MBS especially in Chile, Colombia, and Mexico. more whole. Yet most of the issues remain In part, this success has reflected investors’ GOING LONG 131 BOX 8.1 Covered bonds versus mortgage-backed securities: LAC’s recent experience LAC had a leg up on developing M BS. Many the regulation of �duciary functions, the demise of countries developed—or in some cases improved the monoline insurers, and sheer contagion from the (Brazil, Chile)—their securitization frameworks disturbingly similar U.S. debacle all contributed fur- in the 1990s, including legal structures, true sale ther to the severe downgrades imposed by the rating criteria, accounting rules, and tax treatment. In agencies. As a result, private MBS came to a halt. fact, LAC pioneered the concept of trust in the At the same time, however, Infonavit and Fovissste civil law context that prevails in the subcontinent. (whose activity grew partially in a countercyclical Other technical innovations that laid the ground- way) developed their own MBS programs. These work for securitization included: (a) the use of programs were a big success, not only because of the mortgage loans legally designed to be easily trans- low-risk pro� le of the portfolios, but also because of ferable and sold to institutional investors, such the attractive conditions the public provident funds as Mutuos Hipotecarios Endosables in Chile and were able to afford. Instead, the private sector was Letras Hipotecarias in Argentina; (b) the judicious largely crowded out. Thus, Mexico’s main challenge use of the existing mutual fund legal structure as at this point is to reinstate the conditions for a viable a basis to set up special purpose vehicles (SPVs) private MBS market. in Brazil; and (c) the securitization of future cash Colombia, on the other hand, has not been as flows stemming from leasing of housing contracts adversely affected by the global financial crisis. in Chile. The market that grew the most in this One major reason, aside from the very strict lend- initial period, thanks to this ef� cient legal back- ing norms set by the 1999 mortgage law, has been ground (and temporary monetary stability), was the crucial role of a private central structure, Titu- Argentina. However, the development of MBS larizadora Colombiana (TC). Since 2002, TC has has been far from linear and remains in question been providing comfort to investors by its diligence today. Colombia and Mexico, the two countries in arranging deals and organizing market transpar- where active MBS exist, provide valuable develop- ency. The tax relief enjoyed by MBS investors until ment lessons. 2011 also contributed to investors’ success. In Mexico, MBS grew in parallel with nonde- Although several countries, including Argen- pository mortgage lenders (Sofoles and Sofomes), tina, Colombia, and Paraguay, have a framework the main drivers of the market bounce after the for covered bonds, Chile is the only case where “tequila crisis.� This development was supported by this instrument has been actively used. In fact, guarantees from the state-owned SHF and enhance- the Letras de Credito Hipotecario (LCH)—based ments by U.S. monoline insurers. When commercial on a pass-through mechanism guaranteed by the banks became active again in housing � nance, their mortgage originators—had been in existence since involvement helped diversify the source of MBS 1855. 37 However, the issuer of LCH, a public spe- overseen by SHF. A dual oversight system devel- cialized institution, sank with the financial crisis oped, wherein the Comisión Nacional Bancaria y of the 1930s. When the savings and loan (Sistema de Valores (CNBV) effectively delegated much of Nacional de Ahorro y Préstamo, or SINAP) system the oversight to the SHF, which re� nanced or guar- collapsed, the LCH were revitalized and combined anteed the loans on a contractual basis. However, with the newly created private pension funds and the scheme was severely jostled when the 2009 the widely accepted UF, thereby triggering a remark- economic downturn affected the Mexican housing able development of housing finance. In fact, the � nance market. As Sofoles’ portfolios—which were very long maturities (20 years) and � xed (real) rates largely originated through developers and included appeared in Chile much before such features became a large portion of developer loans—incurred grow- available in many more-advanced economies. LCH ing nonperforming loans, a con�dence crisis devel- were improved by allowing commercial banks, and oped that hindered the rollover of the short-term not only nondepository specialized lenders, to issue debt imprudently issued by some Sofoles. Gaps in them and by designing a more precise process for (continued next page) 132 GOING LONG BOX 8.1 Covered bonds versus mortgage-backed securities: LAC’s recent experience (continued) enforcing the privileged status of bondholders in with strict regulatory norms; and (c) decreasing case of a bank’s insolvency. pro�tability of the LCH business—given the pass- LCH have become much less pivotal, however, through nature of the bonds, banks earn a � xed especially since 2004, progressively losing market commission with no accessory incomes such as share (11 percent of housing portfolios in 2010, treasury reinvestment or active asset-liability man- down from 86 percent in 1995). Three main fac- agers. 38 Recently, LCH regulation was given more tors explain this disaffection: (a) a wave of prepay- flexibility, allowing, in particular, 100 loan-to- ments in 2004–05 in a context in which the value value ratios for the highest rated borrowers and of the prepayment option was not priced properly; adjustable rates. At the same time, the regulation (b) lack of flexibility—LCH can fund only new allowed the funding of housing loans through loans, cannot be restructured, and must comply unsecured “bonos hipotecarios.� preference for debt over equity, arguably due is, almost by de� nition, highly illiquid, not to residual problems in corporate governance least because it typically requires complicated and other hindrances to the development tailor-made structures. This tends to limit of domestic equity markets, as discussed in signi�cantly the appetite for infrastructure chapter 7. Yet problems remain. For example, bonds among mutual funds and defined- in countries such as Brazil, most local cur- contribution pension funds. Moreover, a key rency corporate bond issues continue to be piece of this equation went missing in the short maturity or floating rate. Secondary wake of the global crisis with the demise of market liquidity for corporate debt remains, the monoline insurers, which played a key by and large, a key hindrance to further mar- role in ensuring the rating and enhancing the ket development. And issuing costs continue market attractiveness of the bonds.40 Govern- to be high, limiting access to all but the larg- ments are thus now faced with the dilemma est �rms.39 of whether and how they might need to step Together with mortgage bonds, infrastruc- back in to �ll the guarantee gap. ture bonds provide the other natural source of supply of longer-term �nancial assets. Public- Long-term hedges private partnerships (PPPs) and the growth of institutional investors (pension funds and Except perhaps in Chile—which has patiently insurance companies) have both played a cru- and persistently managed to develop a world- cial role in closing the gap between demand class CPI-indexation scheme with deep, liq- and supply. In many LAC countries, however, uid markets and a well-trusted indexation development banks (for example, BNDES in methodology—indexation remains a stum- Brazil, Banobras in Mexico) are the key play- bling block to long-term � nance in most of ers in providing long-term local currency LAC (see box 8.2).41 Finding a good point of finance for infrastructure projects. Syndi- intersection between what borrowers need cated bank loans and credit from multilateral (affordable � xed interest rates or effective agencies also play an important role, but protection against real wage fluctuations) more often than not in foreign-currency- and what investors want (credible, clean pro- denominated � nance. Infrastructure � nance tection against consumer price fluctuations GOING LONG 133 BOX 8.2 LAC’s indexation experiences in housing �nance LAC countries have struggled over the years with Another approach has been to cover the mismatch the inherent tension between indexes that accom- between assets and liabilities with prudential buf- pany borrowers’ capacity to pay (that is, that are fers. For example, in Argentina in 2004, Banco linked to the real wage) and those that accompany Hipotecario started securitizing fixed-rate peso investors’ purchasing power (that is, that are linked loans while offering investors variable rates. Simi- to CPI infl ation). While the former limit credit risk larly, the Mexican provident funds (Infonavit and and are favored by borrowers, the latter are pre- Fovissste) have issued CPI-indexed MBS based on ferred by investors. Banks are torn between the housing loans indexed to the minimum wage. In both two and can be faced with large asset-liability mis- countries, the credit enhancements needed to sup- matches when they attempt to reconcile both sides port such structures have been very high (25 percent of the market. or more). I n A rgentina, the widening asset- This was the case in Argentina in 2002. As an liability gap eventually deterred lenders from assum- early response to the crisis, loans were indexed to the ing the risk. In Mexico, the provident funds have CPI to alleviate the impact of pesi�cation on lenders. been able to sustain this scheme only because they However, as inflation surged, indexation was switched are not subject to the same capital charges as com- to the wage index. But then, as wages increased dra- mercial lenders and because real wages have not matically, lawmakers reverted to CPI indexation. In fluctuated much over the past 10 years. Colombia, the definition of the UPAC (Unidad de In a few cases, governments eventually took over Poder Adquisitivo Constante) index changed several the responsibility for narrowing the gap by design- times—from a CPI with a cap to a CPI–interest rate ing government-backed hedging mechanisms. In combination to an interest rate on short-term depos- Mexico, the government developed the “UDI [Uni- its. However, when short-term real rates rose in 1998, dad de Inversión] swap� to hedge the real wage this triggered a surge of nonperforming loans and a index against the CPI. The UDI swap had good suc- fall in housing prices, eventually triggering a bank- cess until the correlation between minimum wages ing crisis. In Mexico, a double indexation scheme— and inflation increased, making it less attractive. In indexing loan balances to the CPI, or using variable Colombia, the government established the Fondo de rates but limiting the increase of monthly installments Reserva para la Estabilizacion de la Cartera Hipo- to the increase of wages—was introduced. However, tecaria (FRECH) through which mortgage lenders as the scheme resulted in large losses for banks, it was can buy protection against interest rate hikes (both dropped in 1993 and replaced by variable rates linked arrangements are priced fairly). FRECH’s actual to government bonds. Soon thereafter, however, as usage has been limited, possibly for pricing reasons, interest rates rose during the “tequila crisis,� banks but probably also because of the stabilized macro- stopped extending housing loans. economic conditions. with a deep secondary market) has proved longer-term � xed-rate instruments, remain to be dif�cult. relatively undeveloped in most countries.42 Derivative markets have developed sub- stantially in the largest LAC countries over the past 10 to 15 years. Foreign exchange What are the key policy derivatives have become particularly deep in challenges? the wake of the free floating of the currencies The road to developing markets for long- and, as in Chile, of the regulatory require- term financial contracts is a difficult one. ments on the hedging of pension funds’ dol- This section stays away from policy details lar investments. However, interest rate swaps, and focuses instead on broad directions, the key piece underlying the development of based on the general roadmap followed in 134 GOING LONG this report. According to this roadmap, the returns),44 the most promising route at this role of policy is to help reduce the agency stage is to promote standardization and other or collective frictions that hinder progress ways of organizing information (including toward going long. better benchmarks) that facilitate perfor- Concerning agency frictions, one finds mance comparisons across asset managers or four generic underlying policy issues. (a) How ALMs. The main pitfall to avoid, however, is can the state continue to improve the enabling to exacerbate investors’ focus on short-term (macroeconomic and institutional) environ- returns, which could in turn heighten asset ment so as to facilitate long-term financial managers’ incentives to operate with short contracting? (b) How can the state encour- horizons. This caution is particularly impor- age market discipline, and how far should it tant for pension funds. To prevent such prob- go in doing the monitoring itself on behalf of lems, pension fund supervisors should help investors? (c) What market arrangements are lengthen the benchmarks used to assess pen- more likely, at the current stage of � nancial sion funds’ performance.45 A practical way development in LAC, to provide a successful to do so is to introduce a life-cycle fund in mix of risk spreading and sound monitor- the context of a multifund scheme—a project ing? (d) Finally, how far should regulators go Chile is now putting to the test.46 As pro- in limiting the maturity transformation risk posed by Blake, Cairns, and Dowd (2009), taken by � nancial intermediaries? The fol- regulations that nudge de� ned-contribution lowing considers each of these in turn. pension funds into mimicking the investment As regards the institutional and macro- behavior of a defined-benefit pension fund economic environment, two issues seem to (which has a formal long-term liability that deserve special mention. First, countries will it must try to match with long assets) should need to strengthen contract rights and their also help. enforcement, two areas where (as noted in In addition to encouraging market moni- chapter 6) LAC lags. Second, countries will toring, governments must also decide the need to consolidate the gains made toward extent to which they are willing to play “big using the local currency in long-term con- brother� to investors through public moni- tracts. In this regard, the main lessons from toring and regulation. This involvement may LAC’s indexation experience can be summed help solve problems of consumer protection up in three points. First, indexation cannot and bounded investors’ rationality. One replace macroeconomic stability; when vola- aspect of this effort implies using regulation tility becomes too high and relative prices to encourage pension funds to lengthen their fluctuate too much, it becomes very hard, if investments, including in socially bene�cial not impossible, to � nd indexes that meet all ventures such as infrastructure bonds. The needs.43 Second, good macroeconomic poli- other aspect is to make greater use of smart cies (including inflation targeting) can natu- default options that automatically channel rally limit the volatility of relative prices, to investors into the portfolios that are most the point where a single CPI is probably suf�- appropriate for them. However, the more cient. Third, even in well-stabilized countries governments take charge, the more they such as Chile, a CPI can help boost the devel- can undermine asset managers’ screening opment of long-term � nance, particularly in and monitoring incentives (just following the annuities industry. Persistence will thus the regulations to the letter can become the be needed in issuing and promoting the use name of the game). They can also under- of CPI-indexed bonds for the long-maturity mine market monitoring by investors (who segments of public debt. may think, “why bother?�). And they can As regards the promotion of market dis- generally promote moral hazard (if some- cipline, in addition to continuing to improve thing goes wrong, it is the government’s information to investors (such as increasing responsibility to � x it). In practice, there- the transparency and disclosure of fees and fore, a delicate balance between protecting GOING LONG 135 investors and undermining monitoring has governments can collect or subsidize the to be found. information needed to build indexes that can Regarding the best market arrangements be used for trading and can facilitate long- for combining market monitoring and risk term �nance; such indexes might be housing spreading, a currently dominant theme for prices by city, which can be used to gener- LAC is the choice between covered bonds and ate futures contracts on housing prices or securitization. The former ensure “skin in the insurance products for home equity, or rain- game� but fall somewhat short as regards fall statistics, which can be used to gener- risk spreading. As recently demonstrated, the ate rainfall-indexed insurance products and latter can run into the opposite problem of facilitate the lengthening of maturities in risk spreading (originate-to-distribute) with- agricultural �nance. out suf�cient “skin in the game.� Although Where there are strong economies of scale the global, postcrisis reevaluation of the ben- in collecting information, as in the case of e�ts of covered bonds translated in the region pension funds, unbundling the functions into plans to develop frameworks in coun- of collection and customer service from the tries where the tool is not available (Brazil, portfolio management function (the so-called Mexico), the question is whether further Swedish option) would enhance transparency efforts should also be made at developing and competition in the pension fund indus- securitization.47 In any event, the broaden- try while also reducing costs (hence lower- ing of second-tier � nance facilities that raise ing fees).50 Related policy actions to enhance funds collectively should also be envisaged, market liquidity might include reducing frag- especially in underdeveloped mortgage mar- mentation in securities issuance and trading, kets.48 In addition to reducing participation enhancing securities clearance and settlement costs—by spreading risk and achieving econ- arrangements, organizing securities lending omies of scale—these facilities can also help and borrowing facilities, and promoting con- promote competition and impose uniform tract standardization. In addition, liquidity lending and servicing standards to primary can be enhanced along with monitoring, by, lenders. Whenever possible, however, such for instance, (a) improving valuation meth- funding should be “mutualized� rather than ods and accounting and � nancial reporting “socialized�; that is, the second-tier � nance standards for longer-horizon portfolios and facilities should be private rather than public. (b) promoting (or even subsidizing) the pro- Finally, regarding the regulation of matu- duction of publicly available analyses of asset rity transformation, the use of risk-based cap- managers’ and ALMs’ performance. ital requirements that penalize mismatches in While such an agenda may well be war- duration can go a long way toward encour- ranted, LAC policy makers need to be real- aging ALMs with long liabilities, such as life istic about the feasibility (and desirability) insurance companies, to invest long. In this of adopting a U.S.-style solution that relies case, the regulation is win-win since it has almost exclusively on liquid secondary mar- both stability benefits (reduces the risk to kets to align investors’ desire for an exit insurers) and developmental benefits (pro- option with borrowers’ long-term fund- motes long-term assets).49 ing needs. This solution faces limitations Regarding collective frictions, a � rst ave- because it hinges crucially on size (as dis- nue that governments can take to deal with cussed in chapter 7) and can lead to systemic them is to reduce participation costs. This fragility—the “dark side.� The following can be done through continuing to improve discussion considers these two aspects in a the enabling environment and, where needed, bit more detail. promoting the availability of information As regards the issue of size, as noted in as well as proper infrastructures for collect- chapters 2 and 7, in the absence of suf�cient ing and trading information. Because much size—of markets and of issues—domestic of the information at issue is a public good, capital markets can be easily caught in a 136 GOING LONG low-liquidity trap. This constitutes sobering example, in Mexico, the demand for annui- news for the small domestic capital markets ties from retirees is postponed for decades, in LAC and helps explain why the issuance until new entrants to the system retire. 52 In and trading of Latin American stocks con- Colombia, the required years of contribu- tinue to move to the international � nancial tion and the option for workers to move to centers where global market liquidity clus- the public pension system imply that a rather ters. Thus, where domestic markets are small, small proportion of the forthcoming waves reforms—even if desirable—are likely to lead of retirees will qualify for an annuity. Such to only modest improvements in secondary flaws should be avoided whenever possible. market liquidity. To facilitate the link between pensions and As regards the dark side of market liquid- insurance, governments may also introduce ity, if market arrangements succeed in deep- state guarantees on annuities, with strong ening domestic liquidity (a big “if� for a capital buffers in life insurance companies small country), wedges between private and acting as the �rst line of defense. In addition, systemic interest can widen as private agents the composition of asset holdings in the work- increasingly free ride on liquidity and relax ers’ pension fund as they approach retirement their monitoring efforts. Alternatively, serious should be as similar as possible to the asset problems can be created if, frustrated with the portfolio composition of the life insurance illiquidity of the domestic market for long- company from whom workers will buy their term contracts, the government makes too annuity. In addition to tilting a significant expansive a commitment to providing back- share of overall pension fund investments in stop liquidity (via the central bank or devel- favor of longer-term, illiquid assets as work- opment banks) at too low or no cost. This, ers age, this should also help substantially to in turn, connects with two key issues that reduce annuitization risks.53 will be discussed later in this report: revisit- ing the risk-bearing role of the government in �nancial development (chapter 9) and adjust- Notes ing prudential oversight to deal with the dark 1. This chapter draws heavily on the papers side (chapters 10–13). “Institutional Investors and Agency Issues A second key avenue for dealing with in Latin American Financial Markets: Issues collective action frictions is to mandate and Policy Options,� by Claudio Raddatz (2011), and “Mobilizing Long-Term Resources participation rather than just promote it. for Housing Finance: Trials, Errors, and This approach includes, in particular, par- Achievements in LAC� by Olivier Hassler, which ticipation in privately administered de� ned- are part of the forthcoming Edited Volume that contribution pension funds. When combined will accompany this LAC flagship report. with policies to promote the development of Duration is the weighted average maturity a vibrant annuities industry in the decumula- of a debt contract, where the weights arise tion phase of life-cycle savings, as in Chile, from the contractual periodicity with which it provides one of the most promising levers the interest rate adjusts. If the interest rate for the development of long-term �nance. As is �xed for the life of the contract, maturity the mandatory de�ned-contribution systems and duration coincide. Duration is shorter mature, retirees naturally and increasingly than maturity if the interest rate is periodi- cally adjusted within the life of the contract. demand annuities from life insurance compa- The excess of maturity over duration is larger, nies. In turn, the latter need to match those with more frequent interest rate adjustments. liabilities systematically by investing in long- This chapter uses the labels of “long-term� term assets. and “short-term� debt contracts to denote However, the design problems that may both long (short) maturity and duration. stand in the way of the development of the 2. Recent estimates of worldwide risks and annuities market need to be removed (through returns, by Dimson, Marsh, and Staunton legal reform, among other means). 51 For (2006), based on 19 (mostly high-income) GOING LONG 137 countries, indicate that the risk premium that as a necessary condition for market disci- may be captured by going long is about 450 pline (to solve agency frictions), it becomes basis points. In choosing a proper portfolio problematic when investors react to noisy composition, there are, of course, risk-return signals by running rather than monitoring. trade-offs that need to be taken into account See Calomiris and Kahn (1991), Diamond and that depend on the age and human capital and Rajan (2000), and Huang and Ratnovski characteristics of the investor. For theory and (2011). On the social costs of �re sales and simulations along these lines, see Campbell deleveraging, see Shleifer and Vishny (2011). and Viceira (2002) and Viceira (2010). 10. These are bonds issued by a leveraged inter- 3. See Benartzi and Thaler (2007). mediary and backed by its full balance 4. The securities themselves can, of course, be sheet. Covered bonds have a priority claim structured so as to offer hedging and insur- on the assets of the intermediary in case of ance features by, for instance, incorporating bankruptcy. put options, collars, indexation, and so on. 11. See Shiller (2008) and Kroszner and Shiller 5. Hedge funds are something of a hybrid. (2011) for a detailed discussion of the social Because the funds are leveraged, hedge fund bene�ts of completing certain insurance mar- managers need to hold capital. However, the kets that currently are underdeveloped or risks of the investments are ultimately passed do not exist, for example, the markets for on fully to investors; that is, hedge funds do house-price futures (which can allow inves- not engage in asset-liability management to tors to short sell real estate without having provide downside risk insurance to investors. to sell the underlying asset, for example, the 6. Given that pure asset managers (that is, man- house); livelihood insurance (which can pro- agers of unleveraged funds) have no “skin in tect working people from economic shocks the game� (the gains or losses of the invest- to their income); or home equity insur- ment are fully passed on to the investors), their ance (which can protect mortgage debtors incentives can be aligned with those of the from unexpected losses in the value of their investor only through the way in which they houses). are compensated or through regulation. In 12. Privately administered pension funds are turn, their compensation may be affected by important in Bolivia, Colombia, Costa Rica, the direct incentives provided to asset man- El Salvador, Mexico, Peru, and Uruguay, agers by the companies for which they work which implemented Chile-style pension or by indirect incentives resulting from inves- reforms (that is, they created a mandatory, tors’ decisions to shift their portfolios across privately administered system of de�ned- competing funds. For an extensive discussion contribution pension funds as a pillar of the of incentives and compensation issues, with a national social security system) over the past focus on LAC, see Raddatz (2011). decades. Argentina had a system of that type 7. Since portfolio selection should generally until recently, when it reverted to a pay-as- depend on the investment horizon, short- you-go system. Privately administered but term return maximization is unlikely to be voluntary pension funds are also a prominent optimal from a longer-term perspective (see feature of Brazil’s �nancial system, where a Samuelson 1969; Merton 1969). signi�cant (yet declining) share of assets in 8. Distinguishing the true alpha traders is hin- this industry is held in de�ned-bene�t occu- dered by the bounded rationality of investors. pational pension funds. The rapid growth of Consistent with the �ndings of behavioral these open pension plans bene�tting from tax �nance, even �nancially literate investors advantages has in turn promoted the devel- are found not to monitor adequately the opment of the life insurance industry. behavior of asset managers and may not 13. See Impavido, Lasagabaster, and Garcia- react elastically to performance, measured in Huitron (2010). terms of fund fees and returns. See Impavido, 14. It appears unlikely that the bias toward Lasagabaster, and Garcia-Huitron (2010) for short-term government bonds and depos- an analysis of the related policy implications its could result purely from a lack of other for LAC’s pension funds. instruments. Raddatz and Schmukler (2008) 9. While the reliance on short-term funding is show that pension funds do not invest in all traditionally depicted in the �nance literature assets in which they are allowed to invest, 138 GOING LONG even among equities. Opazo, Raddatz, and they are tactical, follow the pack, and focus on Schmukler (2009) show that Chilean pension the short-term returns. See Didier, Rigobon, funds bid less aggressively for central bank and Schmukler (2011). bonds with longer maturities than for those 21. This is certainly not meant to suggest that with shorter maturities. pension fund reforms were a failure. Indeed, 15. Thus, PFAs that administer de�ned-bene�t the growth bene�ts of the system of indi- pensions appear to invest differently from vidual pension fund accounts have been well those that administer de�ned-contribution recognized. For example, Corbo and Schmidt- pensions. Hebbel (2003) �nd that Chile’s pension 16. For example, in Chile, the return on equity of reform, partly through its positive impact pension funds in 2009 was 24 percent, com- on capital market development, raised GDP pared to 17 percent for banks and 4 percent growth by an additional one-half percentage for insurance companies. point per year, on average, between 1981 17. Various LAC countries (including Costa and 2001. Nor are the limitations of de�ned- Rica, the Dominican Republic, El Salvador, bene�t, mandatory pension funds meant Mexico, and Peru) have changed (or are con- to be LAC speci�c. Indeed, the tendency of sidering changing) the fee structure of their pension funds to herd and to buy and hold pension funds from a pure flow basis (con- is widely recognized the world over. See for tributions) to a pure stock basis (assets under example Blake, Lehmann, and Timmermann management). However, as long as pension (2002). funds do both collections and portfolio man- 22. The fee structure is relatively standard and agement, neither extreme is likely to be opti- consists mainly of a management fee pro- mal as neither fully matches the underlying portional to assets under management, and, cost structure of the funds. to a smaller extent, a combination of entry 18. Cerda (2006) shows that there is a positive but and exit fees. Explicit performance fees are very small correlation between inflows and uncommon. performance. Calderón-Colín, Domínguez, 23. Alpha risk refers to the risk-adjusted excess and Schwartz (2008) show that, for the return for investing in wisely selected securi- Mexican PFAs, there is a substantial fraction ties that perform better than the market. Beta of the population shifting to the PFAs with risk refers to the volatility of a security rela- the lowest returns and highest fees. Berstein tive to that of the market. and Cabrita (2007) show that Chilean work- 24. The simple bonus structure in part reflects the ers are not responsive to price and cost fact that LAC has few or no fat-tailed instru- signals. Srinivas, Whitehouse, and Yermo ments in which mutual funds can (or are (2000) argue that the compulsory nature of allowed to) invest, lacks any standardization contributions and minimum state guarantees of investment styles and segments, and typi- on returns limits workers’ sense of ownership cally has a limited number of funds, which and reduces their incentives to exert market makes benchmarking against a style dif�cult. discipline. Instead, it may induce them (cor- 25. While there is a signi�cant correlation rectly or incorrectly) to assume that the gov- between Chilean mutual funds’ lagged short- ernment is implicitly accountable for securing run excess returns (relative to the industry their pension. Despite recent progress, assess- average) and net inflows of assets, the slope ing the performance of (and benchmarking) of this relation is small, with a 10 percent PFAs remains problematic. excess return resulting in inflows equivalent 19. Several studies have documented the presence to 2 percent of assets (Opazo, Raddatz, and of herding in trading among PFAs in Chile Schmukler 2009). (Olivares 2005; Raddatz and Schmukler 26. At �rst glance, the fact that asset manag- 2008). Strong herding is observed even where ers invest in liquid securities while tending pension fund returns are not regulated, not mainly to hold rather than to trade may least because compensation incentives are seem puzzling. However, it is fully consistent linked to performance vis-à-vis an industry behavior for an asset manager who has no benchmark. incentives to incur the higher screening costs 20. Thus, pension funds tend to be very similar of investing in illiquid instruments, combined to mutual funds in their investment behavior: with risk of early investor redemptions. GOING LONG 139 27. This form of intermediation is analogous 32. In fact, shifting insurance providers is likely to to the separately managed accounts offered be counterproductive for the insured because by �nancial companies in the United States, of preexisting conditions that can imply sig- where, according to some estimates, its size ni�cant switching costs. is similar to that of the mutual fund industry. 33. The obstacles and policy issues involved In LAC, the only way to gauge its imprint is in the development of annuities markets through the unaccounted share of �nancial for various Latin American countries are assets that is not held by domestic �nancial documented in their respective FSAP docu- institutions or foreign investors. A rough ments (http://www.worldbank.org/fsap). For calculation for Chile, based on background the case of Colombia, see also Cheikhrohou research for the latest FSAP (Financial Sector et al. (2006). Assessment Program; under preparation), 34. A securitization platform, Hipotecaria Total suggests that personal brokers and wealth (Hito), has been developed to import a managers might manage as much as 17 per- Danish-style mortgage bond structure. cent of total household wealth, substantially 35. The Sistema Brasileiro de Poupança e less than pension funds (47 percent) but more Empréstimo (SBPE), which encompasses than banks and insurance companies (about the largely predominant Caixa Economica 11 percent each). Federal and 40 other banks, accounts for 28. As of 2004, there were hedge funds located in about half of total housing �nance. Argentina, Brazil, Mexico, and many offshore 36. The taxa referencial also raises issues of locations. In Brazil alone, assets under hedge liquidity risk, as it complicates the covering fund management reached about US$200 bil- of a possible liquidity shortage with capital lion by the end of 2010. Private investment market funding. funds have started to appear in countries 37. LCH were based on similar arrangements like Brazil, Chile, and Colombia. In Chile initiated in Denmark, France, and Germany. and Colombia, they managed assets worth As in France, mortgage-covered bonds were 0.1 and 0.2 percent of GDP, respectively. introduced together with a public housing Venture capital activity is also restricted to a bank. few countries of the region, including Chile, 38. Moreover, commissions were cut, from 3.5 where assets in venture capital funds reach percent down to 1 percent, between 1984 2 percent of GDP (10 percent of the mutual and 2006. fund industry). 39. See Zervos (2004). 29. Admittedly, the demand for annuities is not 40. For references on the role of U.S. monoliners fully mandated in the Chilean pension system and local pension funds in the funding and because pensioners have the option to use guaranteeing of infrastructure bonds in LAC, phased withdrawals. However, by making the see Escriva, Fuentes, and Garcia-Herrero pension available at the time of retirement, (2010). the system creates a natural induced demand 41. For a recent review of debt indexation issues for annuities. Shifting the disability and sur- in emerging economies, see also Holland and vivorship bene�ts to the second pillar further Mulder (2006). contributed to buildup of demand. 42. While foreign exchange derivatives are 30. See Opazo, Raddatz, and Schmukler (2009) generally more liquid, their development is for the case of Chile. often constrained by market asymmetries. In 31. Notice, however, that life insurance com- the case of the major currencies, the demand panies increasingly combine traditional for opposite hedges comes from similar insurance services with access to “savings agents (for example, exporters) located on accounts� whose amounts can be withdrawn each side of the border, but for LAC curren- on demand. Indeed, such features were, in cies, the demand for opposite hedges comes part, responsible for the failure of the Clico mainly from local agents acquiring opposite group in the Caribbean. This blurring of the exposures (for example, exporters versus lines between insurance and banking is one to importers). Thus, the two sides of the market watch carefully going forward. It clearly calls can have dif�culty locating each other when for uniform risk-based regulation across all they do not have similar levels of �nancial leveraged intermediaries (see chapter 12). sophistication. 140 GOING LONG 43. The integrity and credibility of the CPI is, of encouraging intermediaries to lend short, it course, also essential. could shift liquidity risk on the rest of the sys- 44. Shiller’s (2008) six-point reform agenda to tem. As discussed in chapter 12, the solution strengthen the information infrastructure for in this case is to penalize short funding rather mortgage �nance provides a useful reference. than rewarding short investing. The proposed reform areas are subsidizing 50. A challenge under this option could be to comprehensive, independent �nancial advice; keep the quality of customer service from establishing a consumer-oriented government deteriorating, however. �nancial watchdog; adopting and adapt- 51. To be sure, the biggest challenge for LAC’s ing default conventions and standards that pension systems remains their low coverage, work well for most individuals; improving despite their mandatory nature (Gill, Packard, information disclosure on �nancial securities; and Yermo 2004). With the exception of Chile, promoting the creation of large databases of less than half of the economically active pop- �ne-grain data pertaining to individuals’ eco- ulation participates in the mandatory pension nomic situations; and creating appropriate system, and, in many countries, coverage is systems of economic units of measurement below 25 percent. There is also the growing to which �nancial contracts could be linked challenge for de�ned-contribution pension (such as CPI-indexed bonds). funds to produce an acceptable “replacement 45. Academic work on de�ning long-term invest- rate� of income after retirement (see de la ment benchmarks has made substantial prog- Torre, Gozzi, and Schmukler 2007a). These ress but now needs to translate the research challenges, however, fall outside the scope of into concrete and achievable policy propos- �nancial development policy and mainly con- als. In particular, a workable and credible cern social protection policy. decision-making structure, probably based on 52. Workers that were already in the system at the a committee of independent sages, might pro- time of the reform have a “put option�—at vide a proper institutional channel to de�ne retirement, they can “put� their accumulated and periodically review the composition fund to the Social Security Institute (Instituto of the life-cycle portfolio. See Antolin et al. Mexicano del Seguro Social, or IMSS) and (2010) and Castaneda and Rudolph (2011). receive bene�ts under the old pay-as-you-go 46. The life-cycle fund can become the default system. The development of the market for fund, that is, the fund to which workers annuities was further stalled following legal who do not positively choose a fund are changes in 2001–02 that resulted in most dis- assigned. The asset composition of that fund abled individuals electing bene�ts under the would change automatically as the worker old system, contrary to the original intention approaches retirement age. of the pension reform, which was aimed at 47. While mortgage-backed securities provide ensuring that disability, death, and workers’ higher yields, they are less easy to trade than compensation bene�ts would be provided covered bonds—at least in their most com- by specialized private annuity companies. mon form of bullet repayable debt—because Ultimately, the “put option� in the Mexican of the valuation dif�culties they raise. pension system became almost inevitable 48. Such re�nancing structures have already been because af�liates transitioning to the new pen- established in quite a few countries in Latin sion system did not receive, as their Chilean America, including Costa Rica, Ecuador, the counterparts did, the so-called “recognition Eastern Caribbean, Mexico, and Paraguay. In bonds� in compensation of the bene�ts that Brazil, the provident fund FGTS is mandated they had accrued under the under the old pay- to play a similar role. However, the inherent as-you-go pension system up to the date of the conflicts of interest resulting from the fact reform. Instead, recognition bonds in Chile that FGTS is managed by Caixa hinders its created a clean separation between the old openness toward other lenders. and new systems, making it possible for them 49. However, regulators need to take care not to be complementary, rather than substitutes. to treat all maturity mismatches equally and 53. A well-designed default life-cycle pension symmetrically. Penalizing the maturity mis- fund—like the one Chile is now experiment- matches of ALMs with mostly short-term ing with—can also help eliminate that dis- liabilities, such as banks, could have unde- continuity. See Rocha and Thorburn (2006) sirable systemic implications because, by and Rocha, Vittas, and Rudolph (2011). Risk Bearing by the State: A Collective Action Perspective 9 C hapter 2 hinted at important roles prevent markets from resolving agency that the state can play in achieving frictions. The former may justify � rst-tier sustainable financial development. state banking, the latter, second-tier state This chapter explores one such role in par- banking and state guarantees. 2 ticular, namely, that of absorbing � nancial • Such interactions have obscured the rela- risk by providing loans or guarantees. As tive roles to the point that the basic con- in other chapters of this report, the analy- ceptual underpinnings of the role of the sis proceeds along the lines of the � nance state have been too often lost from sight, paradigms by systematically contrasting the triggering polemical, lengthy, and incon- impact and implications of agency (infor- clusive debates. In Latin America and mation asymmetry and enforcement) and the Caribbean (LAC), this has given rise collective (collective action and collective over the past few decades to wide, ideo- cognition) frictions.1 The chapter makes the logically charged policy swings—a play following �ve points: in four acts that this chapter reviews in some detail. • The state has a comparative advantage • Going back to first principles should over the market in resolving collective thus help in rethinking and reorganizing frictions, while the market has a compar- the role of the state in � nancial develop- ative advantage over the state in resolving ment. agency frictions. • These comparative advantages naturally The rest of this chapter is organized as fol- suggest that states and markets should lows. The � rst �ve sections present the con- complement, rather than substitute for or ceptual underpinnings of the risk-bearing compete with, each other. role of the state within the framework of the • In practice, however, this neat separa- �nance paradigms.3 The next section revisits tion of roles can become murky because and interprets LAC’s policy swings over the the paradigms interact; thus, states may past three to four decades concerning the role need to address agency failures that hin- of the state in light of this conceptual discus- der markets from resolving collective fric- sion. The �nal section discusses policy impli- tions or, conversely, collective failures that cations going forward. 141 142 RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE The role of the state in the pure neutrality and taking into account the distri- agency paradigms butional implications of the taxes levied to � nance the state guarantees), systematically This section � rst considers the case of pure concludes that, in the absence of risk aver- agency frictions, assuming for now that lend- sion, state guarantees cannot improve the ers are not risk averse and that there are no market outcome, except when the state has collective action frictions (such as externali- an informational or enforcement advantage ties or coordination problems). According over the private sector, which is, in general, to the literature (which is vast), asymmetric hard to argue (box 9.1). information in credit markets, even without To understand what is at stake, one can risk aversion, can lead to socially inef�cient consider the student loan model of Mankiw outcomes of either underlending or overlend- (1986). This model focuses on the informa- ing. For example, Jaffee and Russell (1976) tion asymmetry problem of adverse selection and Stiglitz and Weiss (1981) demonstrate and assumes that lenders are risk neutral. the case for underlending by showing that Students’ honesty varies over the population. asymmetric information can lead to adverse However, the lender knows less than the bor- selection as higher interest rates attract risk- rowing student; speci�cally, the lender knows ier borrowers; thus, lenders may be better off the mean of the distribution but not each indi- rationing credit below the level that would vidual student’s characteristics. Moreover, be socially desirable.4 reflecting enforcement and informational The appropriate policy response to these frictions, the lender cannot force repayment. agency-driven market failures is not obvi- Hence, the lender must raise the interest rate ous. Most of the literature that finds that on all loans to cover the losses on the unpaid asymmetric information can justify state loans. But, by raising the price of all loans, credit guarantees is of a partial equilibrium the dishonest (those who do not intend to nature; that is, it does not consider the wel- repay) prevent the honest (those committed to fare effects of the taxes needed to �nance the repaying) from borrowing. Because it would guarantees. However, the general equilib- have been socially desirable for the honest to rium literature, which uses an appropriately borrow, society is worse off. stringent welfare criterion (requiring revenue BOX 9.1 Welfare criteria in the theoretical literature on state guarantees The partial equilibrium literature that does not population). With such a criterion, some papers pre- require revenue neutrality � nds that state guarantees dict that state guarantees can lead to an improved can improve things by increasing prudent credit (for equilibrium (for example, Ordover and Weiss example, Mankiw 1986; Smith and Stutzer 1989; 1981; Bernanke and Gertler 1990; Innes 1992; and Innes 1991; Benavente, Galetovic, and Sanhueza Athreya, Tam, and Young 2010). However, others 2006; Arping, Loranth, and Morrison 2010). The do not (for example, Li 1998; Gale 1991; William- literature that takes a general equilibrium view son 1994). The second group of papers incorpo- (and hence imposes revenue neutrality) can be rates the welfare impacts of tax redistribution. The classi� ed into two groups. The � rst group uses a papers in this latter group uniformly conclude that, Kaldor-Hicks welfare criterion that simply looks at without an informational advantage and the ability the total size of the pie but not at its distribution to cross-subsidize, state guarantees cannot produce across the population (that is, it assumes away the a Pareto improvement (for instance, Greenwald and welfare implications of allocating taxes across the Stiglitz 1989; Lacker 1994). RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE 143 What can policy do about this? To answer similar reasons, a private agent might con- this question, one should �rst notice that, in sider offering screening services to the lender the absence of risk aversion, an unsubsidized if he or she was better informed (hence better guarantee (that is, a guarantee priced to able to discriminate between the good loans cover expected losses) has no impact. While and the bad loans) or better able to collect it reduces risk, this is of no consequence to (hence make the dishonest pay for their sins). a risk-neutral lender. The price of the guar- However, an agent with such capabilities (for antee matches the cost of the loan loss pro- example, one able to bene�t from economies visions that the lender would have to incur of scale in putting together an effective sort- in the absence of the guarantee. As a result, ing system for borrowers) would be in the the fairly priced guarantee adds no value and, business of selling services to banks, not in hence, will not affect the lender’s behavior. guaranteeing their loans. By contrast, if the state provides a fully Broadly similar arguments can be devel- subsidized credit guarantee (a 100 percent oped when, instead of adverse selection, the default guarantee with a price equal to zero), problem underlying the failure of risk-neutral the risk-neutral lender saves the cost of loan creditors to lend to honest students is one of loss provisions and is thereby induced to lend enforcement. Supposing, for example, that to all students at the risk-free interest rate. borrowers cannot obtain a loan because From a partial equilibrium viewpoint, absent they lack good collateral; hence, they can- a requirement of revenue neutrality, the sub- not credibly commit to repaying the loan. In sidized guarantee would, therefore, allow that case, viable student borrowers without the social optimum to be reached. However, collateral would be excluded from the loan from a more stringent (and generally war- market, resulting again in a socially inef�- ranted) welfare perspective, the financing cient equilibrium. By replacing the missing of the guarantee and the distribution of tax collateral, it is often argued, a state guaran- payments across the student population also tee could bring such borrowers back into the matter. Unless the students who default also market. The problem with this argument is pay the tax, taxing only the nondefaulting that, absent any change in the students’ own students would make them pay for the sins “skin in the game,� they would confront the of the defaulting students. Thus, although a same commitment-to-repay problem. Thus, subsidized guarantee could be socially justi- unless the guarantee is fairly priced (so as �ed, the nondefaulting, tax-paying students to cover the expected loan losses and other (including those who would not borrow costs), the loan default losses would simply be without the guarantee) would prefer to go shifted to the state (the guarantor). But if the without it. guarantee is fairly priced, risk-neutral lenders Clearly, taxing only the defaulting students would not pay for it because, by de� nition, would lead to a Pareto improvement. But they care only about expected losses and not doing so amounts to assuming that one can about the variance of such losses. Unless the enforce taxation where one cannot enforce a state has an enforcement advantage vis-à- loan repayment. The optimality of the (sub- vis private lenders—which, as was already sidized) guarantee in a Mankiw-type student argued in this chapter, is hard to justify— loan model of adverse selection hinges exclu- there is no case for a state guarantee. sively, therefore, on a differential enforce- The discussion in this section can be sum- ment capacity. This does not make sense in a marized as follows. In a world devoid of risk political system where the rule of law applies aversion and collective action frictions, agency to states as well as to citizens. Any prefer- frictions alone do not in general justify guar- ential collection capacity states may have antees under a general equilibrium viewpoint should be made readily available to everyone that uses an appropriately restrictive welfare through improving the judiciary, as part of criterion. While the market outcome would a more supportive enabling environment. For be inefficient, a state that does not know 144 RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE more than, or enforce better than, the private There is no collective action failure. The dis- sector cannot improve the outcome via credit honest are simply getting away with mischief. guarantees. Indeed, one would generally Even if bargaining were costless, it would not expect the state to have a comparative dis- pay for the honest to buy out the dishonest. advantage in dealing with pure agency fric- Indeed, using the same reasoning as in the tions, rather than an advantage. If the state previous section, the honest would have to had a comparative advantage in this regard, make a transfer payment to the dishonest that the right policy would be to have only state- exactly matches the tax payments that would owned and state-run banks, which patently be required to cover a subsidized state guar- makes no sense.5 More generally, in a world antee or an interest rate subsidy. Similarly, where distortions arise only from agency even though it seems obvious that one should frictions, while the market equilibrium is lend to every student whose return exceeds inefficient, the state cannot improve on it the social cost of funds, a state banker with- by assuming risk, because there is no wedge out an informational or enforcement advan- between private and social interests—prin- tage should not lend and behave exactly like cipals and agents want the same thing that a private banker. society wants, namely, to overcome agency How would adding social externalities and frictions and engage in mutually bene�cial collective action frictions change this conclu- � nancial contracts. The only legitimate role sion? If one supposes that lending to some left for the state in such a world is to improve targeted students (say, the ones studying to the informational and enforcement environ- become primary school teachers) has positive ment so that markets can operate better. social externalities (for example, a good basic education enhances the earning potential from college education in all �elds of study), Adding collective action frictions the market outcome would be inef�cient even This section considers collective action fric- if private lenders could solve agency problems tions that manifest themselves in the form and properly identify all the creditworthy of social externalities—for example, positive students. Private lenders, by pricing all loans externalities to lending that are not internal- uniformly, would fail to lend sufficiently ized by the private lender—and continues to students planning to be primary school to assume that lenders are risk neutral. The teachers because their earnings prospects are literature generally concludes that, in the mediocre, even though those students can absence of information asymmetries, any contribute the most to other students’ earn- credit policy, including guarantees, is inef- ings. The private lender does not internalize fective in improving the equilibrium out- the externality. There is now a clear case of come unless subsidized. 6 Indeed, subsidies a collective action failure. Should students and taxes are generally shown to be the best of all generations and in all �elds of study be policy responses to a market failure arising able to get together and bargain at little or no from uninternalized externalities. However, cost, they would agree on setting aside part the literature concludes that it becomes sig- of the increase in their future earnings result- ni� cantly more dif� cult to design optimal ing from a better primary education in order subsidies where externalities and asymmet- to subsidize the interest rates on the loans to ric information coexist.7 future primary school teachers. One can see what is at stake by noticing As noted, an unsubsidized (fairly priced) first that, in the Mankiw (1986) model of state guarantee in these circumstances pure agency frictions, the dishonest inflict would have no value to a risk-neutral lender. negative informational externalities on the Instead, the state can resolve this externali- honest. However, barring differential taxa- ties-driven market failure by coordinating tion or enforcement capacity, there is no way agents through an interest rate subsidy pro- for the state to internalize such externalities. gram favoring loans to the would-be teachers RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE 145 and paid for by all other students. Of course, credit guarantees. When these frictions are a straightforward wage subsidy to primary relatively light, the state might limit its inter- school teachers funded by general taxation vention to that of a catalyst that brings would be even cleaner than a subsidized loan. together all interested parties and facilitates This illustrates that, as long as there is no the transfers across parties required for a risk aversion, collective action frictions alone mutually bene�cial equilibrium. When the establish the case for tax and subsidy policy frictions are harder to overcome, the state but not for state credit guarantees. can circumvent them through a targeted Another example is the case in which col- tax-subsidy program, which internalizes lective action frictions coexist with agency externalities. However, the implementation frictions while lenders are still risk neutral. of this program may run into agency fric- Where wage subsidies to school teachers are tions. Thus, depending on whether the state not an available option, an alternative could or the markets can better address these latter be a subsidized lending program run by pri- frictions, it might be optimal for the state to vate banks. Since informational frictions subsidize the loans provided by private lend- require that bankers screen potential borrow- ers or to provide the loans directly through ers and monitor their performance, and since a � rst-tier state bank. Remarkably, however, such efforts are costly, targeted interest rate the basic motivation underlying the state’s subsidies dominate targeted and subsidized intervention is always the need to address guarantees (recalling that, if the guarantee collective frictions, which introduce a wedge is fairly priced instead of being subsidized, between private and social interests that the risk-neutral lender would not pay for it). markets cannot resolve on their own. While both policy instruments can similarly expand the level of targeted lending, the interest rate subsidy is preferable because it Adding risk aversion does not distort the lender’s screening and Risk aversion among private lenders can monitoring incentives (the lender retains full now be added to the brew. Arrow and Lind’s “skin in the game�). (1970) article remains the most fundamental But there might also be cases where the and enduring conceptual framing for state risk state’s cost of monitoring whether private bearing. They show that, when risk is spread lenders appropriately screen loan applicants in small amounts over large numbers of inves- according to social criteria is greater than the tors, capital can be priced at risk-neutral cost of simply setting up a �rst-tier state bank prices. They argue that the state’s intertem- that directly provides the subsidized loans.8 poral tax and borrowing capacity gives it a In such cases, the assumption by the state unique ability to spread risk across large pop- of the risks associated with � nancial activi- ulations. Thus, state guarantees (as opposed ties can be justi�ed on the basis of the state’s to subsidies or loans) are naturally called capacity to address agency frictions (that is, for to reduce the cost of risk bearing and to ensuring that the loans are given to the most encourage private investment or lending in the socially desirable borrowers). However, it face of high risk or high risk aversion. is crucial to note that such agency frictions Curiously, the literature on partial credit arise out of an underlying collective action guarantees has mostly ignored the Arrow and failure that prevents markets from internal- Lind perspective. Moreover, in the scant lit- izing externalities. erature on this subject, a dominant theme is a The bottom line for this section is, there- rebuttal of the proposition that there is any- fore, as follows. When social externalities thing unique in the state’s capacity to spread and collective action frictions are added to risk. For example, Klein (1996) argues that if agency frictions in a world devoid of risk the state’s advantage did not lie purely in its aversion, the case for state intervention coercive taxation powers (that is, its capacity becomes clear, but it is hardly in the form of to oblige taxpayers to bear risk through the 146 RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE tax system), then markets would be able to appropriately, the guarantor can monitor spread risk just as ef�ciently. But as Arrow wholesalers and adjust the premium of the and Lind themselves suggest, it may not be guarantee according to how well the lenders possible for the private sector to be completely perform their monitoring. Of course, moni- risk neutral, even when risk is spread through toring the monitor has a cost. Second, the broad ownership. Since the controlling share- guarantor’s capacity to resell the risk to holders of a � rm need to hold large blocks retailers will itself depend on the guarantor’s of stock, and since such holdings are likely ability to convince them that he or she is to constitute a significant portion of their doing a good job at monitoring wholesal- wealth, the costs of risk bearing are not neg- ers and, hence, is offering retailers a fairly ligible, and the � rm should behave as a risk priced risk-sharing deal. averter. Thus, although Arrow and Lind hint This, in turn, will require resolving two at the existence of a link between risk aver- problems. First, retailers need to be able to sion and agency problems (adequate monitor- monitor the guarantor’s own monitoring ing is induced by large stake exposures), they efforts to resolve the moral hazard problem. do not develop the theory, nor has the litera- Therefore, the monitoring story repeats at ture picked up on that theme. this level. Second, even when this informa- Thus, to help analyze whether there is tion friction has been resolved (because indeed something unique about the state’s the information certifying the quality of risk-bearing capacity, risk aversion can be the guarantor’s monitoring is available introduced into the well-known monitoring to retailers at no cost), the guarantor also model of Calomiris and Kahn (1989).9 In this needs to convince retailers to underwrite model, an entrepreneur funds a risky project the guarantee. This will require solving through a mix of retail and wholesale fund- market participation frictions. In view of ing. Projects that are doomed to fail can be the atomicity of their investments, absent liquidated—thereby salvaging some of their a market architecture that allows for pool- value—if they are so identified at an early ing, retailers may not bother to participate. stage through monitoring. Retail investors The remainder of this section focuses on do not monitor because they have too small the � rst problem—that is, the guarantor’s a stake in the project relative to the cost of cost of monitoring the wholesalers—and monitoring. If wholesalers engage in moni- leaves aside the second problem—that is, toring, they can recoup their investments in it assumes that there is no cost to retailers failing projects. But wholesalers would do either in monitoring the guarantor or in so only if they have a suf�ciently large stake participating in underwriting the guaran- in the project (suf�cient “skin in the game�) tee. These latter costs will be reincorpo- to warrant incurring the monitoring costs. rated in the next section. However, because wholesalers are risk averse, In the absence of risk aversion, wholesal- having “skin in the game� inef�ciently raises ers do not need to be paid to bear risk, imply- the cost of funds. A guarantor buying the ing that it would not be socially costly for risk that is concentrated in wholesalers and them to retain “skin in the game� and that spreading it by reselling it in small amounts entrepreneurs would contract enough whole- to retailers can therefore improve the market sale funding to allow wholesalers to recoup equilibrium. fully the cost of the socially ef�cient level of In doing so, however, the guarantor monitoring. Monitoring costs can, therefore, faces two interrelated problems. First, he or be absorbed ef�ciently by risk-neutral whole- she undermines wholesalers’ incentives to salers, and risk would not need to be spread monitor the entrepreneur and the project. out. By contrast, when there is risk aversion This is the standard moral hazard prob- but no cost of monitoring the monitor, the lem faced in insurance markets. To induce risk borne by wholesalers needs to be spread wholesale lenders to monitor the borrowers out among retailers, but this can be done RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE 147 at no cost. Since the guarantor can monitor (cheap) monitoring by a more ef�cient, pyra- wholesalers at no cost, he or she can offer a midal market monitoring arrangement that full guarantee and monitor wholesalers suf- allows for full risk spreading. �ciently well to ensure that they continue to This section has not said anything, how- monitor the borrowers appropriately. At the ever, about the nature, state or private, of same time, since, by assumption, retailers the guarantee. As will now be shown, the can also monitor the guarantor at no cost, comparative advantage of state over private the guarantor can fully pass on (spread) the guarantees depends not on monitoring costs risk to retailers. Hence, the fact that moni- and agency frictions but, instead, on collec- toring costs disappear at the retailer level of tive action frictions (the traditional justi�ca- the monitoring pyramid allows the market tion for state goods) that may, under certain equilibrium to replicate the socially optimal environments, give state guarantees a de�nite solution. edge over private guarantees. In general, however, taking risk away from wholesalers is not cost-free, because it undermines the quality of their monitoring. The question of private or If such a cost is too high, a full guarantee is state guarantees no longer socially optimal. Hence, where the In the previous setting, the absence of moni- cost of ensuring good monitoring by risk- toring or participation costs allowed retail- averse wholesale lenders is high, these lend- ers both to ascertain that the risk-sharing ers would monitor borrowers appropriately deals they were offered by the guarantor only if the lenders retain a significant risk were fairly priced and to participate in the exposure, which implies that risk is no longer risk market, all free of costs. Thus, although fully spreadable. There is, therefore, a funda- the returns they would get from buying mental correspondence between the market’s were small, retailers were nonetheless will- capacity to spread risk and the decline of ing to underwrite the guarantor. With these monitoring costs as one goes up the monitor- assumptions, private guarantors should ing pyramid. naturally emerge, provided that they can The argument in this section can thus be mitigate at a reasonable cost the problem summarized as follows. Unless risk is prop- of monitoring the wholesale lender. Hence, erly spread out, risk aversion, combined with there would be no role for state guaran- agency frictions, introduces a deadweight tees. The only role for the state would be to cost that constitutes a source of market inef- strengthen the enabling environment so as to �ciency. A guarantee may, therefore, be justi- help alleviate the informational (or enforce- �ed as a means to lower the cost of capital ment) frictions that hinder risk spreading. by spreading risk more broadly. However, the However, this conclusion changes once bene�ts of expanded risk spreading should one relaxes the assumption that retailers can exceed the costs of the weaker screening, monitor the guarantor at no cost while still monitoring, and enforcement. For guarantees maintaining the assumption of zero participa- to be desirable, it is thus critically important tion costs. For retailers to ascertain whether that their adverse impact on incentives for guarantors are offering them a fair deal, they monitoring be limited. The latter depends need to certify the quality of the guarantors’ on the extent to which the price of the guar- monitoring, a costly effort. Because it is in the antee internalizes moral hazard, which is a guarantors’ own interest to have their moni- function of the quality of monitoring by the toring certi�ed (they will not be able to sell guarantor and, hence, of the cost of monitor- risk otherwise), and because they can include ing the monitor. When this cost is low, the the certification cost in the price of their market solution approximates the optimal guarantee, guarantors can pay someone (say, solution because it replaces a socially costly a rating agency) to do the certi�cation. How- “skin-in-the-game� requirement for good ever, this pushes the monitoring pyramid up 148 RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE one more layer, as retailers, in turn, need to costs of mobilizing participation will be high, be convinced that the rating agency has done so that adding those costs to the guarantor’s a good job certifying guarantors. If monitor- certi�cation costs might well raise total costs ing the rating agencies could be done without to a point where the guarantee is no longer cost, this would solve the problem. However, viable. this is unlikely to be the case. The potential Thus, the bottom line is that the abil- conflicts of interest between bond issuers and ity of the guarantor to unload risk will very rating agencies that have emerged at the heart much depend on how well-developed � nan- of the postcrisis debate on regulatory reform cial markets are. In mature �nancial systems, are testimony of the dif�culty of solving this the market can resolve participation frictions problem. Such difficulties carry a crucial through a deep network of asset managers. In implication, namely, that market monitoring an undeveloped system, such frictions remain issues (that is, the completion of the moni- unresolved. This is precisely the point at toring pyramid) are in general unlikely to which the state can help to complete markets, be solved without the provision of a public at least temporarily. Because state guarantors good in the form of of�cial oversight over the do not have to market their risk (the distribu- market monitors. Of�cial provision of infor- tion is taken care of through the tax system mation and of�cial oversight thus act as the and state policy), they can effectively resolve last lines of defense to help close the monitor- the collective action failure. ing pyramid. State guarantees can spread the risk all Be that as it may, the key point is that these the more �nely because they can do so across monitoring problems do not depend on the currently living taxpayers and across gen- nature, public or private, of the guarantees. erations within a given jurisdiction. How- Indeed, with one possible caveat (more on ever, even in the case of intergenerational this below), even if retailers faced no costs risk spreading, the state’s advantage derives in participating in underwriting a guaran- again from its capacity to address a collective tee, a state guarantor will face exactly the action (participation) friction, rather than an same monitoring problems and constraints enforcement (agency) friction. Indeed, trying as a private guarantor in terms of convinc- to depict the inability of markets to contract ing retailers that the wholesale lenders will across generations from a pure enforcement be under adequate vigilance. Thus, a state perspective is rather futile. Since it is not pos- guarantee continues to be unjusti�ed. sible to write bilateral contracts with some- Things change further once one relaxes one unborn, “enforcing� such contracts is the last simplifying assumption—the absence meaningless. Instead, the state has an edge of participation costs. Although having retail- because of the intergenerational burden ers participate in the risk market to spread sharing that the political system is naturally risk ef�ciently would be socially desirable, designed to do.10 individual retailers may not internalize this Based on the above conceptual premises, positive participation externality. Instead, the conditions can now be stated under which the participation disincentives faced by retail- public guarantees will be needed either to ers may require the guarantor to incur costs replace or to complement private guarantees (for example, in advertising) to attract inves- in the face of idiosyncratic risk. The main tors. In a well-developed �nancial system, the conclusion is that, where private lenders are guarantor would not have to deal directly risk averse, and even where risks are idiosyn- with retailers. Instead, an additional layer of cratic and therefore diversi�able in principle, agents—the asset managers, such as mutual the state can spread the risk more broadly funds, pension funds, or hedge funds— than the market by resolving participation could pool retail investors for the guarantor, externalities—that is, the state can pool ato- thereby reducing the costs of participation. mistic investors (or taxpayers) that would But in an undeveloped � nancial system, the otherwise not participate in underwriting RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE 149 the guarantee. However, for state guaran- development. That is the case, for example, tees to be desirable, it is critically important when the apparently successful market-led that their adverse impact on monitoring and reduction of agency and participation fric- enforcement incentives be limited. tions may itself introduce—as a result of, There is a clear infant industry argument say, interconnectedness—widening wedges for public guarantees, in the sense that they between private and social interests. The can be justi�ed on a transitory basis when key point is that public guarantees can be � nancial systems are underdeveloped—and justi�ed on a permanent basis in the pres- therefore suffer from high monitoring costs ence of systemic risk, including that which and low participation—but only so long as is endogenously brewed in the process of they aim at crowding in (rather than crowd- � nancial development itself. The rationale in ing out) the private sector. Moreover, if idio- this case, however, no longer derives from syncratic risk is fat-tailed, public guarantees the need to spread risk as broadly as pos- may be justi�ed on a more permanent basis, sible but rather from the state’s capacity to because even the more developed � nancial help coordinate agents’ actions around an markets may not be able to reach the scale ef�cient risk-sharing equilibrium. of participation that would be needed to The argument in support of public guar- atomize and distribute the risk suf�ciently.11 antees under systemic risk has two threads, While more mature systems should enhance depending on whether systemic risk is a the scope for private guarantees by smooth- manifestation of highly uncertain and endog- ing out agency frictions (by helping to close enous risk or a manifestation of correlated the monitoring pyramid) and collective risk. First, in the case of highly uncertain action frictions (by pooling risk through and endogenous risk, the public guarantee asset managers), the spreading of fat-tailed may induce agents to abandon altogether risk may exceed the capacity of private the expected utility-maximizing framework guarantees. Again, permanent public guar- and choose instead a “min-max� criterion antees in mature � nancial systems may be such that they minimize their exposure to the justi�ed so long as they do not unduly erode maximum possible loss. In such a case, Arrow the monitoring incentives of those protected and Lind’s (1970) risk-spreading argument by the guarantees. no longer holds, because the total cost of risk One should note that, in all cases, the bearing remains the same as the population risk-spreading ability of the state and, hence, of taxpayers becomes large, making risk non- the rationale for public guarantees ulti- diversi�able. However, public guarantees are mately rests on the comparative advantage still useful because they help resolve collective of the state in resolving collective action fric- action failures. When all agents minimize tions (the traditional justi� cation for pub- their exposure to a worst-case scenario, they lic goods).12 Public guarantees may have an can end up behaving in a collectively irratio- edge over private guarantees not because the nal way because each individual acts on the state can better resolve agency frictions but, basis of being affected more than the aver- instead, because collective action frictions age. By eliminating such a scenario, public may disable the market’s ability to resolve guarantees effectively function as a coordina- such agency frictions. tion device, much as deposit guarantees and lender-of-last-resort facilities can eliminate self-ful�lling bank runs.13 Systemic risk Correlated risk leads to a further break- Finally, this section considers the case in down of the Arrow and Lind criterion. Even which the risks in question are systemic, if risk exposure is atomized and shared across that is, they are not diversifiable. As dis- a large number of investors, these investors cussed in chapter 2, systemic risk can arise remain risk averse because the returns on endogenously in the process of financial their investment become negatively correlated 150 RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE with their income and consumption. Because and state guarantees. Because it is complex correlated risk applies to an investor decid- and potentially confusing, this interaction ing whether to invest in a private guarantee between collective and agency frictions may scheme as much as to a taxpayer deciding have helped generate the large policy swings whether to vote for a state guarantee observed in LAC’s recent history. scheme, the state has no natural risk aversion advantage in this case. Nonetheless, public guarantees can still improve things by helping LAC’s policy swings—a play in avert the collective action failures that mag- four acts nify the impact of a systemic event. By coordi- Although the importance of state banks in nating agents’ behavior around a collectively LAC fell steeply over the past 40 years, they desirable outcome, public guarantees help remain significant in the region, account- reduce the risk of catastrophic downturns, ing for over 20 percent of banking system thereby smoothing out private consumption, assets (see table 9.1). Moreover, development which, in turn, helps reduce risk aversion. banks, a subset of state banks, continue to The justification for the state to assume be the instrument of choice for states to pur- �nancial risk is always derived from its abil- sue � nancial inclusion and countercyclical ity to resolve collective action frictions bet- credit objectives.14 LAC’s recent history of ter than markets, which is, indeed, the usual wide swings in the perceived developmental justi�cation for public goods. This suggests role of the state in � nance can be depicted as a natural division of tasks, whereby markets a play in four acts.15 take the agency roles but are supported by states wherever markets are unable to cor- Act I: The state can do it better rect the collective failures. States and mar- kets should thus naturally complement each A highly interventionist role for the state other. In practice, however, this neat division emerged in LAC in the 1950s, dominating of labor is complicated by the fact that the � nancial development policy until the mid- state’s role is often justi�ed at the intersection to late 1970s. The premise was that the between the agency and collective paradigms. market, left to its own devices, would be State risk taking may be needed to address incapable of internalizing growth externali- agency failures that prevent markets from ties. Thus, development banks became the resolving collective frictions, as in the case of key policy vehicle to mobilize and allocate �rst-tier state banking reviewed in the second financial resources where markets failed, section, or, conversely, to address collective particularly in long-term � nance, credit to failures that prevent markets from resolv- small and medium enterprises (SMEs), hous- ing agency frictions, as in the case of risk ing � nance, and agricultural credit. While spreading surveyed in the previous section, this rationale rightly put the emphasis on the which can justify second-tier state banking state’s comparative advantage in resolving TABLE 9.1 Development banks in Latin America and the Caribbean: Operative modality of development banks at the beginning of 2009 Development banks Operative modality Number of % Assets (US$) % First-tier banks 66 65.3 376,365 43.5 Second-tier banks 23 22.8 237,762 27.5 Banks of 1st and 2nd tier 12 11.9 251,043 29.0 Total 101 100.0 865,169 100.0 Source: ALIDE database. Note: The database identi�es a total of 101 development-oriented banks in 21 countries in LAC. RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE 151 collective action frictions, in practice, the and more lucrative lines of business, thereby state became directly involved in solving putting them on a potential collision course agency frictions and in this capacity became with private banks. 20 a substitute for (rather than a complement to) � nancial markets. Thus, it is no surprise Act III. Let markets and the state that Act I ended badly, with deeply atro- cooperate phied financial systems and mismanaged state banks that turned into instruments of In the past decade or so, many development political clientelism saddled with yawning banks in LAC have moved away from risk losses.16 taking, into nontraditional catalytic activi- ties aimed at resolving collective action failures. These efforts began to materialize Act II. On second thought, markets in innovative interventions, where develop- can do it better ment banks increasingly combined activi- In reaction to these problems, the conven- ties aimed at solving coordination problems tional wisdom on the developmental role of with activities entailing risk absorption the state in � nance took a 180-degree turn (particularly via credit guarantees). Such toward a frankly laissez-faire view. This “pro- market activism,� analyzed in some shift—a reflection of shifting intellectual detail for Chile and Mexico by de la Torre, winds worldwide—was dramatically stimu- Gozzi, and Schmukler (2007c), included lated in LAC by the debt and �scal crises of (a) the creation of infrastructures to pro- the 1980s.17 The key contention was that, mote participation and to help financial precisely due to agency problems, bureau- intermediaries achieve economies of scale crats can never be good bankers and can in and reduce the costs of � nancial services; 21 fact do more harm than good. Moreover, (b) investment bank–type activities cen- given well-defined property rights, good tered on coordinating various stakeholders contractual institutions, and information around structured � nance schemes, and on availability, markets can take care of most enhancing such structures with state guar- agency failures by themselves.18 Therefore, antees; 22 and (c) partial credit guarantee state efforts should be primarily deployed schemes to promote private sector lending toward improving the enabling environ- to priority sectors.23 ment.19 In response, many development banks redefined their mandates, shifting Act IV. When the going gets tough from first tier to second tier. At the same time, to level the playing �eld, states placed During the recent global financial crisis, the remaining � rst-tier development banks the issue of risk absorption by develop- under the same standards of prudential ment banks came forcibly back to the policy oversight and corporate governance that debate, because, in many LAC countries, apply to private banks and made efforts to development banks performed vigorous enhance the banks’ managerial quality and countercyclical credit activities aimed at sus- independence. However, this action put the taining the flow of credit in the face of rising banks in a dif� cult bind, still at the heart risk aversion among private banks. There is of the challenges they face today. On the by now a strong consensus that these coun- one hand, they are unavoidably compelled tercyclical interventions—which involved to ful� ll an (explicit or implicit) policy man- first- and second-tier lending as well as date that, almost by definition, exposes credit guarantees—were successful over- them to high-risk clientele and often limits all. 24 However, this has reopened conten- their capacity to diversify risks. On the other tious issues that seemed to have been �nally hand, the new and stricter requirements to settled. Development banks are now asking avoid losses induces them to seek less risky themselves whether they should grow bigger 152 RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE even in the good times so as to play a more States can promote private sector participa- forceful role in the bad times. There is thus a tion in guarantee schemes, such as mutual risk of another swing in the pendulum. guarantee associations funded by small local entrepreneurs, or guarantee schemes struc- tured as joint stock companies with private Toward a rebalanced policy participation. The experience across the The main message from the conceptual anal- world with such schemes has been generally ysis in the � rst �ve sections of this chapter is positive, partly because they promote peer that the state should play to its strengths— pressure, a purely private form of resolving helping resolve collective action failures— collective frictions. Indeed, there is some evi- rather than to its weaknesses—dealing with dence that such associations work best when agency frictions. This message implies that they remain purely private, as this fully pre- the state should seek to complement (rather serves incentives for group monitoring and than substitute for) markets, crowding in the limits moral hazard.25 private sector and harnessing its comparative Peer pressure may not work in all cases advantage in dealing with agency problems, and all environments, however. Thus, rather than crowding it out. the third avenue, more controversial and This approach opens three avenues to thorny, involves risk absorption and risk explore. The first, the least controversial, spreading through second-tier state banks, comprises policy interventions exclusively whether through guarantees or long-term geared to solving participation frictions— loans. 26 In either case, it is the interac- along both the intensive margin (the same tion of risk aversion, agency (monitoring) players engaged in more transactions) and the frictions, and participation frictions that extensive margin (the incorporation of new justi�es state insurance, even where risk is players)—without dealing directly with risk. idiosyncratic. As noted, state guarantees Rising �nancial inclusion makes it easier for may be temporary—a strictly developmental the �nancial services industry to lower costs, tool aimed at facilitating risk discovery and expand market liquidity, and diversify risk, circumventing transitional collective action as well as for society to share in the bene�ts frictions—or permanent (in the face of fat- that come from the positive spillovers asso- tailed risk). However, state-sponsored credit ciated with scale and network effects. This guarantee programs do not seem to be typi- observation justi�es the state’s catalytic roles cally, at �rst glance, fully consistent with such portrayed in the previous section. But it can � rst principles. They seem to be permanent also explain the establishment and operation rather than transitory, and they tend to target by central banks of large-value payments sys- well-de�ned, recurrent, limited risks instead tems; the promotion by the public sector of of insuf�ciently understood risks or tail risks clearing, settlement, and trading infrastruc- where the state’s comparative advantage in tures; or the standardization of contracts. risk bearing and spreading could be more It can also justify mandated—or gently fully exploited. Moreover, these guarantee coerced—participation, as in the case of programs are typically justified based on contributions to privately administered pen- asymmetric information, lack of collateral, sion funds or the payment of state employee or externalities instead of risk aversion. What wages through accounts in banks that par- explains these apparent disconnects? ticipate in a shared, open-architecture plat- Political economy offers one explanation. form for retail payments. Given the presence Correcting agency problems that hinder, for of positive externalities, the state can also instance, SME lending (presumably ripe with use well-targeted subsidies as part of these positive social externalities) and lending to interventions. the lower-income households (presumably The second avenue is still catalytic in ripe with equity benefits if not externali- essence but deals directly with risk spreading. ties) sells better in the polity than correcting RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE 153 problems of risk spreading and differential But the tendency of development banks risk aversion. In addition, it is not popu- not to move too aggressively toward the lar for the state to take on risk from private risk frontier reflects not just the political banks, even when doing so is fairly priced, economy. It also reflects legitimate dif�cul- because it smacks of being a bailout. More- ties, particularly with the accurate pricing of over, if the problem is risk aversion among guarantees. As noted, public guarantees are private lenders, where should a risk-neutral welfare-enhancing only as long as and as far state draw the line? Should the state sector as all the risks and incentive distortions they reinsure or guarantee all productive lend- cause have been properly recognized and ing at the risk frontier, not just � nancing to priced in. But risks at the frontier are not well SMEs? Should state guarantees apply to all understood, and they may be abnormally dis- long-maturity loans, irrespective of �rm size? tributed, with fatter tails. Under such circum- Expanding the risk frontier across the board stances, the pricing of guarantees becomes is naturally unpalatable to politicians insofar inherently difficult, as the estimation of as they are held accountable. Indeed, over the expected losses and the decomposing of risk years, parliaments in LAC and many other premiums are subject to much error. Indeed, regions have strictly limited risk taking by there are chilling cases of major failures and state banks. Moreover, state bank managers losses in ambitious state insurance programs are naturally risk averse. They protect their that have aggressively aimed at crowding in capital because they know that they will live the private sector toward the risk frontier. In or die with it. 27 And as already noted, the part, such failures have reflected a misjudg- constraints that development banks face in ment of expected losses. terms of avoiding losses often induce them to Where does all of this lead? First, it is time compete with commercial banks to reap high for state lending and insurance programs to returns for low risks, rather than—as the risk come clean as regards their rationale and the aversion rationale would suggest—to comple- minimal conditions for success. Instead of ment private activity by insuring risk taking justifying state loan and guarantee programs at the frontier. based on goals, as is so often the case and Instead, public guarantees to SMEs or with which it is so hard to disagree, policy to target clienteles, such as those reached makers need to focus instead on alternative through low-income housing or student loan means of achieving these goals. Thus, pub- programs, look like safe bets when they are lic risk-bearing programs should be justi�ed well within the risk frontier. They appear to by comparing their costs and bene�ts with pay for themselves (hence are � scally safe) those of alternative channels of state inter- when well priced and designed.28 Moreover, vention that do not involve any state risk tak- early research seems to indicate that par- ing. Where guarantee programs are deemed tial credit guarantees supplied by states do appropriate, the programs’ objectives, man- provide at least some additionality. 29 Why, dates, and reporting and disclosure require- then, not safely collect the low-hanging fruit ments need to be refocused around a risk instead of shooting for the moon? This pref- aversion rationale and more clearly linked to erence for seemingly low-risk–high-political- the agency or collective action frictions with return programs has been accentuated over which risk aversion interacts. Policy makers the past decade or two in LAC and other also need to explain why the state can achieve regions, especially in the context of medio- that which markets cannot. Either these pro- cre growth, high structural unemployment, grams are really self-sustaining and therefore and tight �scal constraints. Lending or credit should be eventually divested to the private guarantee programs directed to those most sector, or there are hidden risks (fat-tailed affected by economic hardship have provided or systemic) that free markets cannot handle a convenient safety valve to relieve some of well and that need to be explicitly recognized the buildup in political pressure.30 and accounted for. Unless this is done right, 154 RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE state guarantees will likely end up subsidizing capital for speci�c insurance or countercyclical risks, and this is bound to distort incentives risk absorption can help state banks assume and trigger unpleasant fiscal surprises (as more risk in a responsible, bounded manner well as political upheavals) once downsides while protecting their capital from deple- materialize (the recent U.S. experience in the tion. Alternatively, to align incentives, state subprime crisis is, of course, the most obvi- banks can assume a limited part of the risk, ous illustration). the rest being covered by the state through It is thus not enough that public guarantee earmarked capital or other means. Private- programs break even in good times. If priced state partnerships in which the state assumes right, they should accumulate reserves in the most (but not all) of the risk at a fair price good times against potential losses in the bad may help facilitate price and risk discovery. times. For tail risk, this implies charging for Enhanced transparency, better measurement the full expected value of the tail losses. One of risk and returns, and more sophisticated possible approach to facilitate risk discov- checks and balances (for example, through ery is to auction the guarantees according to recurrent assessments by independent evalu- their coverage or price.31 By setting volumes ation units or through occasional, more- rather than prices, guarantors can better strategic reviews by blue-ribbon committees) protect themselves against the risk of major also should all help strengthen the governance mispricing. At the same time, volumes may be of state banks or other state entities engaging adjusted to meet countercyclical objectives. in higher-risk activities. They can be raised in systemic downturns, As already noted, there might also be cases when upward jumps in private risk aversion where, to internalize externalities, states may are more likely to trigger coordination fail- be better off dealing directly with borrow- ures, and reduced in upturns, when risk ers through a first-tier state bank (that is, appetite can swell, fueling excessive credit becoming agents) than through guarantees expansion. To avoid head-to-head compe- provided by a second-tier state bank. In tition with the private sector in providing particular, it has been argued that, in down- primary insurance, the state should prefer turns, partial guarantees may not be suf�- to provide its support through well-targeted cient to overcome bankers’ heightened risk reinsurance against tail risks. aversion. Thus, unless states are willing to As � nancial systems mature—and hence assume most or all of the risk, which could more information exists to assess and price subject them to unacceptably high losses, risk—public loan and guarantee programs guarantee programs may fail to provide an can be better targeted. For example, when effective countercyclical tool. Instead, as the risk scoring methods are available, fairly argument goes, � rst-tier state banking pro- priced public guarantees can be targeted to vides the only reliable channel to increase be at, or just outside, the risk frontier, so as lending in a reasonably safe fashion, that is, to ensure that finance reaches clients and “when the going gets tough, only state banks projects that are too risky for private institu- get going.� As reasonable as this argument tions to lend to without guarantees. When may sound, it can also be turned on its head. loans are made directly by first-tier public The more state banks compete with private banks, making sure the interest rates on the banks, the less private lenders are likely to loans are above market rates can help ensure share information with a state guarantor. that public risk bearing does not crowd out Hence, a noncompeting state sector, one that private risk bearing.32 complements but does not substitute, may in The further one seeks to move away from fact be best able to maintain open access to the private risk frontier, the more caution is the private information that it needs to moni- of course called for. However, risk taking can tor the banks and extend the coverage of the be bounded and state governance protected guarantees during downturns in a fiscally in a variety of ways. For example, earmarked responsible way. RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE 155 An important final question is whether tors to �nance investment projects may enjoy state banks should be supervised as private informational advantages vis-à-vis private banks. In the case of first-tier state banks, lenders, such as knowing more than private the answer is an unquali�ed yes. Since state lenders about state processes and procedures. This may justify MDB guarantees even in a and private banks compete for the same world characterized by pure agency failures business (they are potential substitutes, at with no risk aversion. least to a degree), they should be regulated 6. See, for example, Raith, Staak, and Starke and supervised in exactly the same way as (2006); Penner and Silber (1973); and Lombra private banks. For second-tier state banks, and Wasylenko (1984). the answer is not so clear, however. Because 7. See Stiglitz, Vallejo, and Park (1993); such banks are in the business of ensuring Calomiris and Himmelberg (1994); and against tail risks, the tolerance range involved Bhattacharya (1997). in calibrating their capital under a value-at- 8. The argument that the state may be able to risk criterion will need, by construction, to provide incentives to public lenders more use less strict criterion than that applied to easily than to private ones is in line with Holmstrom and Milgrom’s (1991) result that private banks. At the same time, the empha- increasing the incentives along a measurable sis on uncertain frontier risk requires a dif- performance dimension (costs or pro�tabil- ferent type of supervision. It should rely on ity) reduces the incentives along nonmeasur- high-end, holistic assessments by panels of able dimensions. For a fuller discussion along experts that balance the economywide costs these lines of the role of �rst-tier public banks, and bene�ts of the programs that the state see also IDB (2005). bank engages in rather than ready-made, one- 9. See Huang and Ratnovsky (2011) for an inter- size-�ts-all rules aimed at ensuring �nancial esting extension of the Calomiris-Kahn model stability. challenging the proposition that wholesale investors ef�ciently discipline the bank (or entrepreneur) to whom they lend. Notes 10. Enforcement frictions (the other important type of agency frictions) may also help justify 1. This chapter draws heavily on the paper the need for public guarantees but, as with “Risk-Absorption by the State: When Is It informational frictions, not because the state Good Public Policy?� by Deniz Anginer, has any natural advantage in enforcing con- Augusto de la Torre, and Alain Ize (2011), tracts. As already noted in the previous sec- which is part of the Edited Volume that tion on collective action frictions, there is no accompanies this LAC flagship report. credible reason why enforcement failures can 2. First-tier public banks are those that con- be resolved through the state’s uncontested duct their operations directly with their �nal ability to tax but not through well-formu- customers; instead, second-tier public banks lated private contracts and a well-functioning conduct their operations through commercial judiciary. Instead, the state may have an edge banks. because costly contract enforcement is likely 3. This chapter presents the arguments infor- to require cost sharing, which again faces a mally. A more formal discussion can be found collective action problem. In turn, taxation in Anginer, de la Torre, and Ize (2011). should not be viewed (as in Klein 1996) as a 4. Likewise, overlending can occur. For example, mechanism to oblige unwilling taxpayers to when projects that would be equally pro�t- share risks, but instead as a simple, built-in able if successful have different probabilities coordination mechanism that facilitates the of success, low interest rates can induce bor- participation of all. rowers with low success probabilities to bor- 11. In principle, the Arrow and Lind (1970) row, even though their expected returns are argument continues to apply: no matter how below the social rate of return. See de Meza lumpy the risk, it can still be distributed ato- and Webb (1987, 1999) and Beck and de la mistically, provided there are enough retailers Torre (2006). over which the risk can be spread. In Arrow 5. Notice, however, that multilateral develop- and Lind, the number of retailers can go all ment banks (MDBs) that lend to public sec- the way to in�nity. In practice, however, there 156 RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE is an important difference between a large and Schmukler (2006, 2007c). For the corre- number and an in�nite number. Moreover, sponding evolution of the broader economic and perhaps more important, participation development paradigms in LAC, see Birdsall, frictions limit market depth even in well- de la Torre, and Valencia (2010). developed �nancial systems. Thus, the num- 16. Empirical studies have tended to �nd that ber of retailers over which risk can be spread, public banks have done more harm than even if large, may not be suf�cient. That is good (see Barth, Caprio, and Levine 2001; why there may be a point at which a perma- Caprio and Honohan 2001; IDB 2005; La nent public guarantee may be needed, even in Porta, Lopez-de-Silanes, and Shleifer 2002). mature systems, to bound the risk associated In particular, these studies �nd that greater with unpredictable returns or where there is state participation in bank ownership is asso- some probability, even if very small, of very ciated with lower levels of �nancial develop- large losses. Knightian uncertainty—when ment, less credit to the private sector, wider decision makers cannot determine the proba- intermediation spreads, greater credit con- bilities of events (see Epstein 1999)—is likely centration, slower economic growth, and to have an effect similar to fat tails. The more recurrent �scal drains. However, IDB (2005) uncertain the risk, the more �nely it needs �nds that public banks tend to play a coun- to be distributed, which, in principle, makes tercyclical credit role. For evidence on the more of a case for public guarantees. history of recurring losses of public banks in 12. There may be some exceptions in which the Brazil, see Micco and Panizza (2005), and for public sector may also have a genuine advan- Mexico, see Brizzi (2001). Goldsmith (1969), tage in dealing with agency frictions. For McKinnon (1973), and Shaw (1973) led the example, lenders and entrepreneurs may deal charge in arguing that dirigisme in �nance with each other on the basis of proprietary leads to “�nancial repression.� A number information and relationship lending. Private of cross-country studies try to measure the guarantors need to access this information to impact of �nancial repression (as measured understand the risks. Yet (at least one of) the by real interest rates or a variant thereof) parties involved in these private deals may on growth (see, for example, Easterly 1993; be reluctant to share the information since Lanyi and Saracoglu 1983; Roubini and Sala- they might lose proprietary rents when the i-Martin 1992; and World Bank 1989). information is leaked. In that case, the pub- 17. The laissez-faire view was also consistent lic guarantor with preferential access to pri- with the application in LAC of Washington vate information may reduce its monitoring consensus–style reforms during the 1990s (see costs, giving it an edge. However, one would Birdsall, de la Torre, and Valencia 2010). expect such advantages to vanish as �nancial 18. See Caprio and Honohan (2001); Klapper systems mature and information becomes and Zaidi (2005); Rajan and Zingales (2001); increasingly public. World Bank (2005). 13. See Diamond and Dybvig (1983) and the 19. This paradigmatic shift also put emphasis on more novel contributions of Caballero and �nancial liberalization as a way to achieve ef�- Krishnamurthy (2008) and Caballero and ciency. Thus, by the late 1990s, Latin America Kurlat (2009) on the role of public guaran- reached levels of �nancial liberalization com- tees under uncertainty. More generally, one parable to those in the developed world. See could also argue that the state could behave de la Torre and Schmukler (2007). collectively in a more rational way than indi- 20. In many cases, the resulting tension was exac- viduals when the latter are subjected to sys- erbated by the erosion of the cost-of-funds tematic behavioral biases. advantage that state-owned banks used to 14. The Latin American Association of enjoy, reflecting �nancial globalization and Development Finance Institutions (ALIDE) the rising availability of funds. identi�es a total of 101 development-oriented 21. The interventions examined there include the banks in 21 countries in LAC in 2009 (see following. First is the case of the Internet-based table 9.1). market platform for factoring services created 15. For a more detailed discussion of the evolu- by the Mexican development bank NAFIN. In tion of the views on the role of the state in this platform, SMEs are able to discount com- �nancial development, see de la Torre, Gozzi, petitively the accounts receivable claims they RISK BEARING BY THE STATE: A COLLEC TIVE AC TION PERSPEC TIVE 157 have on large buyers. Second is the electronic and Mistrulli (2009). Lebanon provides an platform implemented by BANSEFI, another interesting example of a seemingly successful Mexican development bank, to help small sav- and pro�table guarantee scheme structured ings and credit institutions (the so-called cajas) as a joint stock company. reduce their operating costs by centralizing 26. A long-term �nance commitment can be back-of�ce operations. Third is the experience viewed as a funding (liability) guarantee that with correspondent banking in Brazil, where provides protection against liquidity risk and a non�nancial public infrastructure with a price volatility, instead of credit default. Thus, large geographic coverage, the post of�ce, was instead of the development bank actually made available to �nancial intermediaries for funding the commercial bank, an equivalent the distribution of certain �nancial services. arrangement would be for the development 22. The speci�c interventions examined in this bank to provide swap and lender-of-last- respect include structured �nance products resort facilities. created by FIRA, a Mexican development 27. Development banks in Mexico, for instance, bank oriented toward agriculture. In one are regulated and supervised on par with such scheme, FIRA set up a structure that commercial banks and are required by law to ensures the provision of adequate working preserve the real value of their capital. capital �nancing for shrimp producers, which 28. See Honohan (2008) and Beck, Klapper, and required the coordinated participation—via Mendoza (2010). a trust fund—of, in addition to shrimp pro- 29. See Larraín and Quiroz (2006); Haines, ducers, the large (international) shrimp dis- Madill, and Riding (2007); and Arvai, Rocha, tributor, commercial banks, and shrimp feed and Saadani (2011). suppliers. FIRA enhanced the structure by 30. See Rajan (2010). providing second-loss default guarantees to 31. This is the approach followed in Chile by commercial banks. FOGAPE. See Benavente, Galetovic, and 23. The interventions analyzed in this regard Sanhueza (2006) and de la Torre, Gozzi, and include the FOGAPE guarantee system in Schmukler (2007c). Chile, which is funded by the state and 32. Some lending and guarantee programs by administered by BancoEstado, a public bank, development banks in high-income coun- and supports lending to small enterprises. tries are structured in this way. The Business 24. See, for instance, ALIDE (2009). Development Bank Canada (BDC) small 25. On the experience of mutual guarantee associ- business loan guarantee program is perhaps a ations in Europe, see Columba, Gambacorta, prime example. See Rudolph (2009). Prudential Oversight: Where Does LAC Stand? 10 T he Latin America and Caribbean oversight as de� ned by the precrisis Basel (LAC) region has weathered the standards, including the institutional and global � nancial crisis fairly well, in accounting and disclosure frameworks, part because of the progress achieved in pru- various aspects of regulation (entry, capi- dential oversight over the past two decades. tal adequacy, credit risk, and other risks), This progress came in the wake of LAC’s and the main components of supervision turbulent macrofinancial history, which (methods, internal controls, and consoli- stimulated substantial efforts to overhaul dated and cross-border oversight). and tighten its prudential oversight. Some- • However, progress has been uneven, and what ironically, this progress was occurring important gaps remain, both across sub- just as � nancial regulators in many devel- regions and across themes. oped countries were bent on easing inter- • The LAC-6 countries (the LAC-7 group mediation through more market-friendly minus Argentina, for which a BCP assess- regimes with less expensive buffers. LAC’s ment was not yet � nalized at the time of efforts were often spearheaded by repeated this writing) were found to outperform assessments of international standards, their peers generally; the smaller, lower– particularly the Basel Core Principles for income LAC countries generally under- Effective Banking Supervision (BCP). This performed; the Caribbean countries lie chapter � rst reviews the progress and iden- somewhere in between. ti� es some of the remaining gaps.1 It then • Two basic legal framework issues—the sets the background for discussing (in the independence of bank supervisors and � nal three chapters of this report) the chal- their legal protection—need special atten- lenges of systemic oversight, issues that were tion in many LAC countries. mostly ignored in the regulatory standards • The regulation of capital adequacy is written before the global � nancial crisis, but uneven, with over half of LAC countries for which LAC needs to prepare. Main � nd- not fully meeting Basel I standards and ings are as follows: only limited progress being made toward meeting Basel II. • Important progress has been achieved by • Bank supervision lags outside the LAC-6 LAC countries in all main areas of bank countries, which illustrates that good 159 160 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? implementation is often the most elusive two dimensions. On the one hand, the rat- component of oversight. ings obtained by LAC countries in their BCP • Issues of consolidated supervision (both assessments are compared. 2 LAC is divided across members of the same financial into three subregions (LAC-6, Caribbean, group and across borders) continue to and other countries) and the 29 BCP ratings loom large; although LAC does not lag its in each BCP assessment are grouped along peers (these issue are truly complex and 10 basic areas of oversight. This provides a tough nuts to crack), the combination of picture of the relative effectiveness of oversight concentrated domestic ownership and when one simply compares across areas. large foreign ownership of � nancial con- On the other hand, an econometric analy- glomerates makes this issue particularly sis is conduced to benchmark the three subre- critical for the region. gions across the world, using the full sample • While the progress made in traditional of 149 BCP assessments that have been oversight will help set the bases for conducted to date (results are synthesized addressing the new challenges of systemic by subregion in table 10.1 and �gures 10.1, oversight, LAC (like everyone else, for 10.2, and 10.3; the methodology is explained that matter) faces a dif�cult and complex in annex 10.A). This analysis of BCP ratings agenda of reforms, with little time to spare is completed with an econometric analysis as the risks rapidly build up. of responses to a 2007 World Bank survey • That being said, LAC’s turbulent past of supervisors that covers a broad range of (which has led to prudential buffers that supervisory practices (annex 10.B). In both are relatively large by international com- cases, controls for economic development parison), an arguably greater willingness (that is, for gross domestic product [GDP] per among institutions to listen to their super- capita, a key underlying determinant of the visors, and some hands-on experience quality of oversight) are introduced. Thus, with simple macroprudential tools and the the basic question this exercise attempts to management of systemic risk have given answer is how the three LAC subregions have LAC a leg up, which the region should performed in relation to their peers at similar exploit to the maximum. levels of economic development. That subre- • While this analysis is backed by a vari- gions may not be performing well in some ety of sources and assessments (not just areas in absolute terms is to be expected, as the BCPs), strong caveats are nonetheless these are tough issues for everybody. In other clearly in order, particularly since many of areas, however, the LAC subregions not only the underlying BCP assessments are dated. may be performing poorly in absolute terms but also may be lagging their peers. These The rest of the chapter is structured as fol- areas clearly require more attention. lows. The � rst section describes the battery The assessments reveal weaknesses that are of sources and methods used for this assess- (or were) common to many LAC countries. ment. The second section reviews the prog- Wide disparities exist across both principles ress and gaps in traditional (that is, before and countries, however. As regards principles, the global �nancial crisis) oversight. The �nal some are clearly easier to meet than others, section briefly describes the challenges ahead with the trickiest principles being met (fully and possible handicaps in developing a sys- or largely) by only about 30 percent of coun- temic oversight capacity. tries (�gure 10.4, panel a). As regards coun- tries, the top three LAC countries are close to being fully compliant as regards their aver- Methodology age rating, while the bottom three are close The analysis in this chapter is based on quan- to being on average materially noncompliant titative as well as qualitative evidence. On the (�gure 10.4, panel b). Not surprisingly, there quantitative side, the analysis proceeds along have been very substantial improvements in TABLE 10.1 Econometric analysis of BCP ratings Institutional Internal Market Corrective Consolidated   framework Authorizations Capital Credit risk Other risks control Supervision discipline actions supervision LAC-6 –0.397 –5.838*** 1.702 –4.067* –5.988* –3.999 –1.781 0.523 1.339 –5.459** (0.814) (0.000) (0.431) (0.081) (0.085) (0.376) (0.426) (0.762) (0.546) (0.020) LAC-6 – Update 1.810 5.009*** –6.006 –1.619 –4.543 1.744 –3.396 –7.625*** 0.198 0.303 (0.662) (0.001) (0.112) (0.530) (0.222) (0.749) (0.173) (0.006) (0.940) (0.915) Caribbean 2.516 1.833 7.061** 3.450 3.627 –0.271 4.133** 1.208 –1.316 5.640 (0.292) (0.179) (0.042) (0.192) (0.243) (0.949) (0.044) (0.615) (0.456) (0.164) Caribbean – 4.057 1.405 0.528 2.371 0.146 4.188 1.205 1.804 3.932 1.438 Update (0.162) (0.473) (0.896) (0.511) (0.967) (0.564) (0.620) (0.583) (0.154) (0.763) Rest of LAC 3.911* 3.788* 3.778* 1.634 3.771* 3.200* 4.989*** 2.006 3.168 4.665 (0.099) (0.084) (0.060) (0.406) (0.085) (0.052) (0.005) (0.309) (0.171) (0.108) Rest of LAC – –2.680 1.976 0.401 3.652 2.384 0.423 –0.940 5.084 –3.910 –4.138 Update (0.325) (0.591) (0.905) (0.205) (0.448) (0.891) (0.739) (0.138) (0.139) (0.297) Observations 144 150 150 150 135 150 150 150 149 112 Source: Heysen and Auqui 2011. Note: LAC-6 countries include Brazil, Chile, Colombia, Mexico, Peru, and Uruguay. Robust p-value in parentheses. Signi�cance: *** p < 0.01, ** p < 0.05, * p < 0.1. 161 162 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? FIGURE 10.1 BCP assessments, LAC-6 At the same time, there is considerable variation across countries, depending on their other risks level of economic development, both across 1.0 the world and within the region. The coef- institutional framework 1.5 consolidated supervision �cient of GDP per capita is always positive 2.0 and significant, which shows that pruden- 2.5 tial oversight is a gradual, collective learning process that advances with overall economic authorizations 3.0 corrective actions development. In addition, the regressions 3.5 also control for nonlinearities by introducing 4.0 the square of GDP per capita. The coef�cient in this case varies signi�cantly in importance supervision capital across principles. This variation reflects the fact that some principles are either less relevant to simpler � nancial systems (hence, not a real source of concern for those coun- market discipline internal control tries) or just tougher to ful� ll (still a source credit risk of concern, but one that is relatively under- benchmark, mean score observed value, mean score standable). Thus, the corresponding regres- sion coefficient can be used to score and Source: Authors’ calculations based on �nancial sector assessment program data from the World rank the principles by order of complexity Bank and the IMF. (�gure 10.6). Strong caveats are in order. As regards FIGURE 10.2 BCP assessments, Caribbean standards and codes, only the BCPs are reviewed and not the other reports of obser- vance on standards and codes (ROSCs) other risks 1.0 reviewed by the World Bank and IMF (for institutional framework 1.5 consolidated supervision example, the standards for insurance, capital markets, audit and accounting, creditor rights 2.0 and insolvency systems, and so on, for which 2.5 a meaningful cross-country comparison is authorizations 3.0 corrective actions more dif�cult due to the much more limited 3.5 number of observations). Second, the BCP 4.0 evidence is clearly limited in scope and needs to be interpreted with caution, as the styles supervision capital and criteria of assessment were not necessar- ily fully uniform across regions. Third, many of the assessments date back a few years and hence do not reflect the most recent advances market discipline internal control the assessed countries may have made. Many credit risk similar caveats also apply to the World Bank benchmark, mean score observed value, mean score supervisory survey, which is also somewhat dated and where some questions can be mis- Source: Authors’ calculations based on �nancial sector assessment program data from the World interpreted; hence, the results may be open to Bank and the IMF. question. Be that as it may, the econometric evidence compliance over time in nearly all cases, as is valuable in that it picks up areas where a many LAC countries have used the results of second look appears to be warranted. This these assessments to program and implement quantitative evidence is complemented with their reform agendas (�gure 10.5). a more qualitative, expert-based review of PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 163 the main oversight reforms recently intro- FIGURE 10.3 BCP assessments, rest of LAC duced, as well as with an assessment of the most important remaining issues. 3 In addi- other risks tion, the chapter uses the results of a recent 1.0 survey on systemic oversight jointly con- institutional framework 1.5 consolidated supervision ducted by the World Bank and the Associa- 2.0 tion of Supervisors of Banks of the Americas 2.5 (ASBA) (annex 10.C).4 authorizations 3.0 corrective actions 3.5 The progress and the 4.0 remaining gaps At � rst glance, �gures 10.1, 10.2, and 10.3 supervision capital indicate very wide differences in overall compliance with the BCP principles across the three LAC subregions, even after con- trolling for different levels of economic market discipline internal control development. LAC-6 countries generally credit risk outperform, while the other subregions benchmark, mean score observed value, mean score underperform, the other LAC countries more strongly than the Caribbean. This sug- Source: Authors’ calculations based on �nancial sector assessment program data from the World gests that much work may still lie ahead to Bank and the IMF. improve bank oversight in the lower-income countries. Moreover, there are many areas where differences across countries are very preconditions for an effective oversight—as, large, resulting in low levels of signi�cance indeed, evidenced by their low complexity for the underlying coef�cients (table 10.1). score (�gure 10.6). A more detailed group-by-group analysis is Although the BCPs do not show LAC’s thus warranted. institutional framework as a whole to be Three groups of principles (institutional worse than that of other regions, two basic framework, authorizations, and market dis- legal framework issues—the independence of cipline) show rather similar patterns. While bank supervisors and bank supervisors’ legal the absolute scores are not perfect, they are protection—are detected by the World Bank relatively high, and except for the non–LAC-6 survey of supervisory practices to be worse countries, the region is broadly on track with in LAC than in other regions (even after respect to its peers. This record suggests that controlling for differences in economic devel- the basic underlying conditions for a function- opment). Indeed, these issues remain prob- ing oversight (clear objectives and responsi- lematic beyond LAC. De jure independence bilities, independence, legal powers, effective includes such key components as budgetary entry and acquisition requirements, and basic independence and operational independence transparency and disclosure—including good (for example, the capacity to issue or with- accounting and auditing practices) have been draw a license and the power to issue pruden- mostly put into place, albeit with some varia- tial regulations). De facto independence also tion across subregions. Overall, this is good requires freedom from political interference news. However, these are also areas where or industry capture. The region has weak- one would expect progress to be the easiest to nesses on both counts. The World Bank– achieve, both because it mostly requires hard- ASBA survey reports that, while most super- wired legal and institutional reforms, rather visors have operational independence, fewer than trickier and softer capacity buildups, have budgetary and administrative inde- and because such areas are truly minimum pendence to set salary scales (40 percent of 164 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? FIGURE 10.4 Compliance with BCP principles by LAC countries a. Share of countries compliant or largely compliant with BCP principles 100 90 80 70 compliance (%) 60 50 40 30 20 10 0 2 4 1) 3 25 20 6) 22 3) 4) 5 19 5) 9 11 21 23 18 10 8 17 2) 12 6 13 M 24 RS 1( 1( 1( 1( 1( 1( Basel Core Principle b. Average compliance with BCP principles for LAC countries 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 country Source: Authors’ calculations based on Financial Sector Assessment Program data from the World Bank and the IMF. Note: This �gure considers the last assessment of compliance with the BCP conducted for each country in the context of a Financial Sector Assessment Pro- gram. The list of core principles corresponds to the 2006 methodology. The assessments conducted under the original 1999 BCP methodology have been mapped into the revised 2006 methodology list of core principles. The term bcpRMS comprises four core principles associated with the supervision of risks that were part of one core principle in the 1997 BCP methodology: risk–management process (CP 7), liquidity risk (CP 14), operational risk (CP 15), and inter- est rate risk (CP 16). Panel a shows the average compliance across country for each principle. Panel b shows the average compliance for all countries with available information. In panel B, the ratings range from 1 (fully compliant) to 4 (noncompliant). Also for panel b, the number of principles in each country ranges from 23 to 28 depending on data availability. respondents indicated the lack of such inde- likely in LAC than in the rest of the world. pendence). Moreover, the World Bank survey The de jure weaknesses are also corroborated of supervisory practices indicates that remov- by a recent survey on governance practices ing the head of supervision by executive deci- in regulatory agencies, which �nds that LAC sion without congressional approval is more regulators lack sufficient independence to PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 165 FIGURE 10.5 Financial regulation and supervision progress in nine LAC countries 25 5 1 (6) 18 20 6 21 17 11 1 (5) Basel Core Principle 23 3 1 (3) 1 (1) 22 9 4 1 (2) 12 1 (4) 13 10 24 19 8 2 RSM –20 0 20 40 60 80 100 percent of countries compliant or largely compliant before FSAP (2000–05) current FSAP: progress (2005–10) Source: Heysen and Auqui 2011. issue regulations and, in some cases, to take FIGURE 10.6 Complexity index certain severe, crucial actions, such as license revocation.5 Nonetheless, signi�cant progress in addressing these issues has been recently other risks 10 achieved in several countries, notably Chile, authorizations 9 8 supervision Colombia, and Mexico. 7 6 Insufficient legal protection for bank 5 4 supervisors is also an important remain- institutional framework 3 consolidated supervision 2 ing issue in most of LAC (except, perhaps, 1 0 for the Caribbean), with less than half of the countries providing adequate legal pro- corrective actions capital tection to supervisors. 6 The World Bank survey indicates that LAC supervisors are more likely to be personally liable for dam- market discipline credit risk ages caused by their actions or omissions internal control in the discharge of their of�cial duties than are their peers elsewhere. Here again, how- Source: Authors’ calculations based on �nancial sector assessment program data from the World ever, many countries—including the Chile, Bank and the IMF. 166 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? the Dominican Republic, and Peru—have requirements under Pillar 1.9 Progress with recently taken important steps to upgrade the implementation of Pillars 2 and 3 has their legal framework. been similarly limited.10 As regards credit A second set of BCP principles (capital risk, many countries have tightened their adequacy, and oversight of credit risk and rules for loan classi�cation and provision- other risks), which focuses on risk manage- ing, but only the most advanced are moving ment, shows much unevenness across the sub- toward a forward-looking, internal-model regions. For example, only one-third of LAC approach that focuses on expected (rather countries had an effective supervisory frame- than incurred) losses. However, several work to determine whether banks have ade- countries have put in place comprehensive quate risk-management policies, processes, approaches that cover all stages of the credit and strategies for the size and nature of their process, thereby allowing supervisors to activities. Both the Caribbean countries and require improvements at an early stage. Some the non–LAC-6 countries underperform on countries have also established global large- capital adequacy. More than half of LAC exposure limits by sector, region, or other countries had capital requirements that did highly correlated exposures, a measure that not meet the minimum Basel I international is particularly relevant in LAC in view of the standards.7 By contrast, LAC-6 countries relatively concentrated banking systems and are on track or even outperform on credit income distributions. Finally, most partially risk and other risks. As the latter are com- dollarized countries in LAC have taken mea- plex areas (�gure 10.6), this suggests that the sures, including higher risk weights or pro- higher-income countries have worked hard visions on dollar loans (especially those to and become pro�cient overall at addressing debtors in the nontradable sector), to address dif�cult risk-management issues. the credit risks associated with borrowers’ However, much remains to be done. Most currency mismatches. LAC-6 countries have implemented most Next come two groups of principles with of the legal or regulatory reforms required average to high complexity scores (supervi- to conform to Basel I, including capital sion and internal control) in which Carib- charges for market and operational risks bean countries and, even more strongly, the (table 10.2). But the implementation of Basel other LAC countries also lag, particularly II has been limited thus far. 8 Only a few as regards supervision. Prima facie, this countries have implemented the new capital could indicate that much of the progress in TABLE 10.2 Capital adequacy requirements for selected LAC-6 countries   Argentina Brazil Chile Colombia Mexico Peru Capital adequacy ratio (% of risk-weighted assets) 8 11 8 9 8 10 Standard Basel I Basel II Basel I Basel I Basel II Basel II Authorization for Basel II internal models No Yes No No Yes Yes Power to require a higher capital adequacy ratio Yes Yes Yes Yes Yes Yes to individual banks based on the supervisory risk assessment Capital charges for: Credit risk Yes Yes Yes Yes Yes Yes Interest rate risk in the trading book Yes Yes No Yes Yes Yes Foreign exchange risk Yes Yes No Yes Yes Yes Other market risks (commodities, equity) Yes Yes No Yes Yes Yes Operational risks No Yes No No Yes Yes Basel II – Pillar 2 No Yes In progress No No In progress Basel II – Pillar 3 No In progress No No In progress No Source: Heysen and Auqui 2011. PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 167 prudential oversight in the lower-income whether domestic or cross-border. These are countries may still be more on the books than complex principles (�gure 10.6) with which in reality, with lack of effective implementa- the whole world is having a hard time. tion continuing to be a more severe problem. Remarkably, however, the LAC-6 coun- Nonetheless, important progress in tries outperform their benchmark, which is strengthening supervisory capacity has been indeed good news. made across the region. First, there has been To some extent, this group of principles a broad, but uneven and of recent vintage, prefigures the challenges of systemic over- shift from pure compliance-based supervi- sight, the main focus of this report. As regards sion toward risk-based supervision.11 Some consolidated supervision, only two of the 31 countries now have well-defined and fully LAC countries were rated fully compliant, integrated risk-based systems and have which is evidence of the complexities of the substantially strengthened capacity by incor- task. And only one-third of LAC countries porating staff with the necessary skills, effectively supervise the � nancial conglom- improving information systems, and upgrad- erates that operate within their jurisdic- ing supervisory processes. Others are still at tion. The absence of legislation granting the the early stages, however, with change being powers to conduct consolidated supervision hindered by a lack of human resources and a of conglomerates and the lack of a compre- weak culture of prevention and enforcement. hensive regulation, including key prudential Important progress has also been achieved in requirements on a consolidated basis, are not many countries in relation to stress-testing the only relevant problems. Opaque conglom- capacity. In addition to developing their own erate structures and insuf�cient cooperation capacity and models, many supervisors are and coordination among supervisors are also now also requiring banks to conduct their widespread problems. As a result, there is own regular stress tests and to use the results substantial scope for regulatory arbitrage.12 to review their risk-management policies. The high level of conglomeration and owner- The weaknesses in the supervision of ship concentration in the region is clearly an internal controls to a large extent reflect aggravating factor (annex 10.D).13 weaknesses in banks’ own risk-management While LAC has made some progress procedures. Here again, however, supervisors toward establishing a proper legal and super- have made important progress. In line with visory framework for consolidated supervi- international standards, most LAC countries sion, most countries still have much work have assigned to banks’ boards of directors ahead of them to make that framework fully the primary responsibility for establishing effective.14 Regulation and supervision are, and overseeing appropriate risk-management by and large, conducted according to license, systems. Many countries have issued detailed even though many licenses can be part of a and comprehensive norms or guidelines on single financial conglomerate. The lack of banks’ risk management. However, some control over holding companies domiciled countries have yet to develop the necessary abroad remains a widespread problem. Most supervisory capacity to assess their imple- countries face severe limitations in accessing mentation. Others have used a less prescrip- information on parallel-owned banks set up tive, rating-based approach. In all cases, a in other jurisdictions. Many countries need to key additional challenge for supervisors is to upgrade their capacity to access information ensure that banks’ board members have the and assess the risks posed by the nonbanking experience and ability to carry out these new entities of a � nancial conglomerate, includ- functions (especially in the local banks). ing through on-site examinations. Other The � nal group (consolidated and cross- common flaws include (a) the inadequate border supervision) includes all those de�nition of a �nancial conglomerate (which principles related to the oversight of inter- leaves important sources of risk outside of connections between � nancial institutions, the scope of supervision);15 (b) the lack of a 168 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? risk–management framework (including pru- supervisors. Last but not least, many host (as dential requirements) and enforcement pow- well as home) supervisors of LAC conglom- ers at the conglomerate level;16 and (c) insuf- erates are unsatis�ed with the scope, depth, �cient powers to control the establishment of and timeliness of the information shared by potentially problematic cross-border opera- their counterparts across borders in the con- tions, to require changes in the conglomerate text of the MOUs. structure when needed, and to mitigate local or cross-border regulatory arbitrage. Both domestic and cross-border coordi- The systemic oversight nation among supervisors require further challenges ahead strengthening. As regards within-border Systemic oversight differs radically from coordination, while some countries still face traditional prudential oversight in that the legal restrictions, the most common prob- whole is no longer simply the sum of the lem is the lack of an effective operational parts. Hence, the traditional institution-by- framework for cooperation that goes beyond institution approach (making sure that each sharing information.17 To address this prob- of the parts is sound and ready to go) is no lem, Colombia and Peru have established, or longer suf�cient, nor is the wider focus on are in the process of establishing, a uni�ed individual � nancial conglomerates. It needs supervisory agency. However, coordination to be complemented in three fundamen- problems can continue to arise even within tal ways: � rst, as � nancial systems become this uni�ed body, especially where oversight increasingly dense and interconnected, con- is done by license. Thus, in Mexico and necting the parts and understanding how Uruguay, important organizational changes the system is “wired� becomes the name of and a movement toward universal banking the game; second, as market failures (par- licenses have also been undertaken to address ticularly collective failures) interact with internal problems of coordination. macroeconomic conditions, endogenous A fortiori, cross-border cooperation unstable � nancial dynamics are generated remains an even greater challenge, which is that require a proactive dynamic response all the more crucial to the region in view of (prudential oversight turns time-dependent); the importance of foreign banks.18 No doubt, and third, as financial development and some significant improvements have taken financial stability become intertwined in place. Most countries have signed bilateral increasingly complex ways, it becomes and multilateral memorandums of under- essential to get ahead of the curve and adjust standing (MOUs) as a framework for coop- developmental and prudential policies as a eration, and colleges of supervisors have been system of interconnected policies. established to oversee the largest interna- For all three reasons, the change in super- tional conglomerates. However, many of the visory thinking needs to undergo a quantum MOUs lack operational teeth, and impor- leap. But, in addition, systemic supervisors tant limitations remain, some of which are need to improve the way they talk to each not adequately captured by the BCP assess- other (this is essential to connect the parts), ments. First, the MOUs are often not effec- to central bankers (this is essential to connect tive in overseeing nonbanking groups or the the dynamics, as well as to connect the parts nonbank entities of a banking group. Second, with the whole), and to �nance ministers and they do not include an adequate cooperation their staff (this is essential to connect stabil- framework for troubled entities. Third, LAC ity and development). A tough and complex countries that host subsidiaries or branches agenda indeed! of large international banking institutions LAC clearly needs to get ready. Thus that are large for the host country yet small far, LAC has mostly escaped the dark side for the home country frequently see their of financial development because the level concerns ignored or downplayed by the home of density and interconnectedness of its PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 169 �nancial systems was still moderate. Yet LAC United States and other developed economies. systems are rapidly evolving. Indeed, early Last but not least, LAC’s rising exposure to signs of potential trouble ahead from heav- world turbulence—an inevitable side effect ily interconnected systems already showed of � nancial globalization—is an increasing up in the global �nancial crisis (for example, source of concern. the problems of Sofoles � nancing in Mexico Thankfully, LAC has a leg up on the or corporate derivative options in Brazil and process. First, its prudential buffers are cur- Mexico). Thus, in view of lead times and rently high. LAC banks largely exceed their longer-term dynamics, now is the time for benchmarks as regards key prudential buf- thinking ahead. While banking crises caused fers (solvency, liquidity, and profitability). by unsustainable macrodynamics may have Indeed, LAC currently has, on average, some become a thing of the past, “good� macro- of the highest reported prudential buffers in management can also set the stage for ulti- the world (�gure 10.7). This consequence of mately more lethal endogenous financial LAC’s turbulent past provides some breath- dynamics, as was recently the case in the ing room to plan ahead. Second, as just FIGURE 10.7 Selected soundness indicators (yearly mediums) a. Bank capital/assets b. Regulatory capital –6 4 3 regulatory capital/RWA (�ltered) –7 bank capital/assets (�ltered) –8 2 –9 1 –10 0 –11 –1 –12 –2 –13 –3 –14 –4 4 5 6 7 8 9 10 11 4 5 6 7 8 9 10 11 log of GDP per capita log of GDP per capita c. Liquid assets/total assets d. Bank’s ROA 0 0 liquid assets/total assets (�ltered) –1 –10 –2 bank ROAA (�ltered) –20 –3 –30 –4 –5 –40 –6 –50 –7 –60 –8 4 5 6 7 8 9 10 11 4 5 6 7 8 9 10 11 log of GDP per capita log of GDP per capita Asia Eastern Europe G7 LAC-7 other advanced economies spline �t Source: de la Toree, Feyen, and Ize 2011. 170 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? emphasized, LAC supervisors have made is viewed as a top priority, closely followed important strides toward improving tradi- by the need to adjust prudential norms to tional oversight, thereby providing the essen- account for systemic risk, making pruden- tial foundation on which a sophisticated tial norms more countercyclical and enhanc- systemic oversight capability may be built. ing supervisory powers to take discretionary Third, LAC’s numerous past crises have given actions. The majority of respondents also its supervisors a de�nite learning advantage. thought that there should be a fundamen- While the roots and dynamics of the systemic tal rede� nition of the roles and functions of crises of the past may differ from those of the supervisors toward making them more pro- future, in the end, both require supervisors to active (annex 10.C). think and act systemically. The reform agenda ahead can be broken Last but not least, LAC’s supervisors are down into three main pieces, each of which well aware of the need for action. This comes is discussed in the following three chap- out clearly from the World Bank–ASBA sur- ters of this report: macroprudential policy vey. There is a broad perception of an increas- (chapter 11, focused on cyclical dynamics), ing likelihood that �nancial crises induced by microsystemic regulation (chapter 12, focused systemic risk could erupt as systems develop on interconnectedness), and systemic supervi- and become more sophisticated.19 Enhancing sion (chapter 13). All three will involve a sea supervisory capacity to assess systemic risk change in approaches and mind-sets. PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 171 Annex 10.A Methodology for the econometric analysis of BCP Ratings Regressions were run for 150 assessments of within the region (so as to detect regional compliance with the Basel Core Principles progress). To improve the �t and simplify the for Effective Banking Supervision (BCP). analysis, the principles are aggregated into The dependent variables take the follow- groups corresponding to 10 different dimen- ing values: 1 (compliant), 2 (largely com- sions of regulation and supervision (annex pliant), 3 (materially noncompliant), and 4 table 10.A.1). A principal component analy- (noncompliant). Hence, lower coefficients sis was used to set the weights of each prin- are associated with better ratings (stronger ciple in the group. Benchmarked values for supervision). Controls include GDP per cap- each group of principles in each region are ita, the square of GDP per capita, whether calculated on the basis of workhorse regres- the ratings correspond to a � rst assessment sions that do not include regional dummies. or to an update, whether the assessment was The benchmarked values are compared conducted based on the initial or revised to the observed values for each group and BCP methodology, a dummy for each of the region. To limit the impact of outliers, medi- three subregions, and, within each region, a ans (rather than means) are used to calculate separate dummy for the updates conducted regional scores. TABLE 10.A.1 BCP consolidation into groups Group 1997 Core principles 2006 Institutional 1.1 Responsibilities and objectives 1.1 1.2 Independence, accountability, transparency 1.2 1.3 Legal framework 1.3 1.4 Legal powers 1.4 1.5 Legal protection 1.5 1.6 Cooperation 1.6 Authorizations 2 Permissible activities 2 3 Licensing criteria 3 4 Transfer of signi�cant ownership 4 5 Major acquisitions 5 Capital 6 Capital adequacy 6 Credit risk 7 Credit risk 8 8 Problem assets, provisions and reserves 9 9 Large exposure limits 10 10 Exposures to related parties 11 Other risks Risk management process 7 11 Country and transfer risks 12 12 Market risks 13 13 Liquidity risk 14 13 Operational risk 15 13 Interest rate risk in the banking book 16 Internal controls 14 Internal control and audit 17 Not included 15 Abuse of �nancial services 18 16 Supervisory approach 19 17 Supervisory techniques 20 Supervision 18, 19 Supervisory reporting 21 Market discipline 21 Accounting and disclosure 22 Corrective actions 22 Corrective and remedial powers of supervisors 23 20, 23 Consolidated supervision 24 Consolidated supervision 24, 25 Home-host relationships 25 Source: Heysen and Auqui 2011.     172 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? Annex 10.B Highlights of the 2007 World Bank survey of supervisory practices This analysis is based on the 2007 Sur- ing, brokering and dealing, mutual funds, vey of Bank Regulation and Supervision insurance, and real estate investment. around the World (updated on June 2008), • Loan classification and provisioning conducted by the World Bank. 20 The sur- (LOANC). Loan classification practices vey includes responses of 142 countries to are stricter in LAC (but not in LAC-7) more than 300 questions covering 12 broad than in the rest of the world. topics: entry into banking, ownership, • Transparency and disclosure (ACCT- capital, bank activities, external audit, PR A): Both LAC and LAC-7 lag in the management, liquidity and diversi� cation adoption of internationally accepted requirements, deposit protection schemes, accounting standards. However, LAC-7 provisioning requirements, accounting and outperforms in other transparency prac- disclosure, remedial actions, and supervi- tices (FSTRANS). Also, LAC banks are sion. Regressions, controlling for GDP more likely to be rated by external rating per capita and including an overall LAC agencies than are their peers in the rest of dummy (or alternatively just a LAC-7 the world (ERCMON). dummy applying to the L AC - 6 coun- • Super vision resources (PSUPPBNK). tries plus Argentina, which in this case is There are more professional bank super- included in the survey), were run for 56 of visors per bank in LAC countries than in the questions in the survey. Only speci�ca- the rest of the world. tions with at least one signi� cant variable • Fr e qu e n c y of o n – s it e i n s pe c t io n s are presented (annex table 10.B.1). The (ONSITEINS). The frequency of on–site main results are as follows: inspections conducted in large or medium– size banks is lower in the LAC region. • Legal protection of bank supervisors • Remedial powers of super visors (INDBANK). Supervisors in both LAC (OSPOW ER). LAC supervisors have and LAC–7 are more personally liable for greater powers to impose corrective actions damages caused by their actions or omis- on supervised institutions. They are also sions as supervisors. more likely to be able to declare a bank • Ind e pend ence of bank super v isor s insolvent or to intervene in a bank’s activi- (INPOLI). The removal of supervisors ties without court permission (DIPOWER). in LAC (but not LAC–7) is less subject Moreover, LAC-7 countries’ supervisors are to congressional approval and more sub- also more likely to be able to appoint a liq- ject to a simple decision by the president, uidator or to supersede shareholders’ rights prime minister, or minister of � nance. without court permission (COURTINV). • Restrictions on bank activities and investments (OVER3AR). Banks in LAC Highe r l e vel of conglom e rat ion (but not in LAC-7) face more restric- (CONG10B). The percentage of capital in tions in engaging in a variety of � nancial the largest 10 banks owned by �nancial con- intermediation and investment banking glomerates is higher than in the rest of the activities, such as securities underwrit- world. PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 173 TABLE 10.B.1 World Bank survey of supervisory practices a. BCL regressions with LAC_DUMMY (those that have at least one significant exogenous variable) Variable LGDPPC LAC_DUMMY Observations R-squared SOR –0.0612** (–2.411) –0.109 (–0.992) 133 0.05 CONG10B 1.329 (0.453) 31.82*** (3.776) 69 0.12 OVER3AR –0.401*** (–4.474) 1.144** (2.537) 130 0.19 BONF –0.147*** (–4.447) 0.617*** (2.784) 131 0.18 NFOB –0.134*** (–3.606) –0.103 (–0.478) 132 0.07 OVERFCR –0.361*** (–4.683) 0.712 (1.271) 129 0.15 OVERBNK –0.549*** (–5.131) 1.591*** (2.730) 128 0.23 OVERAFC –0.745*** (–5.387) 1.766** (1.998) 127 0.22 LIMITFBEO 0.148*** (3.806) 0.0877 (0.656) 132 0.13 OSPOWER –0.270* (–1.869) 0.939** (2.190) 103 0.06 PCPOWER –0.407** (–2.583) 0.629 (1.091) 107 0.08 DIPOWER –0.00142 (–0.0273) 0.526*** (4.177) 130 0.04 COURTINV 0.105** (2.369) –0.195 (–1.351) 126 0.07 LOANCS –23.32 (–1.503) –112.4** (–2.251) 79 0.06 PSTRING –8.248** (–2.088) –4.547 (–0.411) 93 0.09 SUPTENURE 0.401* (1.812) 3.103*** (2.659) 84 0.12 INDPOLI 0.0352 (1.428) –0.205** (–2.330) 129 0.04 INDBANK –0.00405 (–0.235) –0.459*** (–3.778) 129 0.15 INDSA 0.0106 (0.234) –0.746*** (–4.336) 114 0.11 MULSUP –0.00779 (–0.272) –0.364*** (–3.283) 131 0.06 FUNDID 2.687** (2.242) 10.77 (1.542) 101 0.06 GOVBANK –4.175*** (–3.178) 0.826 (0.146) 102 0.08 FSTRANS 0.172*** (3.212) 0.110 (0.479) 123 0.07 ACCTPRA –0.00109 (–0.0631) –0.460*** (–3.867) 128 0.18 ERCMON 0.0803 (1.134) 0.695** (2.140) 73 0.09 ONSITEINS 0.0160 (0.465) –0.394*** (–2.661) 129 0.04 PSUPPBNK 0.0457 (0.189) 4.051*** (3.583) 90 0.10 BNKDEV 0.219*** (10.41) –0.193** (–2.282) 147 0.44 b. BCL regressions with LAC-7_DUMMY (those that have at least one significant exogenous variable) Variable LGDPPC LAC-7_DUMMY Observations R-Squared SOR –0.0600** (–2.357) –0.0558 (–0.327) 133 0.04 CONG10B –0.0381 (–0.0127) 30.87*** (2.814) 69 0.04 OVER3AR –0.416*** (–4.536) –0.0190 (–0.0251) 130 0.14 BONF –0.154*** (–4.575) 0.399 (0.926) 131 0.11 NFOB –0.132*** (–3.551) –0.345 (–1.103) 132 0.08 NBFFOB –0.0776 (–1.998) –0.123 (–0.388) 131 0.03 OVERFCR –0.369*** (–4.644) –0.0856 (–0.0901) 129 0.13 OVERBNK –0.570*** (–5.206) 0.405 (0.369) 128 0.17 OVERAFC –0.769*** (–5.373) –0.0630 (–0.0398) 127 0.18 LIMITFBEO 0.147*** (3.852) –0.142 (–0.698) 132 0.13 OSPOWER –0.298** (–2.069) 1.353** (2.115) 103 0.06 PCPOWER –0.430*** (–2.763) –0.197 (–0.244) 107 0.07 DIPOWER –0.00867 (–0.164) 0.492*** (3.142) 130 0.01 COURTINV 0.108** (2.448) –0.297** (–1.992) 126 0.07 (continued next page) 174 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? TABLE 10.B.1 (continued) b. BCL regressions with LAC-7_DUMMY (those that have at least one significant exogenous variable) Variables LGDPPC LAC_DUMMY Observations R-squared PSTRING –8.014** (–2.022) –28.05 (–1.053) 93 0.11 SUPTENURE 0.314 (1.430) 5.521*** (3.044) 84 0.15 INDBANK 0.00231 (0.129) –0.526*** (–2.985) 129 0.08 INDSA 0.0285 (0.634) –1.035*** (–3.525) 114 0.09 BACCT –0.00721 (–0.222) 0.501*** (8.901) 125 0.04 FUNDID 2.481** (2.079) 26.25** (2.465) 101 0.10 GOVBANK –4.166*** (–3.302) 16.71** (2.035) 102 0.11 FSTRANS 0.170*** (3.189) 0.374** (2.420) 123 0.08 ACCTPRA 0.00566 (0.315) –0.558*** (–3.171) 128 0.11 ERCMON 0.0613 (0.854) 1.244*** (4.040) 73 0.15 PSUPPBNK –0.0677 (–0.255) 2.543** (2.291) 90 0.01 BNKDEV 0.220*** (10.32) –0.302** (–2.373) 147 0.44 Source: Heysen and Auqui 2011. Note: Robust t-statistics are shown in parentheses; Signi�cance: *** p < 0.01, ** p < 0.05, * p < 0.1. PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 175 Annex 10.C Joint World Bank–ASBA survey on systemic oversight The joint World Bank–Association of Super- • Prudential regulations in LAC tend to be visors of Banks of the Americas (ASBA) sur- comprehensive, with the notable excep- vey was aimed at gauging the importance of tion of regulations for cooperatives, systemic oversight issues in LAC, the status although liquidity regulations are not of practices in the region, and the main chal- widely applied. lenges going forward. Questions covered • There is broad consensus on the need to three broad topics: microprudential over- adjust prudential norms to account for sight, management of economic cycles, and systemic risk. Regulation of the cross- macroprudential oversight. The survey was sectional components of systemic risk is, as sent to all the LAC bank supervisory agen- elsewhere, incipient. While LAC is ahead cies that are members of ASBA. The head of other regions in the use of countercycli- of banking supervision of the agency was cal provisions, the use of other countercy- asked to forward the survey to the finan- clical prudential regulations is limited. cial stability or research department of the • Supervisory authorities in LAC consider central bank, even if not an ASBA member. that there should be a fundamental rede�- Of the 31 member countries of ASBA in the nition of the role and functions of the LAC region, 19 supervisors and 9 central supervisors, including making them more banks responded to the survey. Responses proactive. indicated a broad consensus among regional • Supervisory powers to request additional � nancial authorities on the need to enhance buffers to account for the buildup of the current systemic oversight framework. systemic risks appear limited in several Financial authorities see an increasing likeli- jurisdictions. Agencies’ legal mandates, hood that a � nancial crisis induced by sys- political and industry pressures, and lack temic risk could occur as systems develop of adequate legal protection for supervi- and become more sophisticated. Systemic sors hamper the exercise of supervisory risk monitoring needs strengthening, with discretion. smaller countries being less comfortable • In most LAC countries, the stability of the about their monitoring framework. Survey � nancial system is a collaborative effort responses indicate the following: among various regulatory bodies, albeit there appears to be scope for improving • Enhancing supervisory capacity to assess coordination between supervisors and the systemic risks and to identify risks in central bank as well as among different sophisticated products is a top priority supervisors. among supervisors. • Rethinking the organization of supervi- • Regulatory perimeters in LAC are set quite sion is also considered an important fac- wide, and resetting them does not appear tor in improving systemic oversight. to be a top priority at this time. Safety • There are important challenges on cross- net perimeters, in contrast, are much nar- border coordination dealing with systemic rower, mainly covering only commercial risk in the LAC region. Lack of effective banks. arrangements for exchanging information • The most pressing perimeter issues relate across borders and for sharing the resolu- to risk shifting among conglomerates, but tion costs of institutions operating cross- powers to regulate conglomerates do not border are sources of concern among appear comprehensive. regional supervisors. 176 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? Annex 10.D Financial concentration in LAC In most of the largest LAC countries, � nan- Moreover, the relative decline of banks cial institutions have been highly concen- and rise of nonbanks and capital markets trated, and, as table 10.D.1 shows, they intermediaries do not necessarily mean that became more so between 1998 and 2008. the new players are becoming more pre- The 10 largest banks have market shares dominant. Because of the prevalence of large of 75 to 90 percent. The situation is simi- business groups in Latin American coun- lar among life insurance companies and a tries, many of the important players among little less concentrated among general insur- institutional investors have close ties to large ance companies. The 10 largest mutual banks. In some countries, they are actually fund companies generally represent about part of the bank, while in others they belong 80 percent of the market. Among pension to the same �nancial group. In addition, most funds, the concentration is also extremely countries permit nonbank � nancial � rms to high. This suggests that competitive pres- own banks. Thus, ownership concentration is sures may be less intense than in markets high across all market segments (annex table with larger numbers of participants and 10.D.2).21 It is particularly large in pensions, less concentrated distributions, such as the mutual funds, and investment banks. But it is United States. also substantial in banks.22 TABLE 10.D.1 Market share of largest companies and funds in selected LAC countries Percentage of market share Argentina Brazil Chile Colombia Mexico A. Year 1998 Banks Largest 5 48a 58a 59a — 80a All Insurance Largest 10 — 67b — — 100b Pension Funds Largest 2 53c —c 62c 77c 45c B. Year 2008 Banks Largest 5 51 67 73 64 77 Largest 10 74 76 94 86 92 Mutual Funds Largest 10 79 84 87 73 88 General Insurance Largest 10 54 — — 78 — Life Insurance Largest 10 82 62 73 61 88 Pension Funds Largest 10 — 60 100 100 92 Largest 2 — — 55c 52c 33c Sources: Raddatz 2011. Note: a. Caprio and Honohan 2001. b. Srinivas, Whitehouse, and Yermo 2000. c. AIOS (Asociación international de organismos de supervisión de fondos de pensiones). PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 177 TABLE 10.D.2 Ownership concentration in selected Latin American countries a. Percentage of largest 10 institutions related to largest 10 banks (%) Argentina Brazil Chile Colombia Mexico Insurance General 10 — 30 30 40 Life 10 50 40 30 50 Pension funds — 10 40 33 50 Mutual funds 70 60 80 — 80 Investment banks and brokerages Investment banks 0 90 60 33 — Brokerages — 70 — — 30 b. Share of assets of largest 10 institutions related to largest 10 banks (%)   Argentina Brazil Chile Colombia Mexico Insurance General 6 — 34 21 23 Pension funds — 2 53 53 64 Mutual funds 56 67 91 — 94 Invest. banks and brokerages Investment banks 0 98 52 83 — Brokerages — 87 — — 30 Source: Raddatz 2011. Note: Panel a shows the fraction of the largest 10 institutions in each segment (described in each row) that are related to the largest 10 banks in 2008. Panel b shows the share of assets in each segment that are in institutions that are related to the largest 10 banks. 178 PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? Notes And only Brazil and Mexico have initiated the implementation of Pillar 3. 1. This chapter draws heavily on the papers 11. Risk-based supervision calibrates the scope “Recent Trends in Banking Supervision in and intensity of supervision according to the LAC,� by Socorro Heysen and Martin Auqui level of risk individual institutions pose. (2011), and “Survey on Systemic Oversight 12. For example, to exploit supervisory gaps, Frameworks: Current Practices and Reform a conglomerate that has its main business Agenda� by Eva Gutierrez and Patricia and ultimate shareholders in one country Caraballo, which are part of the Edited may choose a second country as its home Volume that accompanies this LAC flagship supervisor. report. 13. The World Bank survey of supervisory prac- 2. Thirty-one BCP assessments were conducted tices con�rms that ownership concentration in the region in the context of the joint World is a particularly important issue in LAC. Bank–IMF Reports on the Observance of LAC countries exhibit a higher than average Standards and Codes (ROSC) and Financial percentage of capital in the 10 largest banks Sector Assessment Program (FSAP) since owned by conglomerates (annex 10.B). 1999—eight countries were assessed twice 14. The World Bank–ASBA survey clearly con- and one, Peru, was assessed three times. �rms that the potential for risk shifting within 3. See Heysen and Auqui (2011) for a more members of a conglomerate, reflecting insuf- complete qualitative discussion of LAC’s �cient powers to regulate them, is viewed as a progress and current stand as regards bank pressing issue. This is particularly the case for oversight. structures where �nancial (and real sector) 4. The survey helps identify the state of systemic companies that belong to the group are not oversight in the region and polls perceptions owned by the bank but by an entity that owns among regional supervisors about the key the bank. Few countries require constituting areas that will need development. a �nancial holding company to control the 5. See Seelig and Novoa (2009). conglomeration of �nancial sector activities, 6. It is often the case that supervisors’ legal pro- and in the majority of cases the holding com- tection ends when they leave of�ce. Such lack pany can be created abroad. Capital require- of continuity can have a seriously chilling ments over holding companies are rare. effect on supervisors. 15. The World Bank–ASBA survey reports that in 7. Lack of compliance is generally associated with seven LAC countries the de�nition of banking the absence of capital charges to cover risks, groups or �nancial conglomerates in the law such as market or operational risks, lower does not include non�nancial group entities. weights for some asset classes, or regulatory In four countries, supervisors do not have the differences about what constitutes capital. In a capacity to presume (based on objective crite- few countries, regulatory forbearance has also ria) which companies form part of the �nan- been a problem. cial conglomerate. In some cases, parent com- 8. The Basel II accord (introduced in 2004) panies or holding companies may also not be has three pillars. Pillar 1 deals with capi- included in the de�nition of a �nancial group. tal requirements, Pillar 2 with supervisory 16. The World Bank–ASBA survey reports that review, and Pillar 3 with market discipline. only eight countries have the power to impose Basel II is more comprehensive than Basel I a special capital requirement for a �nancial and relies on more formal and systematic risk group taken as a whole. While related-party modeling. limits appear to be in place in most cases, 9. While Brazil, Mexico, and Peru have imple- ownership limits exist for only half of the mented the capital requirements under Pillar respondents. 1, the use of internal models is still very lim- 17. The World Bank–ASBA survey reports that ited. As of the end of 2010, only one such formal arrangements were in place to discuss authorization has been issued by Mexico. and resolve potential issues of regulatory 10. Only Brazil has established guidelines requir- arbitrage across �nancial institutions with ing banks to assess their capital adequacy different licenses in only half of the countries. under all risks contemplated in Pillars 1 and About 40 percent of respondents indicated 2 and to have additional capital to cover that they conduct interagency crisis simula- losses generated during moments of stress. tion exercises, but not on a regular basis. PRUDENTIAL OVERSIGHT: WHERE DOES LAC STAND? 179 18. The World Bank–ASBA survey reports the evolves in sophistication, over 50 percent of lack of effective arrangements for cross- the respondents consider it somewhat likely border information exchange and discussion that a U.S.-type crisis could materialize. of common issues to be a very important con- 20. The survey results are discussed in Barth, cern for 75 percent of the respondents. The Caprio, and Levine (2008). lack of effective arrangements to deal with 21. The complexity of ownership structures in cross–border crisis was noted by 63 percent Latin America, which include control pyra- of respondents; the lack of effective arrange- mids and cross-holdings, suggests that there ments to discuss common issues between is likely to be a downward bias in these esti- home and host supervisors was noted by 61 mates. For additional analysis and informa- percent; the lack of effective cross-border tion on �nancial concentration in LAC, see inspections was noted by 57 percent; and the Srinivas, Whitehouse, and Yermo (2000) and lack of effective arrangements for sharing Impavido, Lasagabaster, and Garcia-Huitron the resolution costs of institutions operating (2010). cross-border was noted by 50 percent. 22. Caprio et al. (2007) report that, among LAC-7 19. Most Latin American supervisors and mon- countries, only 10 percent of the banks are etary authorities think it unlikely that a �nan- widely held, and in 70 percent of the cases cial crisis similar to the one recently expe- with concentrated ownership, the controller rienced in the United States could happen is a family. In the rest of the world, the share in their countries under the current stage of of widely held banks is 27 percent, and only �nancial development. However, as the system 33 percent of the banks are family owned. Macroprudential Policies over the Cycle in LAC 11 T he global � nancial crisis underscored remaining as LAC speci�c as possible. Key the need for a new prudential over- messages are as follows: sight policy framework to curb and manage systemic risk and reduce the cost of • The design of a proper macroprudential boom-bust �nancial cycles. Systemic regula- policy framework for LAC needs to be tion has two main dimensions: one stresses based on clear choices across a menu of the links through time (that is, time depen- progressively more ambitious goals. Those dencies), which is the focus of what is de�ned goals range from the most modest aim of here as macroprudential policy; and the other correcting the distortions brought about stresses the links across the � nancial system by traditional prudential norms and other (that is, interconnectedness), which is the macroeconomic and � nancial policies, to focus of “microsystemic� prudential policy.1 the most ambitious objective of dampen- Thus, the discussion of systemically oriented ing “excessive� fluctuations in asset and prudential oversight in this report is broken business cycles (by inducing agents to fully down into three tightly connected chapters. internalize externalities). Between those This chapter reviews macroprudential regu- choices is the intermediate (and currently lation as it applies to Latin America and the more popular) goal of simply making Caribbean (LAC); chapter 12 reviews LAC- � nancial systems more resilient to fluctua- relevant issues in microsystemic regulation; tions (that is, taking the cycle as given). and chapter 13 looks at both dimensions but • The framework also needs to reflect LAC’s from a supervisory (rather than regulatory) peculiarities as regards the risks and vul- perspective that stresses implementation.2 nerabilities the region has faced in the The literature on macroprudential regu- past; in particular, LAC’s � nancial cycles lation and policy has literally exploded over have been more frequent and pronounced the past two years.3 Thus, rather than trying and have ended badly more often than in to duplicate that material, much of which is other regions. highly specific and technical, this chapter, • At the same time, the region is currently as the rest of this report, provides a more facing a particularly perilous mix of holistic and conceptual perspective that tries external pressures (capital inflow surge, to link root causes to policy responses while commodity price boom) that raise the 181 182 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC premium for quickly establishing or con- transactions interact with the forces of the solidating the region’s macroprudential dark side (as surveyed in chapter 2), creating capacity. a class of social inef�ciencies that are neither • In addition to clearly setting the goals of necessarily nor systematically associated macroprudential instruments, other key with short- or long-term infl ation dynam- design issues include the proper mix of ics. The aims of macroprudential policy are rules and discretion, the choice between thus not perfectly correlated with those of more speci�c or more broad-based instru- monetary policy and go beyond the latter’s ments, and the choice between price-based main objective of coordinating and anchor- versus quantity-based regulations. ing agents’ expectations around a low and • Other basic lines of action will include stable inflation target. In sum, the macro�- putting in place a proper institutional nancial imperfections in the economy can- framework and a consistent policy mix; not be solved by a single instrument. monetary, fiscal, capital controls, and Inef�cient � nancial system dynamics can exchange rate policies all need to com- be explained in two fundamental ways: (a) plement and reinforce macroprudential exogenous shocks that get ampli�ed by vari- policy. ous �nancial frictions and (b) dynamics that are endogenous to the �nancial development The rest of this chapter is organized as process itself (mood swings) and get ampli- follows. Following up on the main threads fied by heuristically adjusted expectations laid out in chapter 2, the � rst section briefly and the psychology of the masses, often gen- reviews the implications for macropruden- erated and boosted by �nancial innovations. tial policy of the dynamics of the dark side In the academic literature, the � rst type of in �nancial development. The second section dynamics is typically developed in mod- explores how these dynamics may apply to els of rational expectations but mixed with LAC, � rst by reviewing the � nancial cycles agency or collective action frictions (that is, of the past, and then by taking stock of the frictions associated with the costly enforce- systemic risks that are building up and may ment, asymmetric information, and collec- continue to gather importance. The third tive action paradigms). The second type is section briefly discusses some key issues as associated with Knightian uncertainty and regards the design of macroprudential policy. bounded rationality (that is, the collective The chapter concludes by summarizing the cognition paradigm). main priorities and linkages in LAC’s macro- In the �rst type, an exogenous shock that prudential policy agenda. “comes out of the blue� sharply tightens a financial constraint typically derived from the need for “skin in the game,� or collat- Macroprudential policy and the eral (because of asymmetric information or dynamics of the dark side contract enforcement frictions). The � nan- As in the case of monetary policy, the jus- cial constraint in turn activates massive asset ti�cation for macroprudential policy derives purchases or fire sales that result in large from the fact that economies go through changes in asset prices. In turn, the change socially inef� cient cycles that call for pol- in asset prices further tightens (or relaxes) the icy action. Macroprudential policy focuses � nancial constraint by reducing (or raising) on achieving sustainable financial system the value of the underlying collateral. Finally, dynamics, an objective that is not reduc- the change in asset prices becomes consistent ible to those of other macro (fiscal and with fundamentals because it is validated monetary) policies. In particular, monetary by changes in productivity as assets change policy alone—focused as it is on infl ation hands—between the more informed and and its volatility—is insufficient to man- the less informed (from the specialist arbi- age these cycles because real and � nancial trageurs to the general market participants), MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 183 as in Shleifer and Vishny (1997b); between monetary policy can indeed be deployed to the ones with better technologies and the achieve macroprudential policy objectives— ones with worse technologies, as in Kiyotaki for instance, raising interest rates to tame an and Moore (1997); or between the optimists asset price bubble and avoid a deterioration and the pessimists, as in Geanakoplos (2009). of lending standards motivated by a search Because individuals (or individual � nancial for yield—although not always without costs. institutions) do not internalize these feed- Indeed, monetary policy not only may be a back effects, the upswing is characterized by suboptimal policy to pursue macropruden- a socially excessive expansion of credit and tial policy objectives but also may, by pur- leverage.4 Credit, asset prices, and real activ- suing that policy, end up deviating from its ity all rise in a self-reinforcing loop, and lend- main goal. The use of monetary policy to ing standards weaken as credit is extended to achieve macroprudential policy objectives marginal borrowers.5 These systemic fragili- can, furthermore, have adverse side effects— ties that were built up in the boom deepen in for example, a higher interest rate aimed at the downturn, when feedback effects go in deflating an asset bubble can further stimu- reverse, and a vicious circle develops between late “frothy� capital inflows, which in turn deleveraging, asset sales, weakening loan can increase � nancial system excesses. Con- portfolios, and deteriorating real activity. versely, macroprudential policy can overreach By contrast, in the mood swings story of when, under the guise of pursuing its objec- adverse � nancial system dynamics, the inef- tives, it is really used for other objectives for �cient ampli�cation effects are endogenously which it is not the best tool. This can be the gestated. Faced with the world of the new and case, for instance, when the real intention of unknown (that is, a range of nonreducible deploying macroprudential policy is to cope uncertainty about the future), market partici- with “impossible trinity� tensions arising pants no longer have a steady frame of refer- from misguided exchange rate policy. Finally, ence. As a result, as Keynes (1936) put it, “the there is a fuzzy area where macroprudential market will be subject to waves of optimistic policy and other macro objectives overlap; and pessimistic sentiment, which are unrea- hence, identifying the best policy tool is not soned, yet in a sense legitimate . . . [Because always easy.7 of uncertainty] . . . no solid basis exists for In pursuing its ultimate objective of achiev- a reasonable calculation.� Up the cycle, ing sustainable financial system dynamics, financial innovations and the continuously macroprudential policy may be calibrated improving economy fuel optimism and exu- to pursue four progressively more ambi- berance. At some point, however, a signi�cant tious lines of attack. For a modest � rst line dissonance suf�ces to initiate a mood swing, of attack, macroprudential policy can simply fueled on the way down by acute uncertainty concentrate on correcting those macroeco- aversion.6 In both cases, the adverse dynam- nomic and regulatory policies that unnec- ics are ampli�ed by flights to liquidity—that essarily contribute to the adverse financial are individually protective but collectively dynamics (that is, if one cannot help, one self-destructive—and �nancial failures—that should at least do no harm). For a second line exacerbate the crisis through capacity losses, of attack, macroprudential policy could seek contagion, and interconnectedness. to limit the vulnerability of the �nancial sys- The two types of market failures described tem to adverse dynamics (that is, if one can- above provide a clear rationale for macro- not control the dynamics, one may at least prudential policy as an independent policy be able to make the financial system more arena, different from that of monetary pol- resilient to those dynamics). For a third line icy. It is not always easy, however, to neatly of attack, macroprudential policy could aim discern between the objectives and the use at limiting the buildup of risks by controlling or misuse of monetary and macroprudential ampli�cation effects (that is, if one is unable policy instruments in practice. For starters, to prevent inef�cient cycles, one may at least 184 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC be able to limit their amplitude). For a fourth time, the order of dif�culty associated with and most ambitious line of attack, macro- each is clearly of a different magnitude. Thus, prudential policy may try to maintain the a properly thought-out reform agenda should economy on the right course at all times (that carefully adjust the tools and policies to the is, one can dispel the gestation of adverse objectives that can be realistically pursued dynamics in the � rst place by nipping them at any point in time, taking into account the right in the bud). risks and vulnerabilities a financial system The first line of attack may be built by may face as well as the response capacity maintaining sound fiscal and monetary supervisors may have. The last two sections policies (including avoiding unsustainable of this chapter explore these issues. However, currency pegs), removing the inherent pro- the next section �rst reviews the stylized fea- cyclicality of traditional (Basel I or II) pru- tures of LAC’s past cycles, to ground the sub- dential norms, and curbing other financial sequent discussion of macroprudential policy factors that fuel procyclicality, such as cur- issues in LAC-speci�c realities. rency mismatches and social moral hazard (expectation of bailouts or a “Greenspan put� on monetary policy). Crucial to this LAC’s cycles and vulnerabilities: � rst line of attack is to ensure that pruden- Lessons from the past? tial buffers are really used as buffers (instead This section characterizes � nancial cycles of being untouchable) when bad times hit, in Latin America, drawing from a sample as argued for instance by Goodhart (2010b) of 79 countries with quarterly information and Hellwig (2010). The second line of attack over the period 1970 –2010. 8 Using time requires building up stronger prudential buf- series techniques to date peaks and troughs fers (whether solvency or liquidity); the third and identify booms and busts, this section requires using macroprudential instruments benchmarks cycles of credit, bank lever- as cycle dampeners; and the fourth requires age (de� ned as the deposit-to-credit ratio), introducing (possibly time-dependent) Pigou- stock and housing prices, the real exchange vian taxes (that help internalize externalities), rate, and capital flows by comparing LAC to monitoring and regulating � nancial innova- industrial countries and non-LAC emerging tion, and judiciously tightening or relaxing countries.9 Cycles that did not end with a prudential limits along the path (to help coor- banking crisis are also compared with those dinate or guide the behavior of market partic- that did, and the last credit cycle (that is, the ipants toward socially desirable outcomes). In boom that preceded the global � nancial cri- all cases, policy actions take into account that sis) is compared with the previous cycles.10 the whole is more than the sum of the parts. Reflecting the interconnectedness, spillovers LAC’s �nancial cycles have been both and externalities, and mood swings, what quite frequent and more pronounced matters from the macroprudential perspec- tive is not the riskiness of particular �nancial Based on LAC’s recent history, the uncon- instruments or financial institutions—the ditional probability (that is, the frequency realm of microprudential regulation—but over the sample period) of a credit boom rather their correlation with systemic risk. in LAC is similar to that in other emerg- While the literature on macroprudential ing economies but higher than in the indus- policies is already vast, one seldom � nds a trial countries (figure 11.1). Similarly, the clear and comprehensive characterization of unconditional probability of an equity price goals and objectives along the above lines. boom is higher in LAC than in the industrial Macroprudential tools and policies may countries. Remarkably, however, housing affect several of these (increasingly ambi- price booms are substantially less frequent tious) objectives, but they do it in different in LAC, which is consistent with the sharp ways and to a different extent. At the same underdevelopment of the mortgage market MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 185 FIGURE 11.1 Unconditional probability of booms and crises 16 14 12 10 probability (%) 8 6 4 2 0 banking lending equity price housing price capital flow crisis boom boom boom bonanza event industrial countries non-LAC emerging markets LAC Source: Calderón and Servén 2011. Note: This �gure presents the frequency of banking crisis episodes in the sample of countries. It computes the unconditional probability of banking crisis as the number of years where a banking crisis takes place divided by the number of years in the entire sample of the country. Banking crisis episodes are identi�ed as in Laeven and Valencia (2008). An analogous calculation is made for the frequency of lending booms, equity price and housing price booms, and capital flow bonanzas. Following Claessens et al. (2011a, forthcoming), these �nancial booms are de�ned as the top quartile of the upturn in credit, stock prices, housing prices, and gross capital inflows in the world sample. identi�ed and discussed in chapter 6. How- high inflation episodes, and recurring bank- ever, cycles (particularly credit) have been ing and currency crises. more protracted and abrupt, especially dur- ing downturns (table 11.1). For instance, the Credit cycles in LAC have tended to median drop in credit per capita in LAC dur- precede output cycles and follow ing peak-to-trough phases is approximately asset cycles 18 percent, four times larger than observed in industrial countries (4.4 percent) and over All macro variables are highly correlated twice the magnitude seen in other emerging with credit cycles but more particularly with markets. Likewise, cyclical fluctuations in output (more than 70 percent concordance) bank leverage, housing prices, and the real (�gure 11.2).11 However, real credit tends to exchange rates exhibit greater amplitude in precede output at turning points, particu- LAC, especially in cyclical downturns. These larly at the beginning of the downswing.12 facts echo the history of macroeconomic This pattern of precedence is even more pro- instability of the region, as reflected by nounced in output downturns that coincide unsustainable �scal and external positions, with a banking crisis (� gure 11.3). At the 186 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC TABLE 11.1 Main features of real and �nancial cycles Sample: 79 countries, 1970q1–2010q4 Average duration Median amplitude (%) Median slope (%) Upturn Downturn Upturn Downturn Upturn Downturn Real GDP Industrial countries 3.17 3.98 2.9 –2.4 0.6 −0.6 Latin America 3.36 3.76 5.6 −5.5 1.3 −1.4 Non-LAC emerging markets 3.34 3.61 7.5 −5.1 1.8 −1.6 Real credit per capita Industrial countries (IND) 5.08 6.12 3.9 −4.4 0.9 −0.9 Latin America 4.40 7.10 9.9 −17.7 2.4 −2.8 Non-LAC emerging markets 4.39 5.31 7.4 −7.1 2.2 −1.8 Credit-GDP ratio Industrial countries (IND) 4.56 7.15 3.5 −5.5 0.9 −1.0 Latin America 4.94 8.31 7.6 −15.9 2.0 −2.5 Non-LAC emerging markets 4.03 6.53 5.2 −7.4 1.2 −1.3 Bank leverage Industrial countries (IND) 4.31 10.33 3.3 −11.3 0.9 −1.5 Latin America 3.42 14.64 4.9 −31.4 1.9 −3.3 Non-LAC emerging markets 3.29 13.13 3.7 −20.4 1.3 −2.6 Stock prices Industrial countries (IND) 3.99 6.23 19.3 −35.9 5.3 −7.0 Latin America 4.09 6.42 36.3 −56.7 8.9 −11.2 Non-LAC emerging markets 4.08 6.21 33.4 −55.4 7.2 −8.9 Housing prices Industrial countries (IND) 3.94 7.06 2.9 −4.5 0.9 −0.9 Latin America 4.50 8.33 6.8 −19.7 2.4 −2.7 Non-LAC emerging markets 3.91 7.89 3.5 −8.1 1.0 −1.4 Real effective exchange rate Industrial countries (IND) 4.05 5.45 3.5 −6.7 1.0 −1.3 Latin America 3.94 6.01 6.4 −12.3 2.1 −2.3 Non-LAC emerging markets 3.72 5.89 5.4 −9.8 1.5 −2.0 Gross capital inflows (ratio to GDP) Industrial countries (IND) 3.29 5.17 3.7 −5.3 0.9 −1.1 Latin America 3.45 5.20 3.0 −4.2 0.6 −1.0 Non-LAC emerging markets 3.23 4.98 3.0 −4.4 0.8 −1.0 Source: Calderón and Servén 2011. Note: The table reports the average duration (in quarters) of the different cyclical phases (downturns and upturns) for real and �nancial variables. The statistics for amplitude and slope refer to sample median across episodes (averages for those statistics are not reported but are available from the authors upon request). The duration of downturns (recessions or contractions) is the number of quarters between peak and trough. Upturns (or recoveries), on the other hand, are de�ned as the early stage of the expansion that takes place when either the real or �nancial indicator rebounds from the trough to its previous peak. The amplitude of the downturn is the distance between the peak in real output and its subsequent trough, while 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. same time, lending booms are more likely reversal is a reminder of the limited role to be preceded by booms in asset prices.13 played so far in LAC by the availability of Not surprisingly, in the case of equity prices, mortgage credit in affecting housing prices. this result suggests that the latter are early But this is likely to change in the future as predictors of improving economic condi- mortgage markets develop, and is some- tions. It is somewhat remarkable, however, thing for which the authorities already need that housing prices in LAC tend to precede to be preparing. Also somewhat surpris- (rather than follow) credit booms. This order ingly, LAC’s capital flow bonanzas are less MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 187 FIGURE 11.2 Synchronization between real output and �nancial cycles 0.80 0.70 0.60 concordance index 0.50 0.40 0.30 0.20 0.10 0.00 real output, real output, real output, real output, real output, credit stock prices housing prices real exchange rates gross inflows event industrial countries non-LAC emerging markets LAC Source: Calderón and Servén 2011. Note: The �gures presented here represent the concordance statistics for the cycles of the corresponding pair of variables (Harding and Pagan 2002a). The concordance index takes values between 0 and 1 and measures the fraction of time that the two cycles are in the same cyclical phase. The concordance �gures presented are computed over the period 1970q1–2010q4. likely to be followed by lending booms than other regions. Nearly half of LAC’s banking in other regions (figure 11.4), which may crisis episodes followed lending booms, and reflect the greater buffering role of exchange more than half of these crashes were pre- rates. Indeed, real exchange rates move more ceded by either equity price booms or capi- strongly with the credit cycle than capital tal inflow bonanzas. Moreover, when crises inflows, a trend that has sharpened during have taken place, their costs have been very the latest cycle (�gures 11.5 and 11.6). high, whichever way one wants to measure them.14 In addition, the scale of the boom is a signi�cant predictor of the occurrence of LAC’s �nancial cycles have more often crises: bigger booms are more likely to end ended in crises badly. In particular, the likelihood that an Financial crashes have occurred more fre- upturn in real credit per capita will end up quently in LAC than in other regions (�g- in a � nancial crash is positively associated ure 11.1). For starters, the unconditional with the length and the size of the credit probability of a banking crisis is higher in upswing (table 11.2).15 Furthermore, the size Latin America (4.6 percent) than in indus- of the credit boom remains a good predictor trial countries or other emerging markets of financial crashes even after controlling (2.7 percent and 3.4 percent, respectively). for the size of the upswing in asset prices The frequency of banking crises following (stock prices and real exchange rates) and lending booms is also higher in LAC than in capital flows. Indeed, the magnitude of the 188 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC FIGURE 11.3 Behavior of credit during downturns in real economic activity: Are �nancial cycles leading real cycles? a. Credit per capita behavior during financial crisis versus noncrisis periods Middle-income countries Latin America and the Caribbean 0.10 0.10 0.05 0.05 0.00 0.00 growth rate (yoy) growth rate (yoy) –0.05 –0.05 –0.10 –0.10 –0.15 –0.15 –0.20 –0.20 –0.25 –0.25 8 7 6 5 4 3 2 1 8 7 6 5 4 3 2 1 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 T T T– T– T– T– T– T– T– T– T– T– T– T– T– T– T– T– T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ real GDP peak real GDP peak no financial crisis financial crisis b. Credit per capita behavior during the current cycle versus average historic cycles Middle-income countries Latin America and the Caribbean 0.20 0.20 0.15 0.15 0.10 0.10 growth rate (yoy) growth rate (yoy) –0.05 0.05 –0.00 0.00 –0.05 –0.05 –0.10 –0.10 –0.15 –0.15 8 7 6 5 4 3 2 1 8 7 6 5 4 3 2 1 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 T T T– T– T– T– T– T– T– T– T– T– T– T– T– T– T– T– T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ T+ real GDP peak real GDP peak previous cycles current cycle Source: Calderón and Servén 2011. Note: This �gure depicts the year-on-year growth rate (or variation) in credit and asset prices around 17 quarter windows centered around peaks in real GDP (T). The period T when the peak in real GDP takes place is identi�ed using the Bry-Boschan quarterly algorithm (Harding and Pagan 2002b). Panel a shows the evolution of credit during peaks in real output associated to episodes of banking crisis (as de�ned in Laeven and Valencia 2008). Panel b distinguishes peaks in real GDP during the current cycle (2006–10) and the average of previous output cycles. credit boom is the dominant crisis predictor; not simple. On the one hand, the duration, once it is considered, the occurrence (or the amplitude, and intensity of credit cycles in magnitude) of booms in other � nancial vari- the world, and particularly in Latin America, ables does not signi�cantly increase predic- have declined over time, especially during the tive power.16 past decade (figure 11.7). This worldwide How much of a guide for macropru- dampening might be attributed to what has dential policy design are LAC’s historical become known as the Great Moderation. patterns of � nancial cycles? The answer is But it was accentuated in LAC by the much MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 189 FIGURE 11.4 Unconditional and conditional probability of lending booms 0.16 0.14 0.12 0.10 probability 0.08 0.06 0.04 0.02 0.00 lending boom lending boom / lending boom / lending boom / equity price boom housing price boom capital flow bonanza event industrial countries non-LAC emerging markets LAC Source: Calderón and Servén 2011. Note: Unconditional probabilities are computed as the frequency of lending booms over the sample. The conditional probability of lending booms reports the frequency of booms in asset prices or capital flows occurring in t, t–1 or t–2 that end up in a banking crisis in period t. Finally, note that �nancial booms and capital flow bonanzas are de�ned as cyclical components of the corresponding �nancial variables exceeding 1.75 times their standard deviation. improved macroeconomic policy frame- and vulnerabilities.18 Much of the region works, lower debt burdens, and reduced is currently experiencing a double tailspin currency mismatches. This improvement in push coming from outside and resulting LAC’s macro� nancial “immune system� not from sharp worldwide cycle asynchronicity. only reduced the volatility of output over On the one hand, there are large increases time but also made the region more resil- in commodity prices, associated with high ient to global shocks.17 Given this signi�cant Asian growth, that are resulting in a terms of change, a simple extrapolation of frequen- trade bonanza. On the other hand, the con- cies and magnitudes of past cycles would not tinuously depressed economic conditions in be a sound basis for macroprudential policy industrial countries are producing very low going forward. On the other hand, however, world interest rates, which are in turn result- the importance of credit cycles relative to ing in a global liquidity fl ood. Such outside other cycles, as well as some of the sequenc- pressures are combining with the inside pres- ing results, are likely to survive as long as sures resulting from still stimulative home the basic � nancial structure does not change policies (both �scal and �nancial). As a result, too much. several countries in LAC are experiencing Be that as it may, the backward-looking unusually strong capital inflows. Partly as analysis of cycles needs to be complemented a result, some are in the midst of sustained by a more forward-looking survey of risks credit booms (for example, Brazil and Peru), 190 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC FIGURE 11.5 Real exchange rate behavior during downturns in real credit a. Real exchange rate behavior during �nancial crisis versus noncrisis periods Middle-income countries Latin America and the Caribbean 0.06 0.10 0.08 0.04 0.06 0.02 0.04 growth rate (yoy) growth rate (yoy) 0.02 0.00 0.00 –0.02 –0.02 –0.04 –0.04 –0.06 –0.06 –0.08 –0.08 –0.10 8 7 6 5 4 3 2 1 T 1 2 3 4 5 6 7 8 8 7 6 5 4 3 2 1 T 1 2 3 4 5 6 7 8 T– T– T– T– T– T– T– T– T+ T+ T+ T+ T+ T+ T+ T+ T– T– T– T– T– T– T– T– T+ T+ T+ T+ T+ T+ T+ T+ real GDP peak real GDP peak no banking crisis banking crisis b. Real exchange rate behavior during the current cycle versus average historic cycles Middle-income countries Latin America and the Caribbean 0.05 0.14 0.04 0.12 0.10 0.03 0.08 0.02 growth rate (yoy) growth rate (yoy) 0.06 0.01 0.04 0.00 0.02 –0.01 0.00 –0.02 –0.02 –0.03 –0.04 –0.04 –0.06 8 7 6 5 4 3 2 1 T 1 2 3 4 5 6 7 8 8 7 6 5 4 3 2 1 T 1 2 3 4 5 6 7 8 T– T– T– T– T– T– T– T– T+ T+ T+ T+ T+ T+ T+ T+ T– T– T– T– T– T– T– T– T+ T+ T+ T+ T+ T+ T+ T+ real GDP peak real GDP peak previous cycles current cycle Source: Calderón and Servén 2011. Note: This �gure depicts the year-on-year growth rate (or variation) in the real exchange rate around 17 quarter windows centered around peaks in real credit per capita (T). The period T when the peak in real credit per capita takes place is identi�ed using the Bry-Boschan quarterly algorithm (Harding and Pagan 2002a). Panel a shows the evolution of the real exchange rate during peaks in real credit associated to episodes of banking crisis (as de�ned in Laeven and Valencia 2008). Panel b distinguishes peaks in real credit during the current cycle (2006–10) and the average of previous credit cycles. others in the midst of incipient credit booms strong Asian economic pull calls for rais- (for example, Colombia). ing interest rates, yet this confl icts head-on As explained in chapter 6, the region’s with the low world rates imposed by the policy response has been much complicated depressed high-income countries. Managing by the world cycle asynchronisms. The this confl ict has been further hindered by MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 191 FIGURE 11.6 Foreign capital and credit downturns a. Behavior of gross capital inflows during �nancial crisis versus noncrisis periods Middle-income countries Latin America and the Caribbean 0.06 0.06 0.04 0.04 0.02 0.02 0.00 0.00 growth rate (yoy) growth rate (yoy) –0.02 –0.02 –0.04 –0.04 –0.06 –0.06 –0.08 –0.08 –0.10 –0.10 –0.12 –0.12 1 2 3 4 5 6 7 8 8 7 6 5 4 3 2 1 8 6 4 2 T 2 4 6 8 T T+ T+ T+ T+ T+ T+ T+ T+ T– T– T– T– T– T– T– T– T– T– T– T– T+ T+ T+ T+ real GDP peak real GDP peak no banking crisis banking crisis b. Behavior of gross capital inflows during the current cycle versus average historic cycles Middle-income countries Latin America and the Caribbean 0.09 0.09 0.06 0.06 0.03 0.03 0.00 growth rate (yoy) growth rate (yoy) 0.00 –0.03 –0.03 –0.06 –0.06 –0.09 –0.09 –0.12 –0.12 –0.15 –0.15 –0.18 –0.18 8 7 6 5 4 3 2 1 T 1 2 3 4 5 6 7 8 8 7 6 5 4 3 2 1 T 1 2 3 4 5 6 7 8 T– T– T– T– T– T– T– T– T+ T+ T+ T+ T+ T+ T+ T+ T– T– T– T– T– T– T– T– T+ T+ T+ T+ T+ T+ T+ real GDP peak T+ real GDP peak previous cycles current cycle Source: Calderón and Servén 2011. Note: This �gure depicts the year-on-year growth rate (or variation) in the ratio of gross inflows to GDP around 17 quarter windows centered around peaks in real credit per capita (T). The period T when the peak in real credit per capita takes place is identi�ed using the Bry-Boschan quarterly algorithm (Harding and Pagan 2002b). Panel a shows the evolution of the ratio of gross inflows to GDP during peaks in real credit associated to episodes of banking crisis (as de�ned in Laeven and Valencia 2008). Panel b distinguishes peaks in real credit during the current cycle (2006–10) and the average of previous credit cycles. the region’s relatively high interest rates— the large and mounting real exchange rate the lasting legacy of its turbulent past—and appreciations. Altogether, the rising � nan- (as shown in chapter 4) by its relatively more cial vulnerabilities and constraints faced by open capital account compared to other monetary policy make a very strong case emerging economies (particularly the Asian for quickly establishing or strengthening the economies). Thus, monetary policy has region’s macroprudential policy capacity. In struggled more than in other regions with turn, the latter will need to address several 192 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC TABLE 11.2 Size of �nancial booms and the probability of crisis: Probit analysis Sample: 79 countries, 1970q1–2010q4 Dependent variable: Financial crash [1] [2] [3] [4] Credit Real credit 1.1555** 1.3246 2.5876* 2.6228* (amplitude of preceding upturn) (0.518) (1.027) (1.364) (1.405) Capital flows Ratio of non-FDI inflows to GDP .. −0.1269 −1.2177 −1.3084 (amplitude of preceding upturn) (1.844) (2.306) (2.308) Asset prices Real exchange rate .. .. 2.1246 2.2336 (amplitude of preceding upturn) (2.208) (2.291) Stock prices (real) .. .. 0.1658 0.0984 (amplitude of preceding upturn) (0.523) (0.560) Leverage of the banking system Credit-deposit ratio .. .. .. −0.3075 (T-P amplitude) (0.528) Constant −1.4536** −1.3826** −1.8141** −1.7686** (0.163) (0.256) (0.345) (0.366) No. of observations 298 144 116 110 Log likelihood −108.4 −58.0 −38.3 −36.8 Pseudo R-squared 0.0623 0.0581 0.2084 0.193 Source: Calderón and Servén 2011. Note: Robust standard errors are shown in parentheses. Signi�cance level: * = 10 percent; ** = 5 percent; *** = 1 percent. key design or policy issues, as discussed in speci�c? Should the macroprudential policy the following section. tools be price based or quantity based? This section sheds some light on these questions by taking into account the dif�culties coun- Some key macroprudential tries have found in addressing these ques- policy design issues tions, and their experiences in doing so. The macroprudential perspective has moti- vated a number of regulatory proposals. Sweet spot versus overactivism Most of them involve adjusting regulatory instruments to the changes in risk over the In considering policies to manage systemic cycle, essentially tightening regulation in the risk over the cycle, one must first keep in upswing and relaxing it in the downswing. mind that the � nancial cycle in part reflects This approach raises a number of key ques- fundamentals that are themselves procyclical. tions that affect virtually all proposals. How Investment opportunities and credit demand ambitious should policy makers be? Should rise in the upswing, while the riskiness of macroprudential regulations seek to protect prospective borrowers declines. Indeed, the the � nancial system only by adjusting pru- Schumpeterian “creative destruction� that dential buffers over the cycle in the most ef�- characterizes market economies itself intro- cient way? Or should macroprudential regu- duces a real cyclicality that is intertwined lations go beyond, and also seek to dampen with �nancial fluctuations. Thus, part of the the cycle itself? Should the adjustment be financial cycle may reflect efficient adjust- based on rules, or left at the regulator’s dis- ments to changes in the underlying real fun- cretion? Should the trigger for macropruden- damentals (that is, real business cycle effects). tial adjustments be aggregate, or institution- Or it may reflect excess fluctuations that MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 193 are best taken care of through traditional FIGURE 11.7 Main features of the credit cycle over time monetary policy. Moreover, distinguishing between financial cycles and trends is dif- a. Average duration of the cycle of real credit per capita ficult, particularly in emerging economies 8 with lagging financial depth (LAC’s case), 6 as rapid � nancial expansion may just reflect 4 no. of quarters sustainable catch-up growth. Clearly, there- 2 0 fore, there is a risk of macroprudential policy –2 overkill. Macroprudential policy should not –4 seek to flatten the cycle altogether. It should –6 focus, rather, on � nding the “sweet spot�— –8 that is, containing only the “excess� volatility –10 caused by socially undesirable �nancial activ- 9 9 9 9 9 9 9 9 9 –8 –9 –0 –8 –9 –0 –8 –9 –0 80 90 00 80 90 00 80 90 00 ity (calming the excessively stormy seas) and 19 19 20 19 19 20 19 19 20 enhancing the resilience and stability of the industrial countries Latin America non-LAC emerging markets �nancial system (protecting the boat from the rough seas). Identifying this type of �nancial b. Median amplitude excesses is a tricky task. The fact that only a 0.15 0.10 small fraction of credit booms (about 1 in 10) 0.05 end up in crashes is a further warning against 0.00 the dangers of overactivism.19 amplitude –0.05 –0.10 Buffers or dampeners? –0.15 –0.20 As noted, prudential buffers (capital, generic –0.25 provisions, and liquidity) need to become –0.30 available to financial institutions in the 9 9 9 9 9 9 9 9 9 –8 –9 –0 –8 –9 –0 –8 –9 –0 80 90 00 80 90 00 80 90 00 downturn, when funding costs rise and their 19 19 20 19 19 20 19 19 20 industrial countries Latin America non-LAC lending portfolio and pro�tability deterio- emerging markets rate. Fixed minimum requirements (on capi- tal, loan loss provisions, and so on), however c. Median slope 0.04 high, are not helpful in this regard, since 0.03 they have to be met continuously through- 0.02 out the cycle. 20 Instead, the ability to draw 0.01 down without penalty during downswings 0.00 slope the buffers that were built in the upswing –0.01 should help both better shield the � nancial –0.02 system and dampen the amplitude of the –0.03 cycle. At the same time, by giving institu- –0.04 tions self-protection (capacity to absorb –0.05 9 9 9 9 9 9 9 9 9 losses), the ability to use the buffers in bad –8 –9 –0 –8 –9 –0 –8 –9 –0 80 90 00 80 90 00 80 90 00 19 19 20 19 19 20 19 19 20 times weakens the � re sale externalities (by industrial countries Latin America non-LAC reducing the need to sell assets). Moreover, emerging markets countercyclical buffers that are usable by the upturn downturn institution can also help in bad times by off- setting the procyclical deleveraging effect of Source: Calderón and Servén 2011. risk-weighted, Basel II–type, minimum capi- Note: This �gure reports the average duration of the different cyclical phases (downturns and tal requirements. 21 upturns) for real credit indicators per region and per period. The statistics for amplitude and slope refer to sample median across episodes. Countercyclical capital buffers have been amply advocated in both the academic and 194 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC the policy literatures. 22 However, there is at Another option is to contain the risk present only limited analytical research, and buildup itself and address the risk spillovers no actual experience, on their design and by constraining � nancial institutions’ lever- effectiveness. Thus, it is not clear whether age or short-term � nancing; or by directly the size recommended by the Basel Commit- targeting credit growth over the cycle tee on Banking Supervision (2.5 percent of through liquidity or reserve requirements. assets, see BCBS 2010a) would be adequate Thus, some have suggested introducing a to substantially contain deleveraging in the maximum, adjustable leverage ratio. 25 Like- downturn. Even less is known about their wise, different schemes have been suggested contribution to dampening aggregate vola- to penalize short-term wholesale funding tility. While a recent study by Angelini et al. (noncore liabilities) and encourage the hold- (2011) �nds that Basel III–style countercyclical ing of systemically liquid assets. 26 In recent capital would reduce gross domestic product years, many emerging countries (included (GDP) volatility by as much as 20 percent several LAC countries with inf lation- relative to the no-buffer scenario, much work targeting regimes, such as Brazil, Colombia, remains to be done to assess the robustness of and Peru) have resorted to discretionary these results. changes in banks’ reserve requirements to In contrast, countercyclical provisioning moderate credit cycles (�gure 11.8). 27 Their has seen actual use, notably in Spain, and primary impact on credit derives from the has more recently become popular in several fact that reserve requirements are unremu- LAC and Asian countries. 23 Although the nerated; hence, they introduce an implicit objective is similar to that of countercyclical tax that widens the loan-deposit interest capital buffers, the main difference is that rate differential. 28 In practice, however, the provisions have been traditionally seen as cost effectiveness of reserve requirements buffers against risks that have already been in dampening the amplitude of the lending recognized (expected losses), while capital is cycle (relative to other potential macropru- intended as a buffer against risks that have dential instruments such as countercyclical not been recognized (unexpected losses). capital requirements) remains to be fully The underlying logic is that credit risk is demonstrated.29 incurred during expansions, when credit portfolios are being built up, even though Rules or discretion? the expected losses the provisions intend to cover have not yet been identi�ed in speci�c The limited experience with most propos- loans. By giving banks incentives to extend als on countercyclical regulation, the fact loans more carefully in the upswing and that many of them remain untested, and the by limiting the growth in bank pro�ts and role of mood swings in � nancial dynamics capital, dynamic provisioning should help all suggest the need for regulatory discre- restrain credit expansion. However, its pri- tion. Indeed, rules are more congenial to mary purpose is not to prevent credit booms a world of rational expectations than to per se, a task that might require prohibitively one of uncertainty and bounded rational- high provisioning rates. Thus, while most ity. However, the more discretion the regu- observers agree that the Spanish dynamic latory authorities exercise, the higher the provisioning scheme helped banks weather political hurdles. Thus, the cooling down the global downturn better, there is much of the boom might be too little and come less agreement as regards its impact on the too late. Indeed, it is important to recall cycle. 24 In retrospect, it was clearly insuf�- that countercyclical regulation would cient to tame the lending cycle or to prevent have been possible in many countries in the boom in real estate prices. Nonetheless, the run-up to the global financial crisis, it may also be argued that, without it, the yet very few made use of them. Hence, credit boom and real estate bubble would all things considered, a “set it and forget have been even bigger. it� countercyclical regulation based on MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 195 well-defined rules is more likely to help FIGURE 11.8 Reserve requirements and reference rates deflect pressures on the regulator. But it is doubtful that rules-based countercyclical a. Brazil prudential norms would suf�ce, and a sig- 20 60 ni� cant scope for discretion would likely need to be built into any robust macro- 15 45 prudential policy framework, an issue to which chapter 13 returns. percent percent 10 30 Regardless of whether rules or discretion predominate, matters are complicated by the lack of well-de�ned, easy-to-agree-on macro- 5 15 prudential policy levers as well as indicators to which the changes in that policy can be 0 0 automatically tied. In contrast to monetary 06 07 08 09 10 11 20 20 20 20 20 20 policy whose objective (price stability) can reference interest rate demand deposit be represented by a well-measured sum- savings deposits time deposits mary index, � nancial stability is an elusive b. Colombia and multidimensional concept that is much 12 40 harder to measure and monitor. Moreover, the trigger signal should be timely (ahead 9 30 of the buildup of financial imbalances), percent percent should become available quickly, and should 6 20 offer an early indication of cyclical turning points, particularly at the beginning of the 3 10 downturn. A variety of indicators have been suggested for this purpose, ranging from GDP growth to the rate of growth of overall 0 0 06 07 08 09 10 11 credit, the rate of growth of institution-spe- 20 20 20 20 20 20 ci�c credit, the credit-to-GDP ratio, or asset reference interest rate saving accounts prices (particularly housing). 30 While they current accounts all show a considerable degree of comove- 12 c. Peru 40 ment over the cycle, blindly following one or the other can result in very different paths 9 30 of regulatory tightness. 31 Unavoidably, a variety of � nancial and real indicators thus percent percent would have to be used, as in the case of 6 20 monetary policy, to get a more nuanced per- spective and reduce the risk of doing more 3 10 harm than good. A further difficulty, as noted, concerns 0 0 the distinction between trends and cycles. 06 07 08 09 10 11 20 20 20 20 20 20 Sustained increases in credit might reflect reference interest rate average healthy financial development (the bright side), rather than an unsustainable cyclical boom (the dark side). Thus, mechanically Source: Calderón and Servén 2011. gearing regulatory stance to changes in these variables could have the unintended result of Speci�c or broad? retarding � nancial deepening. This problem is even trickier in LAC because the region The question of whether macroprudential still has quite a way to go before it catches up instruments should be specific or broad- with its benchmark. based has multiple facets. A first facet is 196 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC whether the changes in regulatory stance of risk, and in virtually all cases they have over the cycle should be top-down (guided relied on judgment. On the whole, there is by institution-specific indicators) or bot- little systematic evidence on the extent to tom-up (guided by systemwide indicators). which they have been effective at restraining Both alternatives have been defended in the housing booms, although there is some anec- literature, and both have been applied in dotal evidence that they may have increased practice. 32 There is, in addition, the ques- banks’ resilience to falling property prices. tion of whether, once triggered, the macro- Beyond questions on their effectiveness, such prudential norm should be applied equally sector-speci�c interventions are not devoid of across the board or differentiated accord- other problems. First, they are hard to cali- ing to institutions’ speci� c characteristics. brate.36 Second, their practical effectiveness Across-the-board application may have may be limited because the more selective greater dampening effect on the cycle but, the intervention is, the more vulnerable it is by punishing the more prudent lenders, it to circumvention.37 Third, the more selective could undermine individual discipline and the measures, the larger the risk of straying encourage risky behavior. In addition, con- into the type of distortionary credit alloca- ditioning the application of the norm on tion policies that characterized the � nancial institution-speci�c indicators of risk buildup repression epoch a few decades ago. is arguably the right approach if one aims at internalizing externalities. Price based or quantity based? A second facet is whether the norms should aim at containing all types of lending A last but not less important issue is whether or should focus on speci�c types of lending the countercyclical prudential norms should and risks. A particularly relevant example be price based or quantity based. Liquid- is the housing sector. Concern with this sec- ity regulations, for example, can take the tor is well justi�ed since housing prices are form of quantity constraints, such as mini- closely synchronized with credit cycles. Thus, mum liquidity requirements, or price-based a �xed ratio of the size of the loan to the mar- norms, such as Pigouvian taxes that discour- ket value of the real estate (the loan-to-value age short-term (noncore deposit) funding. ratio, or LTV) in mortgage lending standards The more price based the norm, the easier can be an important source of procyclicality. it is to apply across the board—and the As housing prices rise in the boom, a given broader its application, the lower the risk of LTV allows credit to expand more in the distortions and regulatory arbitrage; how- upswing, further fueling the asset price spi- ever, the harder it also becomes to properly ral. The higher the LTV and the more com- calibrate it. Indeed, a case can be made that petition there is between lenders, the bigger the higher the uncertainty about the impact this effect. 33 A similar reasoning applies to of taxes and the more pressing the need for the cycle-independent ratio of debt service immediate results, the stronger the com- to income (DTI), owing to the procyclical parative advantage of quantity-based regu- behavior of personal income. lation. 38 In less mature financial systems, Thus, a variety of tools have been deployed the relative advantage of quantity-based in emerging countries in recent years at times regulation is arguably stronger because it of rapidly growing property lending and hous- would carry a lower risk of distortion and ing prices. Among those tools, caps on LTV the impact of price-based regulation would ratios have seen considerable use, especially be more uncertain. in Asia.34 Limits on DTI ratios have likewise been applied in several countries to help con- tain lending to lower-quality borrowers. 35 The path ahead In many cases, these adjustments have been In de� ning the path ahead, the region will only loosely related to quantitative indicators need to keep in mind three basic points: MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 197 (a) there is a particularly strong (and, in sev- be the � rst order of business no matter what. eral countries, urgent) case for macropru- If that approach does not help, at least it dential policy in the region, (b) yet there is won’t hurt. In fact, it is likely to help achieve much uncertainty as to the effectiveness of the higher-order goals—certainly the buff- macroprudential policy tools (which puts a ering goal but also the dampening goal. premium on research), and (c) the deploy- The second order of business should be to ment of macroprudential regulatory norms enhance the stability of the � nancial system is not devoid of costs. In the end, it will all through better and stronger (liquidity and boil down to � nding the sweet spot at which solvency) buffers. This does not mean, how- the gains are clear and the norms address ever, that one should give up altogether at the areas where the bright side of � nancial this stage on the objective of dampening the development begins to interface with the cycle. As demonstrated by the use of reserve dark side. This is even harder because some requirements, loan-to-value ratio, and debt- of the costs of macroprudential policy may service-to-income ratio, a case can also be be visible and materialize immediately (for made for cautious yet � rm experimentation example, the immediate reduction in the of instruments aimed at this latter goal. credit available to the marginal but viable However, the goals should generally inter- borrowers), whereas others (the distortions, act with the means. Thus, because a judg- regulatory arbitrage, and overall possible ment is likely to be necessary when assessing slowdown in the pace of sustainable � nan- � nancial stability and the need for interven- cial development) may only become evident tion, a purely rules-based scheme of the type in the longer run. Nor will the bene�ts of proposed by Basel III might be more feasible macroprudential policy, in the form of if used strictly as a buffer rather than a damp- reduced incidence of crises, become imme- ener. Instead, if one really wishes to affect the diately obvious.39 cycle, doing it through a purely rules-based In this context, three important criteria scheme is most likely a pie in the sky. If it for macroprudential policy design and imple- cannot be done for monetary policy (no cen- mentation come to mind: (a) the overall strat- tral bank in the world has ever been able to egy needs to be organized in accordance with blindly follow a Taylor rule based on output the four lines of attack identi�ed earlier; (b) gaps, or the Friedman rule of a constant rate the institutional framework that is put in of money supply growth), the odds against place should promote good judgment, trans- doing it for macroprudential policy seem parency and accountability, and organized overwhelming. The scheme would survive decision making; and (c) all the eggs should only if it were bland enough, hence unlikely not be put in the same basket (macropru- to have much impact one way or the other. In dential policy is not a magic bullet) and the practice, one is thus probably left with a mix strategy should emphasize complementarities of rules and discretion, the former aimed at with other policies. The second of these items buffering the system, the latter at affecting its (which lies at the interface between regulation path over the cycle. and supervision) is dealt with in chapter 13. Another example of interaction between The other two are discussed next. ends and means can be applied to the choice of speci�c versus broad-based instruments. As noted above, if one mainly aims (at least The lines of attack in a �rst phase) at dampening the cycle rather The uncertainties described above militate than internalizing externalities, a better case in favor of a gradualist approach linked to might be made for a simpler instrument that the mentioned lines of attack. Thus, elimi- applies uniformly across �nancial intermedi- nating as much as possible the procyclicality aries. The objective of inducing the internal- and inconsistencies inherent in the precrisis ization of externalities could then be pursued Basel-style prudential norms would seem to through Pigouvian taxes that aim to address 198 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC systemic risk. They will do so by focusing Fiscal policy also has a key role to play, on the wedges between private and social however. It can help mitigate the amplitude of interests behind both excessive cyclical fluc- the � nancial cycle and align the private and tuations and on the dark side of interconnect- social incentives for risk taking. Indeed, the edness, as discussed in chapter 12. stabilizing potential of countercyclical �scal policy has been underscored by the global cri- sis. The accumulation of �scal buffers during The macro policy mix the good times allowed a number of emerg- Macroprudential policy tools are not the ing countries, notably in Latin America, to only ones available to policy makers for adopt an expansionary stance when the cri- managing systemic risk. Monetary and �scal sis hit.45 Yet, discretionary �scal policy often policies (and possibly capital controls) also faces considerable delays, which underscores have a potentially important impact on the the need to build up self-deploying automatic � nancial cycle and the buildup of aggregate stabilizers, particularly in LAC.46 In addition, � nancial risk. Indeed, the global crisis has �scal policy can play a key role in internal- revived interest in the powers of monetary izing externalities through Pigouvian taxes policy to aid � nancial stability. In addition (thereby putting into place the last and more to restraining credit through higher inter- ambitious perimeter of defense).47 However, est rates, monetary policy can also affect in spite of the solid theoretical justi�cation of the quality of lending through the so-called this kind of taxes, little is known about their risk-taking channel.40 Unduly low monetary likely effectiveness in practice, how tax rates policy rates lead banks to expand their lend- should be set, and to what extent they should ing to higher-risk borrowers. Empirical evi- be varied over the cycle. Thus, a simpler and dence tends to support this view.41 more immediate step in the same direction Before the global crisis, a consensus view would be the removal of �scal incentives that seemed to have emerged in favor of “the Tin- favor debt � nancing over equity � nancing. bergen principle� of separating monetary Limiting the incentives for leveraging could stability from financial stability. Targeting have a major effect on the resilience of the asset prices (beyond what was needed to �nancial system.48 target inflation) was viewed as impossible Another policy area where some action or, if possible, counterproductive.42 How- might be needed to complement macropru- ever, an alternative view is now emerging in dential policy is capital controls. As shown which monetary policy can dampen boom- in the above section (“LAC’s cycles and vul- bust episodes if it reacts to credit growth, nerabilities�), capital flows and credit growth asset prices, leverage, or other indicators of show a significant degree of comovement � nancial risk.43 This means that monetary over the cycle, and �nancial crises in emerg- policy should become more restrictive dur- ing markets have been frequently preceded by ing a credit and asset price boom and more capital inflow booms. In particular, capital responsive to indicators of � nancial stabil- controls may be used to change at least the ity. However, how to manage in practice the composition, if not volume, of inflows toward possible conflicts between price stability and less risky forms.49 However, though there � nancial stability without undermining the is a massive literature on capital controls, credibility of central banks is at this point a there are few robust �ndings as regards their largely unresolved issue. What is clear, how- effectiveness, largely owing to the almost ever, is that monetary and macroprudential insurmountable dif�culty of establishing the policies should be viewed as complements, proper counterfactual scenario. 50 Perhaps not substitutes. This view has prompted calls one of the few conclusions on which every- for central banks to assume both tasks in a one agrees is that capital controls quickly fully integrated manner and using more than develop leakages, all the more so in the more one instrument.44 advanced and globally integrated financial MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 199 markets. Finally, it is important to stress that Latin America� by César Calderón and Luis the aggregate vulnerability of most emerging Servén (2011), which is part of the Edited economies to sudden capital flow reversals Volume that accompanies this LAC flagship has declined substantially in recent years, report. 3. For a recent excellent general review of the lit- owing in particular to the attainments of net erature on macroprudential policy, see Galati creditor positions in debt contracts as well as and Moessner (2011). Ostry et al. (2011) to reduced �nancial dollarization.51 review the closely related issue of managing A last area where good coordination with capital inflows. Terrier et al. (2011) offer a macroprudential policy is also required is very detailed analysis of macroprudential pol- as regards the exchange rate regime. The icies and instruments in the Latin American region’s transition to floating-rate regimes has context. enhanced its macroprudential policy inde- 4. See Lorenzoni (2008) and Jeanne and Korinek pendence at the same time as its monetary (2011). An excessive expansion of credit and policy independence. A floating-rate regime leverage can also be fueled by moral hazard. naturally creates a wedge between the foreign For instance, when an innovation opens new opportunities (the upside widens) or a macro- and the local currency. The more volatile the systemic shock wipes out a large part of the exchange rate, the higher the risk of invest- �nancial intermediaries’ capital (the down- ing or borrowing in foreign exchange for side shrinks), agents may take increasing risk local residents. Hence, floating-rate regimes because they have increasingly little to lose not only should (and indeed have, as shown relative to what they can gain. in chapter 3) reduce the dollarization of local 5. See Dell’Ariccia and Marquez (2008) and financial intermediation, but also, by the Adrian and Shin (2010). same token, should limit the scope for off- 6. All in all, the mood swings story is more shoring by reducing the substitutability of rounded and self-contained than the �nancial local and foreign currency loans. The lower ampli�cation story. However, it is also more ad substitutability should reduce the scope for hoc. It �nds its roots in Keynes and Minsky and is generally consistent (albeit somewhat loosely) regulatory arbitrage across borders, thereby with the behavioral �nance literature. Recent enhancing the scope for a more active macro- formalization attempts can be found, for exam- prudential home policy, even when the latter ple, in de Grauwe (2009) and Lo (2009). is asynchronous with that of rest of the world. 7. As an illustration, one can consider a surge However, to the extent that macroprudential in capital inflows that are not driven by fun- policy can help limit exchange rate volatility damentals. They can lead to an “excessive� by assisting monetary policy (for example, appreciation of the currency (unnecessary loss by limiting the need for monetary tight- of competitiveness), which is a relevant mac- ening, hence, exchange rate appreciations roeconomic policy concern, as well as lead to during the booms), some underlying tensions the buildup of systemic risk in the �nancial will need to be managed. A well-integrated sector, which is a relevant macroprudential policy concern. macroeconomic policy is therefore again of 8. The sample consists of 23 industrial countries the essence. and 56 developing countries, of which 15 belong to LAC. Notes 9. The reader should be cautious with the inter- pretation of the housing price dynamics in 1. The postcrisis discussion of the required LAC because the information is sparser than reforms to systemic oversight generally for industrial and other emerging-market includes microsystemic aspects within the economies. The lack of availability of a hous- broader category of macroprudential reforms. ing price index for the countries in the region However, a terminological distinction is made restricts the LAC sample to Chile, Colombia, in this flagship report because it helps orga- Mexico, and Uruguay. nize the discussion more neatly. 10. Cycles ending up in a banking crisis are de�ned 2. This chapter draws heavily on the paper as those whose peak-to-trough phase coincide “Macro-Prudential Policies over the Cycle in 200 MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC or are followed by a banking crisis within two multilateral institutions and think tanks. See, years. Episodes of banking crisis are de�ned for instance, Eyzaguirre et al. (2011), de La as in Laeven and Valencia (2008), that is, as Torre et al. (2011), Izquierdo and Talvi (2011), a situation in which the following four condi- and Cardenas and Levy-Yeyati (2011). tions are met: (a) rising nonperforming loans 19. This leaves open the question, however, as exhaust banks’ capital, (b) asset prices col- to whether it is socially desirable to mod- lapse on the heels of run-ups before the crisis, erate cycles even when they do not end up (c) real interest rates are sharply raised, and in crashes. This question probably has two (d) there is a large reversal or slowdown in answers. First, from an ex ante perspective, capital flows. what matters is to contain the risk of an 11. The concordance indexes reported in �gure eventual crash down the road. If one can- 11.2 measure the proportion of time that two not a priori tell which cycles will end badly, series share the same cyclical phase (that is, then all cycles should call for a precaution- it measures the strength of the contempora- ary response, which puts more weight on the neous comovement). However, they do not downside. Second, �nancial cycles may have provide information on whether �nancial social costs (for example, an over- or under- cycles tend to precede real cycles. An event extension of credit) even when they do not study analysis was used to evaluate statistical threaten �nancial stability. precedence. 20. See Goodhart (2010b) and Hellwig (2010). 12. Whether peaks in real credit or asset prices 21. See Gordy and Howells (2006); Repullo and precede peaks in real output are detected Suárez (2009); and Repullo, Saurina, and based on regressions of (year-on-year) growth Trucharte (2010). in credit on a 17-quarter window centered on 22. See Kashyap and Stein (2004); Hanson, the peak in real GDP (T). The period from Kashyap, and Stein (2010); Shleifer and T 8 to T represents the run-up to the down- Vishny (2011), Brunnermeier et al. (2009), turn in real economic activity. On average, and Goodhart (2011). the period from T to T 4 may capture the 23. Countercyclical provisioning schemes have downturn in real output whereas that after been introduced under various forms in T 4 may capture the start of the recovery Bolivia, Colombia, Peru, and Uruguay. The period. Figure 11.3 plots the estimated coef- systems vary considerably in terms of design. �cients of these regressions. See Fernández de Lis and García-Herrero 13. An analogous regression analysis is conducted (2010) for a comparative analysis of the cases just as the one described above but with T rep- of Colombia and Peru. Wezel (2010) provides resenting the peak in real credit per capita. a detailed description of the Uruguayan sys- 14. Chapter 6 provides estimates of the long-run tem. Galindo and Rojas-Suarez (2011) and costs of crises in terms of �nancial disinterme- Terrier et al. (2011) offer broad overviews of diation. As regards the �scal cost of resolving these systems. banking crises, the estimate provided for LAC 24. See Saurina (2011). by Laeven and Valencia (2008) is an average 25. See Goodhart (2010a). of 16 percent of GDP. 26. See Brunnermeier et al. (2009) and Shin and 15. Similar results are found by Barajas, Shin (2011). Dell’Ariccia, and Levchenko (2009). 27. Compared to (interest earning) liquidity 16. Similar results are found by Schularick and requirements, reserve requirements are an Taylor (2009) and Jorda, Schularick, and inef�cient prudential way to buffer banks’ bal- Taylor (2010). ance sheets against liquidity shocks. Reserve 17. The arguably structural improvement in requirements can also be used as a monetary LAC’s macro�nancial policy frameworks over policy instrument (to mop up money supply), the past decades has been amply discussed in although less ef�ciently than open market policy and academic circles, particularly in operations. Reserve requirements can more view of LAC’s strong performance during the rightfully claim to be a macroprudential tool recent global crisis. See, for example, de la when their monetary impact is fully offset by Torre et al. (2010a, 2010b, 2011). open market operations. Even then, however, 18. More detailed and recently updated accounts it is often not clear if the countercyclical use of LAC’s macro�nancial risks and vulnerabili- of the reserve requirements reflects macropru- ties can be found in regional reports by the dential concerns or is rather a cheaper way of MACROPRUDENTIAL POLICIES OVER THE CYCLE IN LAC 201 �nancing the accumulation of international 37. See Park (2011). reserves resulting from foreign exchange inter- 38. Weitzman (1974) provides an early version vention. In this latter case, reserve requirements of this argument, which has been recently can be seen as an attempt by the monetary applied to the case of prudential regulation authorities to defeat the “impossible trinity�— by Haldane (2010b), among others. The rela- that is, to pursue an independent monetary tive merits of prices versus quantities for the policy, along with exchange rate targets, in a case of liquidity norms have been recently context of open capital accounts. On this latter analyzed by Perotti and Suárez (2011). point, see Montoro and Moreno (2011). 39. See Viñals (2011). 28. If market �nancing or central bank loans are 40. The term was coined by Borio and Zhu not perfect substitutes for deposit �nancing, (2008). It was more recently analyzed by the introduction of reserve requirements can Adrian and Shin (2011). also reduce the supply of loans by subjecting 41. See Jiménez et al. (2008) for the case of Spain; banks to liquidity and interest rate risk. Vargas Ioannidou, Ongena, and Peydró (2009) for et al. (2010) claim some empirical support for Bolivia; and Delis and Kouretas (2011) for this mechanism in the case of Colombia. the euro zone. 29. Montoro and Tovar (2010) present a gen- 42. See Bernanke and Gertler (1995) and eral equilibrium model in which reserve Svensson (2010). requirements help stabilize the business cycle 43. See Bordo and Jeanne (2002a, 2002b), in the face of demand shocks but not sup- Christiano et al. (2010), and Woodford (2011) ply shocks. The Peruvian authorities report for a dissenting view. that a 1 percentage point increase in reserve 44. See for example Mishkin (2011) and Claessens requirements affects the output gap as much et al. (2010). as a 25 basis points increase in interest rates. 45. The empirical evidence shows that this In Brazil, Carvalho and Azevedo (2008) �nd stance helped reduce considerably the real that reserve requirements affected banks’ cost of the crash. See Didier, Hevia, and shareholders by affecting bank pro�ts. Schmukler (2011) and Corbo and Schmidt- 30. See Drehmann et al. (2010). Hebbel (2010). 31. For example, using data for industrial coun- 46. See Debrun and Kapoor (2010) and Claessens tries, Repullo and Saurina (2011) �nd that et al. (2010). the credit-to-GDP ratio often behaves coun- 47. The literature in this respect is in full bloom tercyclically, and that in some countries it and includes models that connect Pigouvian lags the cycle rather than leading it. Also, as taxation to credit booms (Jeanne and noted by Goodhart (2011), it is not clear if Korinek 2011), maturity mismatches (Perotti the empirical regularities that favor one trig- and Suárez 2011) or noncore liabilities (Shin ger over another would survive once regula- and Shin 2011). Angeletos, Lorenzoni, and tion is allowed to change over the cycle, as Pavan (2010) develop a model of asset price this might induce behavioral changes on the booms under information externalities in part of �nancial intermediaries. which procyclical asset taxes can improve 32. See for example Goodhart and Persaud welfare by narrowing the gap between (2008) and Drehmann et al. (2010). market-determined prices of assets and their 33. See Borio, Fur�ne, and Lowe (2001). fundamental values. 34. Two examples are Hong Kong SAR, China, 48. See Hellwig (2010) and Goodhart (2011). and, more recently, China. 49. Capital controls may also be of use when cap- 35. These adjustments have sometimes been ital inflows bypass regulated �nancial institu- accompanied by other ad hoc measures, such tions and directly accrue to the non�nancial as increases in risk weights applied to property private sector, thereby indirectly putting the lending for the calculation of regulatory capital local �nancial system at risk. (a step taken, for example, in India in 2005), as 50. See Demirgüç-Kunt and Servén (2010) and well as by direct controls such as credit ceilings, Ostry et al. (2011) for references. actively employed in some countries (notably 51. Calvo, Izquierdo, and Talvi (2003) �nd China) to restrict bank lending to housing. empirical evidence supporting the hypothesis 36. Thus, the Republic of Korea has seen a sharp that liability dollarization raises vulnerabil- reversal in its housing market following a ity to sudden stops and reversals in capital recent tightening in LTV policy. flows. Microsystemic Regulation 12 T he previous chapter dealt with the egating the oversight of the unregulated to dynamic dimensions of systemic the regulated or to auxiliary institutions— oversight—that is, how to control might become desirable in the future. the evolution of the system through time. • LAC’s current silo-based approach to This chapter focuses on the cross-sectional setting the regulatory perimeter—where dimension of systemic oversight—that is, regulation is linked to the license, that is, how to use microsystemic policies to address to the type of institution, such as commer- the spillovers and externalities infl icted by cial banking, insurance, investment bank- individual agents and institutions on the rest ing, or asset management—which also of the system—as that dimension relates to has served the region well thus far, will Latin America and the Caribbean (LAC).1 also need some revisions to limit regula- The chapter � rst looks into the issue of how tory arbitrage as well as to achieve better far the (prudential) regulatory net should economies of scope. extend and how uniform it should be. It then • At the same time, the oversight of � nan- explores three key facets of cross- sectional cial conglomerates—an inherent challenge systemic regulation, namely, systemically for silo-based regulation—will need to be important � nancial institutions (SIFIs), sys- strengthened considerably, both across temic liquidity, and financial innovation. domestic silos and across borders. Key highlights of the chapter are as follows: • These improvements will need to be matched with the establishment of effec- • While LAC’s current approach of pushing tive frameworks to deal with the failure the outside perimeter of prudential regu- of � nancial conglomerates and individual lation as far away as possible has served SIFIs. the region well thus far, it is likely to face • As with other regions around the world, the increasing challenges (including high region will need to review its strictly pru- demands on the already stretched human dential liquidity norms—many of which and �nancial resources of regulatory agen- are currently set from a purely idiosyn- cies) as � nancial systems mature. cratic (maturity mismatch) perspective—to • Thus, carefully designed policies that make them consistent with a systemic reduce the oversight load—possibly by del- perspective. 203 204 MICROSYSTEMIC REGULATION • The region will need to revamp its liquid- conduct� regulation; the latter underlies ity regulations to include a more systemic the bulk of prudential regulation, a distinc- perspective and review its systemic liquid- tion on which the “twin peak� approach ity access policies to better meet the chal- to regulation is typically based. 3 Market lenges of the future. This aim will most conduct regulation focuses on market integ- likely require a revision of the criteria for rity and consumer or investor protection access to the central bank liquidity facili- issues and comes primarily in the form of ties, as well as of the terms. disclosure requirements, accounting stan- • Fi nal ly, wh i le t he reg ion’s c u r rent dards, conduct-of-business rules (includ- hands-on approach to the regulation of ing anti-money-laundering regulation), and financial innovation has also served it governance and � duciary responsibilities. well thus far, that approach may need to Prudential regulation, by contrast, focuses be revisited to allow the forces of innova- on the safety and soundness of individual tion to more fully join the bright side of � nancial institutions and the stability of the � nancial development. � nancial system. To the � rst two justi�cations for pruden- The rest of this chapter is organized as tial regulation—asymmetric information follows. The � rst section briefly reviews the and collective action—one must add the pos- rationale for regulation and the flaws of the sible need for regulation aimed at addressing approach that prevailed prior to the global potential market failures under the collective �nancial crisis. The second and third sections cognition paradigm or curbing market power discuss the outer and inner boundaries of (a potential fifth finance paradigm). The regulation, respectively. The next three sec- former may justify the (arguably polemical) tions discuss the regulation of SIFIs, systemic regulation of �nancial innovation.4 The latter liquidity, and innovation, respectively. The can justify the regulation of anticompetitive �nal section concludes. behavior. Traditionally, competition-oriented regulation was not seen as connected to, or as The rationale for prudential a part of, prudential regulation. Indeed, less competition was generally viewed by pruden- regulation tial regulators as promoting �nancial stabil- The case for � nancial sector regulation can ity, rather than undermining it. However, the be readily connected to the finance para- global crisis has brought a new twist to this digms depicted in chapter 2. The “asym- debate, in the form of the possible need to metric information� paradigm justi�es the regulate SIFIs differentially. Hence, in addi- need for prudential regulation aimed at pro- tion to the traditional antitrust policy goals tecting the uniformed (or unsophisticated). 2 of curbing market power, excess pricing, Collective action failures (the collective and monopoly rents, a prudential regulation action paradigm) justify the need for pru- rationale emerges—to control the negative dential regulation aimed at protecting the spillover effects and moral hazard associated interests of the community (or, equivalently, with broad interconnectedness and the too- of the � nancial system taken as a whole). big-to-fail (TBTF) syndrome. Public policy can take two forms, one The bene�ts of regulation need to be bal- aimed at facilitating � nancial participation anced with its costs, however. Besides the (inclusion along the intensive and extensive cost of official oversight itself, there is the margins) through uniform rules of trans- opportunity cost of the �nancial activity that parency and disclosure (a public good), and may not take place because of regulation, the the other aimed at internalizing spillovers distortions of regulatory arbitrage, and the and externalities infl icted by individual par- risk that of�cial oversight may be perceived ticipants on the system. The former under- as a seal of approval promoting moral haz- lies what is generally known as “market ard and leading to the socialization of risks. MICROSYSTEMIC REGULATION 205 These costs of regulation have traditionally boundary effects are nonetheless already led to drawing a line in the sand, a regulatory apparent. They are expected to become grad- perimeter that separates the fully regulated ually more intense as � nancial systems con- deposit-taking banks from the unregulated tinue to evolve. Hence, the question of where (or lightly regulated) other � nancial entities to set the outer boundary of prudential reg- (including investment banks). This line-in- ulation applies to LAC as much as to other the-sand approach rested on two premises regions of the world. that have been deeply questioned by the recent subprime crisis: first, that the well- informed investors would exert effective mar- The outer boundaries: ket discipline on the market-funded �nancial Illuminating the shadows institutions, and second, that buffering and There are many options as to where to set ring-fencing of the deposit-taking institutions the perimeter of prudential regulation. Some through prudential oversight would protect of these, ranked by descending order of the “core� from systemic risk. comprehensiveness, include regulating (a) Consistent with this traditional line- all leveraged � nancial intermediaries (except in-the-sand approach, there was a general the very small); (b) all leveraged financial trend toward less intrusive supervision or intermediaries except those that borrow outright deregulation (at least the absence of only from regulated institutions; and (c) all new regulation) for the unregulated finan- commercial banks plus any other leveraged cial entities, instruments, and markets that financial intermediary that is considered were rapidly growing at the outer edges of to be a SIFI. 5 The third option is the one the commercial banking core. However, as implicitly embedded in the U.S. financial the subprime-induced global financial cri- reform law (box 12.2). It would involve set- sis amply demonstrated, this approach was ting the severity of regulation and intensity frankly inconsistent with the control of sys- of supervision based on an assessment of temic risk (de la Torre and Ize 2010). The the risk that a particular institution poses line in the sand promoted massive regulatory to overall � nancial stability. To limit exac- arbitrage, as differential returns encouraged erbating regulatory arbitrage (the so-called investors (informed or not) to leave the world boundary effects), the norms could in prin- of the regulated and join that of less regu- ciple be based on a continuous risk scalar lated, highly leveraged, and short-term-funded that avoids any major jump at the border intermediaries where systemic risk was totally between the SIFIs and the other institutions. uninternalized. The exponential growth of However, implementing such a smooth shadow banking and rapidly rising intercon- transition is likely to be quite difficult in nectedness of � nancial entities, regulated or practice, because openly naming some insti- not, spread systemic risk, putting in place the tutions SIFIs gives rise anyway to an abrupt conditions for the global crisis. border line, with institutions on one side of While these developments were especially the border (the SIFIs) implicitly bene�tting evident in the United States, the underlying from a publicly recognized TBTF advantage. issues are universal. For example, in LAC While one could try to eliminate this advan- they came to light in full bloom in the case of tage by imposing a prudential surcharge for the Mexican Sofoles (box 12.1). At the same crossing the border (that is, for becoming time, as noted in chapter 3, LAC’s � nancial a SIFI), getting it right could be devilishly deepening is taking place to an extent out- hard. The problem may be compounded by side the domain of deposit-taking institu- the fact that this is a dynamically evolving tions. While much of this expansion thus problem and that moving institutions across far appears to have accompanied a gradual borders is bound to have important, possibly broadening of the demand for � nancial ser- destabilizing, signaling effects. Hence, there vices, as opposed to regulatory arbitrage, is a risk that the two groups—SIFIs and 206 MICROSYSTEMIC REGULATION BOX 12.1 The Mexican Sofoles The Sofoles (Sociedades Financieras de Objeto eventually led to an explicit and complete prudential Limitado) are limited-purpose, non-deposit-taking deregulation of the Sofomes (Sociedades Financieras institutions with activities in a variety of sectors, de Objeto Múltiple), a newly created multipurpose including mortgages (and the associated � nancing and more flexible license meant to replace that of the of construction), consumer and small and medium Sofoles. The Sofomes were subject only to licensing enterprise (SME) lending, microcredits, and loans to requirements, not to prudential regulation. the agriculture and education sectors. In some sectors However, the Sofomes were hard hit by the (particularly mortgages), a large part of their funding global financial crisis. Some of them had started is through public development banks. In others (such to depend heavily on the issue of short-term com- as the automotive sector), their � nancing is fully pri- mercial papers, with much of the proceeds invested vate. The Sofoles were created in the context of the in risky and illiquid real estate investments. Under North American Free Trade Agreement (NAFTA) to the favorable assessments of local rating agencies, allow the similarly unregulated U.S. � nance institu- the papers were gobbled up by a variety of inves- tions to operate freely in Mexico. They grew rapidly tors, local and foreign. However, the subprime crisis after the 1995 Tequila crisis, when banks stopped brought the system to an abrupt halt and triggered lending altogether, reflecting public development a run of Sofoles debt. A relatively large mortgage policies as well as the banks’ capacity to rapidly provider, Sofol (Hipotecaria Crédito y Casa SA) col- detect and exploit pro�table market niches thanks to lapsed owing to soaring bad loans and mounting a close-to-the-�eld, savvy approach to lending. The re� nancing dif� culties, while Metro� nanciera SA, Mexican banking law established that the Secretaría a major lender to builders, restructured under pre- de Hacienda y Crédito Publico (SHCP) was in charge packaged bankruptcy protection after defaulting. In of authorizing and regulating the Sofoles. And while turn, Sofoles’ � nancing problems immobilized the the charter of the Comisión Nacional Bancaria y commercial paper market. Further contagion was de Valores (CNBV) included them within the list of avoided thanks only to the timely intervention of institutions under its oversight, there was a lack of Sociedad Hipotecaria Federal (SHF), the develop- clarity as to the exact mandate of the CNBV. The ment bank that had initially nurtured the growth of concerns about the discrepancy between Sofoles’ the mortgage Sofoles. Once the market abruptly shut de facto light regulation and the perception by the its doors, SHF brought the Sofomes back in its fold public that they were fully regulated and supervised through emergency credit lines. non-SIFIS—may become fossilized. If so, oversight (including the severity of regulation) the scheme would effectively end up having could (and probably should) vary depending some of the same flaws as the ill-fated, line- on the systemic importance of each insti- in-the-sand approach that led to the crisis. tution, these variations would be strictly An arguably preferable option would be to formula based and apply across the whole avoid making any open distinction between population of institutions within the perime- the SIFIs and the other institutions and to go ter, without any a priori distinction. However, for an ample perimeter, that is, choose option a major drawback of this approach is that a, to regulate all leveraged �nancial interme- ensuring that even the small institutions are diaries (except, perhaps, for the very small). effectively supervised puts a very high burden While a boundary effect always exists, under on of�cial oversight, thereby raising the cost option a, it would be simply illegal to set up of supervision across the board. Supervisors leveraged � nancial intermediation activities may soon realize that, though all leveraged without a license and being subject to pru- intermediaries are made subject to pruden- dential oversight. Although the intensity of tial oversight by force of law, there may not MICROSYSTEMIC REGULATION 207 BOX 12.2 Reforming the regulatory perimeter: United States versus the European Union Under the Dodd-Frank Act, the newly created Commodity Futures Trading Commission with Financial Stability Oversight Council (FSOC) regulatory powers on over-the-counter derivatives. responsible for macroprudential surveillance is At the same time, Title II of the Dodd-Frank Act entrusted with discretionary powers to submit non- empowers the Federal Deposit Insurance Corpo- commercial bank � nancial companies to supervi- ration with an orderly liquidation authority for sion by the U.S. Federal Reserve (that is, to become noncommercial bank SIFIs, aimed at reducing the a SIFI) if the council assesses that the bank’s fail- TBTF moral hazard associated with them. ure or its activities pose a serious risk for � nan- By contrast, the redefinition of the regulatory cial stability. The council may require any bank perimeter has received comparatively little atten- or nonbank � nancial institution with assets over tion in Europe, probably reflecting in large part its US$50 billion to submit certi� ed reports as to the predominant universal banking model, which has company’s � nancial condition, risk management limited regulatory arbitrage and the concomitant systems, and transactions with subsidiaries that growth of shadow banking (nonetheless, regula- are regulated banks, plus the extent to which any tory arbitrage also occurred through the buildup of the company’s activities could have a potential of off-balance-sheet exposures with less stringent disruptive effect on � nancial markets or the over- prudential requirements). Instead, efforts are cur- all � nancial stability of the country. To strengthen rently concentrated on adding a supranational layer supervision of holding company subsidiaries, the to existing supervisory arrangements, as exempli�ed Federal Reserve is authorized to examine nonbank in the creation of a European Systemic Risk Board subsidiaries that are engaged in banking activi- responsible for the monitoring, identi� cation, and ties. The proposals also aim to � ll regulatory gaps prioritization of systemic risk. At the same time, the by requiring that hedge funds and private equity reform proposals aim at developing a single set of advisers register with the Securities and Exchange fully harmonized rules across Europe, thereby limit- Commission (SEC), and providing the SEC and ing cross-border regulatory arbitrage. be enough human and �nancial resources to Because the regulated institutions would be perform that job adequately across the entire subjected to full prudential oversight, includ- universe of leveraged intermediaries (com- ing a possible systemic surcharge, they would mercial banks, investment banks, insurance pass on (through their lending rate) this sur- companies, cooperatives, credit unions, and charge to the unregulated institutions that so forth). Supervision on the cheap, moreover, borrow from them. At the same time, the can easily back�re, leading to a false sense of supervisor could adjust the intensity of over- security and moral hazard among depositors sight (and the prudential surcharges) of the and investors. regulated intermediaries, taking into full Option b would address in large part this account the risks that regulated institutions latter problem. Like option a, option b would would be assuming in their lending to the prudentially regulate the institutions that unregulated. In this way, the regulator would borrow from the public or in the market.6 effectively delegate the oversight of the unreg- Unlike option a, however, it would not regu- ulated to their regulated creditors, in a typi- late the �nancial intermediaries that borrow cal principal-agent relationship. This scheme only from the regulated ones. Thus, option would avoid regulatory arbitrage and limit b would reduce the cost of supervision (both the cost of oversight. At the same time, by the oversight cost and moral hazard cost) by opening the �eld to the new, smaller (pruden- effectively creating a two-tiered structure. tially unregulated) entrants, it would enhance 208 MICROSYSTEMIC REGULATION competition and promote innovation in niche activities conducted by non� nancial entities markets. Thus the small, unregulated inter- that do not neatly fall under the de�nition of mediaries could develop on the fringes of the “intermediation� established in the �nancial larger, regulated intermediaries. When suc- legislation. Indeed, illegal intermediation has cessful, they could become fully regulated happened recently, for example, in Colom- or be purchased by regulated intermediaries, bia, one of the LAC countries with the widest much as happened in Mexico with the most regulatory perimeter, as some illegal pyramid successful Sofoles. schemes mushroomed and eventually col- This approach would be somewhat similar lapsed. The case of intermediation by non�- to the manner in which the oversight of hedge nancial companies seems to be spreading fast funds was structured after the Long-Term in LAC countries, especially through depart- Capital Management failure, with prime ment stores that provide consumer loans, brokers exercising control of hedge funds typically via credit cards, and fund them by, through appropriate counterparty risk man- say, issuing commercial paper. In the case agement, and regulators concentrating on the of Chile, for instance, regulatory perimeter close supervision of the prime brokers. It may issues have been recently brought to the fore- be argued that this approach failed to prevent front as a result of an alleged fraud, where a hedge funds from building up highly leveraged (prudentially unregulated) non�nancial com- positions, with the subsequent increase in sys- pany that issued credit cards was automati- temic risk once the funds sold their positions. cally (and without the consent of the debtor) The counterargument is that—provided the rolling over the nonperforming credit card hedge funds were subject to minimum trans- debt to avoid raising loan-loss provisions. parency (auditing) requirements—the regula- Perimeter issues in LAC are likely to tors could have raised the risk weights of the become increasingly complex and dif�cult to loans by the regulated intermediaries to the control as � nancial systems mature. Many hedge funds, based on the leverage and sys- countries in LAC already have an unresolved temic implications of the latter. The failure problem of how to adequately oversee the of oversight, if there was one, was not a fail- numerous smaller institutions, such as credit ure of the scheme under which the agent was co-ops and micro� nance institutions, where operating but one of misunderstanding as to prudential oversight is required by law. In what systemic risk was all about. view of capacity and resource constraints, Except for the � nance companies, which the challenges of expanding regulation into are unregulated in several countries (for the outer edges of the system are daunting. example, El Salvador, Mexico, Peru, and Pan- In many LAC countries, full uniform over- ama), �nancial regulation in LAC has thus far sight would add hundreds of institutions to basically followed option a. Most supervisors the load carried out by supervisors. Resource have cast their net very broadly, while making considerations have already led a number of efforts (often not very successful) to apply reg- LAC countries to set up auxiliary models ulations uniformly within the perimeter. This wherein regulatory responsibilities are del- approach has thus far functioned relatively egated to nonsupervisory agencies, including well inasmuch as it has limited regulatory industry associations.8 While these and other arbitrage at the edges of the perimeter. How- models of auxiliary supervision usually entail ever, it has not fully eliminated arbitrage, incentives for these entities to cooperate with neither inside the perimeter (see below) nor their regulators, the risk of a lack of regula- at the outer edge.7 Indeed, even when all lev- tory independence and effectiveness looms eraged �nancial institutions are prudentially large.9 Restricted legal mandates and capac- regulated, boundary effects (hence regulatory ity to exert supervisory discretion are further arbitrage opportunities) at the outer edge of complicating factors. A possible alternative to the regulatory perimeter can take the form of such a model, more along the lines of the two- outright illegal �nancial activities or �nancial tiered regulatory scheme suggested above, MICROSYSTEMIC REGULATION 209 could be to allow the small credit co-ops to importance of differential regulation for remain prudentially unregulated as long as similar � nancial activities performed under they only fund themselves with their mem- different licenses has been well recognized bers under a clearly mutualized loss-sharing by standard setters and supervisors, includ- arrangement that makes them rather similar ing in LAC.13 The failure of the Caribbean to mutual funds. CL group, where the line between insur- Another complicating factor is that only ance and banking-type operations became a few LAC countries are endowed with the blurred, provides a particularly vivid illus- statutory discretion to extend the perimeter tration of the potential severity of these as needed, particularly to any �nancial entity problems (annex 12.A). However, the pace whose possible failure could threaten � nan- of reform (in LAC as much as in the rest of cial stability. In most countries (Uruguay the world) has thus far been slow, in part being the only exception in the region), this reflecting the complex trade-offs involved. prevents the authorities from flexibly redraw- There are two basic issues. First, although ing the “line in the sand� in the face of a a functional focus—whereby the same risk- rapidly changing �nancial landscape. Even if based regulation is applied based on the type laws were to be amended, empowering regu- of � nancial activity and across all leveraged lators to extend the perimeter as needed, in financial intermediaries—may be ideal in doing so the regulator may be confronted principle, it can be dif�cult to implement in with powerful pressures.10 practice. Second, even when the same regu- Last but not least, systemic risk may also latory framework applies across the board, emerge outside the �nancial sector. This may the question remains as to whether the same happen when the failures of large non� nan- institution should be allowed to conduct cial entities affect the �nancial intermediaries all � nancial activities or whether the latter that lent to them or undermine the markets should be compartmentalized across sepa- in which they (or their borrowers) operate.11 rate silos. Of course, if extending the perimeter to all As regards risk-based regulation, the cal- � nancial institutions is already challenging culation of value at risk (that is, the amount enough, extending it to non� nancial corpo- of capital that the firm needs to ensure its rations is outright unfeasible. This can be solvency with a certain probability over a viewed as further pushing the case in favor certain period) should in principle be identi- of an option b solution, in which supervisors cal across licenses. Basel II provides a general engage in some form of delegated supervision. model to follow, and the insurance regulation Through adjustments of risk weights, super- in the European Union is already moving in visors can induce the regulated institutions that direction, although implementation has to internalize the risks they are taking (and proved to be technically demanding and data contributing to spread to the rest of the sys- intensive. In the region, very few countries tem) by lending to borrowers (including the have implemented Basel II in full, and only large, unregulated corporations) that engage Chile thus far is in the process of introducing in socially risky activities.12 a risk-sensitive solvency capital requirement for insurance companies. As regards universal licensing versus silos, The inner boundaries: Silos the debate is more conceptual than techni- versus universal licensing cal. The proponents of universal licensing The global crisis also highlighted regula- argue that it not only eliminates the “inner� tory arbitrage within the perimeter, includ- regulatory arbitrage (that is, the arbitrage ing between investment and commercial between differently regulated activities) but banking books, between on- and off-bal- also improves efficiency by allowing uni- ance sheets, and across licenses (particu- versal intermediaries (that combine credit, larly between banking and insurance). The investment, securities, and insurance) to fully 210 MICROSYSTEMIC REGULATION diversify their risk and exploit economies probably more challenging in the short run of scale and scope. This has bene�ts at the but may also have better chances of long- micro level (intermediaries can better hedge term success. and diversify their exposures) as well as at In LAC, the prevailing silo-based regula- the systemic level (participation costs are tory culture appears thus far to have served reduced and risk is more ef�ciently allocated the region relatively well, particularly as across the population).14 However, follow- it has been paired with a generally more ing the global crisis, an alternative view has hands-on and prescriptive approach to regu- emerged in favor of silos.15 The argument is lation. However, as �nancial systems mature, that universal licensing can result in a loss of the approach is showing increasing strains. institutional diversity, as all institutions end In some cases, it has led to arti�cial barriers up behaving similarly and accumulating the between activities or to possible duplications, same risks. Instead of limiting the exposure with the same activity being conducted under of individual institutions through diversi�ca- different regulations in different silos—as is tion, such uniformity increases systemic risk the case, for instance, for deposit-like instru- (if one institution goes, everybody goes). It ments issued by insurance companies. More is also argued that economies of scale taper generally, it has promoted the growth of large off past US$100 billion in assets, a size well and complex conglomerates. On the bright below that of the larger international banks side, these conglomerates can be viewed as a (Haldane 2009a). Moreover, the disecono- desirable hybrid between silos and universal mies of resolving the failure of large institu- banks that allows for effective economies of tions tend to offset the economies of scale scope and scale (the conglomerates can—and that they may achieve in their operations. increasingly do—operate as integrated insti- Others have advocated a silo-type public tutions that share information systems and utility approach that ring-fences retail bank- central services), as well as risk diversi�cation ing from the rest of a �nancial conglomerate at the conglomerate level. Moreover, by alter- when the latter fails.16 ing the regulatory limits on cross-silo claims Overall, this debate is far from being and liabilities, the supervisor can in principle settled. As discussed in chapter 2, one may control the extent of risk exposure and diver- take the view that the more a � nancial sys- si�cation even at the level of the individual tem matures, the more interconnected it is silos. On the darker side, however, the con- likely to become, independent of the �nancial glomerates are dif�cult to effectively regulate structure the regulator seeks to impose. At and supervise, even more so as the scope for the same time, the � nancial structure itself interconnectedness increases, the identi�ca- is likely to evolve endogenously, irrespective tion of related parties becomes more compli- of the regulation. As amply demonstrated by cated, and reputational effects provide new the global �nancial crisis, investors will leave avenues for contagion. in troves the presumed islands of safety to The region’s weaknesses in consolidated migrate where the returns (hence the risks) are regulation identi�ed in chapter 10, particu- located. Hence, systemic problems are likely larly the lack of comprehensive supervisory to arise no matter what. At the same time, the powers to conduct consolidated regulation more barriers one builds, the more opportu- and supervision, compound this problem. nities one provides for regulatory arbitrage. Most countries do not require conglomerates Thus, a silo approach may provide some to formally establish themselves as a � nan- measure of systemic safety in the shorter run, cial holding company (FHC), and even when but also create a false sense of comfort leading they do, many countries lack the power to to an ultimately less-safe system. Instead, impose capital requirements on the FHC.17 an approach that removes arti�cial barriers FHCs can often be created abroad and fall altogether but internalizes systemic risks uni- under foreign (rather than domestic) super- formly through microsystemic regulation is vision, which exacerbates the cross-border MICROSYSTEMIC REGULATION 211 supervisory coordination weaknesses also Brazil, Chile, Colombia, and Mexico each identified in chapter 10. These gaps leave holds about 20 percent of total banking sec- open the possibility of multiple gearing (capi- tor assets (and 30 percent in Peru). Thus, if tal insuf�ciency disguised by the use of the one de� nes as a SIFI any institution hold- same capital by more than one of the mem- ing more that 10 percent of banking sector bers of the conglomerate or by a member and assets, LAC countries have the second high- the holding), a problem that is exacerbated est number of SIFIs in the emerging world when the conglomerate includes unregulated (�gure 12.1).18 � nancial entities domiciled abroad. In these There is a consensus on the need to adjust cases, judging the adequacy of capital com- the regulation of SIFIs to take into account mensurate with the risks being borne by the the contribution that these institutions pose group as a whole is a dif�cult undertaking to systemic risk, including as regards the and an important source of systemic risk, TBTF syndrome.19 However, as already given the interconnectedness and size of some noted, calculating ex ante the systemic risk of these �nancial groups. contribution of each SIFI is quite a challeng- In parallel with these inner perimeter ing exercise, even more so when the markets issues, there are also lively discussions across for the equity and debt securities issued by the the region on what the most appropriate SIFIs are relatively underdeveloped and illiq- supervisory architecture is. The two con- uid, as is the case for most LAC countries. structs are of course not independent. A uni- Moreover, larger buffers are unlikely to versal license can be naturally paired with a totally eliminate risks of default and the asso- single supervisor. Multiple supervisors (each ciated TBTF problem. Restricting the size of administering a different license) naturally go institutions (a solution envisaged by some) together with a silo-based system. But differ- could hamper ef�ciency and is politically dif- ent institutional arrangements can be found �cult, all the more so when the institutions to address the challenges posed by cross- are foreign owned. Even if � nancial institu- license regulatory arbitrage and the control tions’ internal growth were not limited in of � nancial conglomerates, from line super- LAC, external growth could be limited by visors with explicit coordinating mechanisms prohibiting mergers and acquisitions among (such as Chile), to uni�ed supervision under the large institutions.20 However, exceptional a single agency different from the central mergers and acquisitions would probably be bank (such as Colombia), and uni�ed super- needed to preserve �nancial stability in times vision housed within the central bank (such of severe systemic distress, which further as Uruguay). While all alternatives have pros underlines the complexities and inherent ten- and cons, a single supervisor should in princi- sions associated with this issue. ple have some natural advantages in unifying Thus, as in the rest of the world, the best the prudential treatment across different busi- defense against TBTF is probably to put in ness lines, regardless of whether the business place suitable legal and procedural arrange- lines stem from the same or separate licenses. ments to resolve in a non-destabilizing Ultimately, however, different organizational manner the failure of SIFIs. Following the architectures can all work, provided there are �nancial crises of the 1990s and early 2000s, effective coordination arrangements. bank resolution frameworks in many LAC countries were reformed. 21 To be sure, sev- eral of these frameworks remain untested, The SIFI problem and implementation remains an issue since The SIFI problem looms large in LAC. In the reforms have mostly focused more on addition to the presence of large and com- the legal than on the operational aspects. 22 plex financial conglomerates, many insti- Even where suitable bank failure resolution tutions can on their own be considered frameworks exist, however, arrangements to SIFIs. The single largest bank in Argentina, resolve the failure of �nancial conglomerates 212 MICROSYSTEMIC REGULATION FIGURE 12.1 Average number of banks with more than 10 percent of total banking assets, 2006–09 3.5 3.0 2.5 average number of banks 2.0 1.5 1.0 0.5 0.0 East Asia and Europe and Latin America Middle East South Asia Sub-Saharan Paci�c Central Asia and and Africa the Caribbean North Africa Source: Cortes, Dijkman, and Gutierrez 2011. are generally not in place throughout LAC. law. Then, being able to treat a group as a In effect, in most of LAC, the resolution of group “in death� requires being able to also �nancial institutions other than banks is still treat it as a group in life. Thus, supervisors subject to the general bankruptcy code. need to have the powers to conduct effec- As in the rest of the world, LAC will need tive consolidated supervision and regulation, to introduce legal reforms that grant authori- which, as already noted, is a substantial chal- ties the necessary powers to conduct an lenge in the region. Finally, the resolution orderly resolution of large, complex �nancial of cross-border � nancial conglomerates will conglomerates and nonbank SIFIs.23 The use require enhanced cross-border harmoniza- of contingent capital and bail-in debt as well tion of procedures. The fact that regional and as statutory bail-ins may also be appropri- global conglomerates usually operate in LAC ate. 24 Another tool that could be attractive through cross-border subsidiaries (rather for many LAC countries is the constitution than branches) should facilitate the task. of prefunded resolution funds with charge- Nevertheless, the lack of harmonized failure- to-bank levies.25 “Living wills� can also play resolution procedures and of burden-sharing a useful role, though they can hardly be con- agreements could prompt local authorities sidered an alternative to integral resolution to block flows among group entities in order reform. to ring-fence assets. As illustrated by the The resolution of complex � nancial con- failure of CL Financial (annex 12.A), this glomerates poses certain technical, legal, and issue looms particularly large in the Central political challenges speci�c to LAC. For start- America and Caribbean subregions. Dif�cult ers, the civil law does not make things easy, as as it would be to set up, a prefunded reso- the room for flexibility in the resolution pro- lution fund in these subregions would help, cess tends to be curtailed by administrative but regional coordination would be needed to MICROSYSTEMIC REGULATION 213 avoid distortions and double taxation in set- and central bankers a hedge in implement- ting such a fund. ing a new framework for systemic liquidity. However, as systems mature and interbank markets continue to deepen, pressures toward Systemic liquidity: Norms and holding higher-yield assets and exploiting access market funding more fully will no doubt As demonstrated by the global �nancial cri- also continue to build up. In this context, the sis, the growing interconnectedness of � nan- breadth of access to liquidity support facili- cial institutions and markets as �nancial sys- ties (lender of last resort, or LLR) is likely to tems mature is increasingly likely to give rise become an increasingly key component of to systemic liquidity disruptions. The latter policy. The key questions in this regard are can affect both funding liquidity (the ability how far to extend the LLR perimeter in terms to raise cash through borrowing or selling of institutions (that is, which institutions of assets) and market liquidity (the ability to should have access in addition to banks) and trade an asset at short notice without affect- instruments (that is, which instruments may ing its price; see Adrian and Shin 2010; be used as collateral, in addition to public Brunnermeier and Pedersen 2009). As noted securities). The systemic linkages increasingly in the chapter 11, the liquidation of � nancial taking place through funding markets also assets at �re sale prices under funding liquid- highlight the possible need for LLR facilities ity constraints is a major channel of ampli�- to support critically important markets, and cation of systemic shocks in downturns. But not just individual liquidity-distressed �nan- interdependencies may also build up at the cial institutions. interface between � nancial institutions and In LAC, the discussion on the scope of � nancial markets and infrastructures. 26 central bank liquidity support is particularly Not surprisingly, the reform of liquidity relevant but also particularly complex given regulation has been the subject of lively dis- the region’s history. Reflecting the region’s cussions. It has become broadly accepted that history of macroeconomic volatility and high the traditional prudential framework based inflation, most central banks operate under on maturity mismatches—that allows inter- very restrictive LLR frameworks. Liquidity mediaries to borrow short if they lend equally facilities are usually available only for a nar- short—is inadequate for systemic purposes.27 row range of �nancial intermediaries, chiefly Indeed, the new Basel III framework—which deposit-taking banks, which constitute the introduces a liquidity coverage ratio (LCR) backbone of LAC � nancial systems to date. and a net stable funding ratio (NSFR)— Nonetheless, the increasing importance of focuses instead on securing stable funding nonbank � nancial intermediaries and broad sources. And as regards the asset side, the funding markets described in chapter 3 will framework differentiates between the truly probably warrant a broader definition. 28 systemically liquid assets and the other assets, Collateral eligibility (which also tends to be including those that, while being short term, on the conservative side, that is, typically are not liquid in times of systemic stress. restricted to public sector debt securities) Under current conditions, LAC banks and pricing arrangements (for example, deep should be able to meet these new liquidity haircuts and high margin calls) may also requirements with relative ease, given the need to be reviewed. Understandably, many relatively modest share of nondeposit market central banks in the region are reluctant to funding and the relatively high share of pub- move toward more flexible arrangements out lic sector securities in their balance sheets. of concern for the possible impact on their Moreover, LAC’s previous experience with hard-fought gains in credibility. Possible ave- systemically oriented regulation—such as nues to manage these inherent conflicts might reserve or liquidity requirements on foreign include temporary relaxations of counter- currency deposits—should give its regulators party and collateral eligibility requirements 214 MICROSYSTEMIC REGULATION under circumstances of systemic distress, In LAC, most countries have followed the introduction of stronger analytical and thus far a relatively conservative approach to governance LLR frameworks, and the cre- � nancial innovation whereby new � nancial ation of private liquidity pools and insurance products need to undergo regulatory preap- arrangements.29 proval before they become part of banks’ permissible activities.32 By and large, such a policy has resulted in regulators approving Financial innovation: Did LAC only plain vanilla products that they “under- have it right? stood� well. With the bene�t of hindsight, As noted in chapter 2, financial innova- one could say that this approach has worked tion is an essential component of both the relatively well, with banks and other �nancial bright and dark sides of � nancial develop- institutions staying mostly clear of complex ment. While it plays a fundamental role in derivatives with a potentially high degree of expanding the quality and reach of � nan- toxicity. The extent to which such conserva- cial services, the global � nancial crisis gave tive behavior was induced by regulation or a disturbing glimpse of a world in which simply resulted from more limited market � nancial innovation could spin largely out opportunities and/or more limited risk tak- of control. 30 The innovators did not care ing, or from a more traditional risk manage- about the possible systemic perils of their ment culture among bankers, is not always creations (that is, the collective action par- clear, however. adigm at work), and much of the investing Be that as it may, as financial systems public did not understand the risks associ- mature, such a tight regulation of innovation ated with them (that is, the collective cog- (as noted above) is likely to show increas- nition at work). To make matters worse, ing strains. In this context, LAC regula- innovations were favorably sanctioned by tors’ limited independence might become the rating agencies and largely ignored by more problematic, and regulators’ need for the regulators. staying ahead of industry practices might While one should thus think that regu- become increasingly taxing, given resource lation should have a potentially important constraints. A more flexible approach that role to play in controlling � nancial innova- matches instruments to risk management tion, regulatory intervention faces the same needs and capacities across industry licenses cognitive problem that all investors face. is likely to be needed. Yet, doing this effec- Distinguishing between “good� and “bad� tively without promoting regulatory arbitrage innovations is no trivial matter and depends will not be an easy task. At the same time, in part on whether end users fully understand and perhaps more important, it may be dif�- the risks associated with the innovations’ use cult to reconcile a tight control of innovation and, when they do not, whether they should with the growing need for more sophisticated be protected from their own mistakes. Faced market instruments (such as credit swaps) with such uncertainties, regulators may be that can help supervisors assess risk by put- easily pulled in opposite, equally inappropri- ting a price on it (more on this in chapter 13). ate directions, either rubber-stamping inno- Hence, some system for “piloting� � nancial vations under pressure from the industry, or innovation before it is mainstreamed would putting a freeze on innovations under the pull be worth considering. Involving credible of risk aversion.31 Thus, the pitfalls associated private sector representatives in the process with giving regulators the powers and man- of approving innovations would also help date to regulate innovation may end up either dampen the tendency of of�cial regulators to contributing to moral hazard by sanctioning be unduly risk averse. risky regulations or stifling �nancial develop- All in all, a revision of LAC’s approach ment through excessive conservatism. to the regulation of financial innovation is MICROSYSTEMIC REGULATION 215 probably an important but quite complex task dif�culties of resolving SIFIs, managing sys- lying ahead. There is little doubt that such a temic liquidity, and putting � nancial inno- review will be required. However, unlike in vation to work on the bright—rather than other areas, there is perhaps less urgency here dark—side, one gets a complex yet essential in changing what has so far worked relatively agenda for reform. well. Hence, there should be enough time for Tackling both micro- and macrosystemic well-pondered reforms. risk calls for an eclectic approach that builds on ex ante prudential regulation as well as ex post safety net and resolution frameworks. The regulatory agenda ahead As regards the perimeter of regulation, the Financial systems in LAC, aside from some region will need to find its own preferred Caribbean countries, have weathered the road, building on its strengths but aware of global � nancial crisis largely unscathed. The the challenges posed by regulatory costs and region’s wide outer perimeter of prudential the rising threat of � nance in the shadows. regulation; its closely circumscribed, silo- On resolution issues, the region will need to based inner regulatory perimeter; and its address the current weaknesses as regards the relatively hands-on approach to regulation bankruptcy framework for nonbank � nan- (including that of �nancial innovation) have cial intermediaries. It will also need to face all contributed to this outcome, in part by its demons in rethinking the financial sys- keeping the emergence of a “shadow� bank- tem’s safety net. It is simply not an option ing system at bay. However, the few local- for central banks to let the problem pass. On ized episodes of severe stress in the recent �nancial innovation, a new balance between global crisis are probably early manifesta- prudence and boldness will need to be found. tions of a world yet to come, with ever-rising Now is the time to push through the needed exposure to systemic risk through increasing reforms, both to help deal more efficiently interconnectedness and complexity. Already, with the emerging challenges and to be better the provision of � nancial services through prepared to deal with an eventual fallout. The conglomerates—a potential fulcrum for the evolving international financial standards buildup of systemic risk—is raising dif�cult should provide a useful guide that is likely issues in the inner perimeter. And the cur- to be more evolutionary than revolutionary. rent policy of pushing the outer perimeter of However, countries will need to re� ne and regulation as far as the eye can see is likely adapt regulation and institutional setups to to face increasing challenges. If one adds the their particular circumstances. 216 MICROSYSTEMIC REGULATION Annex 12.A Challenges posed by large, complex �nancial conglomerates: The case of CL Financial in the Caribbean CL Financial is one of the largest conglom- In February 2009, Trinidad and Tobago erates in the Caribbean, with interests in authorities issued a rescue package to CL insurance, banking, energy, agriculture, and Financial. This included the provision of real estate, and operations in several coun- US$800 million (4 percent of GDP) to replen- tries (annex �gure 12.A.1). The holding is ish CLICO’s assets. CLICO and BAT were incorporated in Trinidad and Tobago, with intervened and their management replaced. assets estimated at about 75 percent of the The new government that took of�ce after country’s GDP in 2006. In recent years, life the 2010 elections announced that the short- insurance subsidiaries expanded aggressively term investment business of CLICO would by selling deposit-like products to the public, be separated from its traditional business, promising a high rate of return, and invested and the latter would merge with BAT. Insur- the proceeds into U.S. real estate and other ance policies would be honored, backed by ventures whose values sharply declined dur- the statutory fund, but annuity and short- ing the global credit crisis. term investment holders would only be As a result of the global credit crisis, partially compensated. The plan was chal- CL Financial started experiencing liquid- lenged by investors but upheld by the courts. ity pressures. Its investment bank subsid- A special fund was established for the credit iary, CIB, faced a high level of withdrawal unions. requests, and its insurance subsidiaries, The ECCU governments stressed that Colonial Life Insurance Company (CLICO) the collapse of CLICO and BAICO was a and British American Trinidad (BAT), had regional problem requiring a regional solu- trouble meeting statutory fund require- tion. In April 2009, the Caribbean Commu- ments. British American Insurance Com- nity (CARICOM) governments established a pany (BAICO) incorporated in the Bahamas liquidity support fund for the policyholders and CLICO International Life incorporated of CLICO and BAICO to which Trinidad in Barbados (CLICO Barbados) also entered and Tobago contributed US$50 million. As into dif�culties. of June 2009, BAICO was insolvent, with a T he col lapse of C L Fi na ncia l a nd de�cit of US$287 million. BAICO was put its insurance subsidiaries represented a under judicial management in September major challenge for the financial stabil- 2009, and the ECCU governments developed ity of the Eastern Caribbean Currency a resolution strategy in November 2009 to Union (ECCU; composed of Antigua and prevent the systemic consequences of liquida- Barbuda, Dominica, Grenada, St. Kitts tion. It is expected that BAICO will ultimately and Nevis, St. Lucia, and St. Vincent and be liquidated and the annuity policyholders the Grenadines). The exposure of ECCU will be creditors in the estate. CLICO Bar- policyholders and depositors to CLICO bados still operates independently. The gov- and BAICO is estimated at about 17 per- ernments of the ECCU resumed talks with cent of ECCU GDP. The ECCU insurance the new government of Trinidad and Tobago and banking sectors are closely intercon- to investigate how the country could support nected since (a) many � nancial institutions the resolution process further. Trinidad and invested directly in annuity products from Tobago has agreed to explore the creation of CLICO and BAICO, and (b) the companies a CARICOM Fund with the support of the provided insurance as collateral on bank- ECCU and Barbados to provide relief to poli- ing assets such as mortgages. cyholders of CLICO and BAICO. FIGURE 12.A.1 CL Financial annual report, 2007 finance and energy, agriculture and communications insurance real estate banking forestry, and services manufacturing, retail Bahamas CLICO (Bahamas) British America Ltd. Insurance Co. Ltd. Belize CLICO (Belize) Ltd. CLICO CLICO International CLICO International CLICO Agricultural Barbados Life Insurance Life Insurance CLICO Property Mortgage and Development Co. Ltd. Co. Ltd. Finance Co. Development Co. Ltd. Co. Ltd. CLICO (Guyana) CL FINANCIAL Guyana Caribbean Ltd. Resources Ltd. OECS CLICO (Turks and *Eastern Caribbean *Eastern Caribbean Caicos) Ltd. Financial Holding Co. Home Mortgage Bank Ltd. Suriname CLICO (Suriname) Ltd. *Europa LLC IBIS Asset United Management Ltd. CL World Kingdom Brands *United Systems United CL Capital Market CLF Latin *Southern and Software Inc. CAMMB States Inc. America Chemical Corp. *United Image CLICO Technology Inc. CLICO Investment Home Trinidad and (Trinidad) Ltd. Bank CLICO Primera Oil and CL *One Caribbean Mortgage Bank HCL Group Tobago Ltd. Energy Co. Gas Ltd. Communications Media Ltd. CMMB Ltd. Colfire Co. Ltd. Angostura Methanol Holding Ltd. Republic Bank CL Holdings Ltd. Ltd. Permanent *Caribbean Nitrogen Co. Ltd. Health Net *Laqtel Ltd. Source: CL Financial Annual Report 2007 (http://www.cl�nancial.com) Note: * = associate or joint venture company; OECS = Organisation of Eastern Caribbean States. 217 218 MICROSYSTEMIC REGULATION Notes the case of the cajas (savings and loan institu- tions) in Mexico, where oversight functions 1. This chapter draws heavily on the paper are delegated by the CNBV to the federations “Micro-systemic Regulation: A LAC of cajas. By law, however, the CNBV retains Perspective� by Mariano Cortes, Miquel ultimate responsibility for oversight of the Dijkman, and Eva Gutierrez (2011), which is cajas and its outcomes. part of the Edited Volume that accompanies 10. According to the results of the joint World this LAC Flagship Study. Bank–ASBA survey, about 60 percent of 2. In the words of Dewatripont and Tirole respondents viewed limited legal mandates as (1994), the state (the regulator) “represents� hindering effective systemic oversight and the the interests of the small investor against use of discretion; at the same time, 40 percent those of the more informed (or more sophis- of respondents indicated that political and ticated) market participants who could be industry pressures were also an important taking advantage of the little guys, eventually hindrance in the use of discretion. limiting their access to �nancial services. 11. One such illustrative example is that of the dif- 3. On the “twin peak� approach to regulation �culties encountered by Comercial Méxicana, see Taylor (1995, 2009). a Mexican hypermarket group, which at the 4. Notice that the fourth paradigm, costly height of the �nancial crisis suffered sharp enforcement, does not justify prudential reg- derivative losses in toxic foreign exchange ulation. It does justify, however, the role of (FX) derivatives. Its subsequent default in the the state in facilitating contract enforcement autumn of 2008 contributed (together with (hence commitments) through improving the the Sofoles) to the breakdown of the Mexican enabling environment, notably the collateral commercial paper market. regime and the judiciary. 12. In any event, it is of course essential that 5. Any form of borrowing constitutes leverage in supervisors carefully monitor (and assess the this de�nition, whether borrowing from retail systemic implications of) the risks accumulat- depositors, from large depositors, from other ing outside the �nancial sector. �nancial institutions, or through the capital 13. For example, a capital charge on a credit risk markets. Unleveraged institutions that engage held by a bank becomes a technical provision in pure asset management, such as mutual when transferred to an insurance company. funds, would be subject to market conduct See Joint Forum (2010). The World Bank– regulation but not to prudential regulation. ASBA survey of LAC supervisors indicates 6. Option b would require all leveraged �nancial that the perimeter issue that is perceived to intermediaries to have a license but, unlike in be the most pressing is that of potential risk option a, not all would be prudentially regu- shifting between the entities of �nancial con- lated. glomerates, both cross-border and domestic. 7. As revealed by the World Bank–ASBA survey, 14. Several high-income countries have contin- the global crisis appears to have heightened ued to express a clear preference for universal the awareness among LAC supervisors of the licensing, even after the global crisis. Thus, importance of regulatory arbitrage. Indeed, the Hong Kong Monetary Authority contin- all the supervisory agencies that responded to ues “to see merit in the model of universal the survey indicated that they are considering banking� (see Chan 2011). Similarly, while extending the perimeter to the hitherto unreg- recognizing the need for stricter regulation of ulated, unsupervised intermediaries. the largest banks because of their massive rel- 8. This model is especially popular for deal- ative size, the Swiss Central Bank has recently ing with the cooperatives and small savings noted the advantages of universal banking and loans, whose supervision is particularly for diversifying and attracting high-net-worth labor-intensive given the small scale and geo- individuals. graphical dispersion of the institutions (for 15. See, for example, Haldane (2009a) for the example, El Salvador, Mexico, and Peru). United Kingdom. Similar views in the United 9. This is the case at present in Paraguay regard- States have led to pressures toward fully rein- ing the supervision of cooperatives—a sys- stating the Glass-Steagall Act, eventually giving temically important segment of the �nancial way to the “Volcker Rule,� whereby “banking system—where the board of the supervisory entities, which bene�t from federal insurance agency is selected by the industry. It is also on customer deposits or access to the discount MICROSYSTEMIC REGULATION 219 window, are prohibited from engaging in senior unsecured in the latter) that contain proprietary trading and from investing in or contractual clauses prompting conversion sponsoring hedge funds and private equity to common equity when the institution is no funds, subject to certain exceptions.� longer viable, diluting existing shareholders’ 16. See, for example, Kay (2010). This is the equity and applying a haircut on debt hold- approach proposed in the United Kingdom ers. Contingent capital conversion could also by the Independent Commission on Banking, be triggered when the prudential ratios of the the so-called Sir Vikers Commission (see institution begin to deteriorate in the form Morrison & Foerster 2011). of an automatic prompt corrective action. 17. As noted in Joint Forum (2010), de�nitional Statutory bail-in provisions empower liqui- differences as to what constitutes a group and dating authorities to write-down or convert how entities enter for the purpose of calculat- debt into equity as part of the resolution ing group capital also create problems for the procedure outside normal bankruptcy pro- assessment of risk and capital adequacy. cedures. See FSB (2010) and BCBS (2010c, 18. This is the de�nition proposed by Thompson 2010d). (2009). 25. The purpose of such funds is to enhance the 19. The Financial Stability Board (FSB) and effectiveness of the resolution framework by the Basel Committee on Bank Supervision ensuring that funds are readily available (its (BCBS) have stated that SIFIs, particularly rationale is akin to that of deposit insurance the global ones, should have higher loss- funds with resolution powers). absorbing capacities. See BCBS (2010b) and 26. In Brazil, for example, a drastic increase in FSB (2010). market volatility boosted margin require- 20. The U.S. reform, for example, prohibits merg- ments at the stock and futures exchange, call- ers or acquisitions if the resulting company’s ing for a large posting of liquid collateral at market share (in terms of liabilities) would a time of severe liquidity stress. See Mesquita exceed 10 percent. and Torós (2010). 21. Legal reforms to the bank failure resolu- 27. See, for example, de la Torre and Ize (2011). tion regimes took place in the 1990s and 28. Such a broadening of the LLR perimeter has early 2000s in Argentina, Bolivia, Colombia, already taken place on occasions of high Dominican Republic, El Salvador, Guatemala, �nancial sector stress. For example, in the Mexico, and Peru, among others. The reforms mid 2000s, an emergency decree allowed provide better resolution tools, including the securities brokers in Colombia to repossess transfer of assets and liabilities to an exist- their government securities directly at the ing bank (purchase and assumption) or to a central bank to address market disturbances. bridge bank. 29. Private liquidity pools and insurance arrange- 22. Starting in 2008, �nancial crisis simulation ments are particularly attractive in the case of exercises were conducted in several LAC coun- highly dollarized countries. tries with the support of the World Bank and 30. Haldane (2010b) �nds that the real economic the Financial Sector Reform and Strengthening bene�ts of the precrisis surge in �nancial Initiative (FIRST). In most cases, the exercises innovation have been unimpressive once cor- underlined the need to (a) develop a frame- rected for the risks that were incurred. work to assess the potential systemic impact 31. The global �nancial crisis has spurred aca- of failing institutions; (b) improve operational demic interest in the merits and perils asso- readiness to resolve an institution through pur- ciated with �nancial innovation and the chase and assumption or bridge bank tools; implications for regulatory policy. See, for and (c) improve coordination between the example, Gennaioli, Shleifer, and Vishny different �nancial authorities and formulate a (2010) and Lerner and Tufano (2011). communication strategy to calm and reassure 32. This approach is reflected in the relatively investors and the public. high scores of LAC countries as regards the 23. See, for example, Squam Lake Working second Basel Core Principle, which deals with Group on Financial Regulation (2009). the permissible activities of banks. More than 24. Contingent capital and bail-in debt are 90 percent of countries have been graded at securities (subordinated in the �rst case and least largely compliant. Systemic Supervision 13 T raditional prudential supervision them seamlessly with the exercise of good starts from the view that strong indi- judgment. vidual banks lead to strong banking • Systemic supervision will require not systems. Hence, it monitors the behavior of only developing new analytical capabili- individual � nancial institutions. The global ties but also taking a different approach crisis exposed the fallacy of composition to gathering, analyzing, and circulating embedded in this view by showing that the information, including a better use of mar- risk to the system is greater than the sum ket signals and more effective communi- of the risks faced by individual institutions. cation. To some degree, the Latin America and • Implementing this ambitious agenda will Caribbean (LAC) region has grown quite require new tools and, perhaps more familiar with systemic events, whether important, new powers and a new organi- homegrown or imported. 1 W hile L AC zational arrangement. supervisors have learned from these events • LAC’s experience with systemic events and and, partly as a result, have put in place a generally more hands-on oversight culture stronger supervisory framework, there is no (including as regards � nancial innovation) room for complacency. So much is new in should provide a good basis to build on. systemic supervision that sustained efforts • However, LAC also faces special chal- by LAC supervisors to learn and adapt will lenges, including high concentration in be needed. 2 Main highlights in this chapter �nancial ownership (that promote intercon- are as follows: nectedness and too-big-to-fail) and some shortcomings in independence, legal pro- • Systemic supervision is all about making tection, and legal philosophy that weaken the connections through time, between the grounds for supervisory discretion. the parts and the whole, and between sta- • While the region is already well on its way bility and development. toward establishing financial stability • Systemic supervision should symbiotically committees, dif�cult issues of functional- complement systemic regulation by test- ity, transparency, and accountability still ing the rules and, where needed, fusing need to be addressed. 221 222 SYSTEMIC SUPERVISION The rest of this chapter is organized as that inflicts negative externalities on the rest follows. The first section deals with the of the system. new connections, the next section with the new approach, the third section with the new tools, and the fourth section with implemen- Connecting through time tation. The � nal section concludes by map- Traditional prudential regulation, of the ping the agenda ahead. type prevailing before the global crisis, was mostly time invariant and, hence, mostly procyclical (that is, worsening inefficient The new connections systemic dynamics rather than leaning Systemic supervision is all about making the against the wind). Instead, to the extent that right connections. These include connect- it extends beyond simple compliance and ing the parts and the whole, connecting the into the realm of true risk-based supervision, parts through time, and connecting develop- systemic supervision must become dynamic ment with stability. and forward looking, as both risks and risk exposures vary over time in an endogenous way. Hence, the supervisor’s task must start Connecting the parts and the whole where the regulator’s ends. Even when regu- A tendency among supervisors is to equate lation has been dynamically adjusted along good systemic supervision with good consol- macroprudential lines to correct for pro- idated supervision. Clearly, the latter helps cyclicality, gaps are likely to be left between the former, as it connects the activities of regulation—which, by construction, is a the members of a � nancial group, thus help- given—and the risks and exposures—which ing identify potential channels of systemic vary over time, often in unexpected ways. contagion. Given that many � nancial � rms As discussed in chapter 11, the uncertainty are parts of conglomerates, a narrow stand- as regards the risks faced by the financial alone view of individual regulated entities is system and its vulnerabilities (which limit obviously insuf�cient to obtain a full picture the scope for rule-based regulation) further of the conglomerate’s risks. Thus, the gaps enhance the need for such state-contingent in consolidated supervision that were iden- supervisory discretion. Thus, the supervi- ti�ed and discussed in chapters 10 and 12 sor’s task is to complement the regulator’s, clearly will need � xing. However, systemic �lling up the gaps that a set rule unavoidably supervision goes much beyond consolidated leaves in an uncertain and rapidly changing supervision. While focusing on a broader environment. entity (the financial group), consolidated supervision still follows an idiosyncratically Connecting stability with development oriented approach. Systemic supervision, instead, looks at the � nancial system as a Another key issue in formulating a new vision whole, recognizing and addressing the inter- for supervision is striking the right balance connectedness of all � nancial sector play- between financial stability and financial ers, whether they belong to the same group development. If supervisory responsibilities or not, and seeking to understand how the are to be broadened to include systemic risks, system is wired, including the nature of the there will be a need to � nd the appropriate interactions between the parts and between trade-off between the necessary restrictive- the parts and the whole. Although the ness of the supervisory stance and the pos- actions of each institution or � nancial con- sible stifl ing of � nancial development. This glomerate may be individually optimal and is particularly important in LAC, because seem acceptable to a traditional supervisor, many countries, in part in response to their they may contribute to driving the system prior experience with crises, are perceived along a socially inappropriate and risky path to have a more hands-on, less permissive SYSTEMIC SUPERVISION 223 supervision (including as regards the con- regulations so much that prudential over- trol of � nancial innovation) than is (or was) sight ends up losing its teeth. Granting the case in most developed countries. Some supervisors discretion requires con�dence argue, based on the experience of develop- that they are sufficiently independent of ing � nancial systems, that a relatively rigid political pressures, have the expertise and framework may be needed initially but that technical capacity to make sensible judg- greater supervisory flexibility may be appro- ments, and have the legal powers to enforce priate once markets have developed.3 Others those judgments. The latter is an even argue that strict supervision can be consis- greater challenge in civil law countries, such tent with further � nancial deepening, even as in Latin America, where administrative in already well-developed systems (a view law restricts supervisors’ actions only to that the recent global crisis seems largely to those actions that are speci� ed in the law have vindicated).4 In any event, a right bal- and accompanying regulations. 6 Nonethe- ance needs to be found because the modali- less, as discussed below, setting up the right ties of supervision (that is, how, when, and institutional arrangements (including, of where supervision is made more stringent) course, adequate transparency and account- are likely to have an important impact on ability) for a decision-making framework � nancial development. should go a long way toward meeting this challenge. A new approach The new approach will need to deviate from Top-down versus bottom-up the traditional one in a number of ways, Systemic supervision examines risks from including as regards the boundary between a top-down perspective and blends this supervision and regulation, the intermingling analysis with a bottom-up analysis. The of top-down and bottom-up approaches, the intermingling of bottom-up microsupervi- roles of off-site versus on-site supervision, sion with top-down systemic supervision the use of markets, and the provision of pub- provides an information feedback loop that lic analysis and information. allows for an enhanced assessment of sys- temic risks. Indeed, for systemic supervi- sion to be truly effective, understanding the Regulating versus supervising actions of individual institutions and their Systemic regulation implies rules that apply tolerance to various types of risks is neces- to all institutions. Moreover, as discussed sary. One of the weaknesses in the finan- in chapter 11, these rules may be based on cial stability analysis published thus far by aggregate, rather than institution-speci� c, central banks has often been the absence of parameters. For example, while a systemic the supervisors’ perspective on what is hap- regulation might establish a uniform maxi- pening at individual institutions. Combin- mum loan-to-value ratio, systemic supervi- ing bottom-up traditional supervision with sion may need both to deviate uniformly top-down systemic supervision provides from the rule (by changing this ratio) and key synergies. Top-down supervision gains to give supervisors the discretion to apply from the insights and information gathered stricter standards to an individual institu- by bottom-up supervision, and both add to tion because its actions pose a greater risk and gain from new perspectives on systemic to the system. Thus, the inherent tensions risks, thereby refining their assessments. and complementarities between regulation However, the skill sets for top-down sys- and supervision are an essential part of the temic analysis are very different from those “rules versus discretion� debate. 5 The chal- needed for traditional bottom-up idiosyn- lenge is to build suf�cient discretion into the cratic supervision. At the same time, coor- supervisory process but without relaxing dination between the two approaches is very 224 SYSTEMIC SUPERVISION important, particularly in countries where trigger liquidity problems. 9 At the same bottom-up supervision is outside the central time, however, a good argument can be bank (more on this below). made that gaps in market discipline result, in part, from the insuf�cient exploitation of market signals, as well as the incompleteness Off-site versus on-site of those signals. Hence, the question is not A closely related (but conceptually distinct) so much whether market discipline is good issue is the relative emphasis on off-site ver- or bad, but rather how supervisors can make sus on-site supervision. At first sight, one better use of market signals. Where the lat- might think that, because it involves the ter are weak, the issue is how to strengthen forest more than the trees (with a greater them. In recent years, various market instru- focus on systemwide analysis and research), ments, such as credit default swap spreads, systemic supervision is more about off-site have emerged and can help detect some- than on-site. Hence, supervisory agencies thing that supervisors may not have been might need to refocus to some extent and aware of; or instruments may matter simply primarily concentrate on strengthening their because they—as well as the prices of � nan- off-site analysis. Indeed, if on-site supervi- cial institutions’ debt and equity securities, sion is limited to checking compliance with where available—gauge the state of expec- regulations, it will have nothing much to tations, which in turn helps predict pricing add to the identi�cation of systemic risks.7 and the availability and cost of funding.10 In However, the adoption of systemic super- LAC, more and better use of such signals, vision calls for a rede� nition of the role of when available, is clearly desirable. Indeed, on-site supervision, rather than a downgrad- several central banks, for example, in Brazil, ing. Although much of the work of mapping Chile, and Mexico, have been already doing interconnectedness and monitoring links work along these lines. But there is clearly between the balance sheets of important ample scope for furthering such types of institutions can be done off-site, it can best analysis.11 be veri�ed by on-site inspectors. At the same But limited capital market development time, on-site inspections are the best place in LAC, resulting in weak market signals, to understand in full depth the evolving constitutes a severe limitation. Hence, a key business models, the links with other parts question for the region is whether policy of the �nancial system, and the risks associ- makers can help develop market instruments ated with them. The insuf�cient communion that promote risk discovery. The latter gen- of on-site and off-site supervision can thus erally requires costly monitoring. Hence, lead to poor, or outright fl awed, analysis. for market participants to invest suf�ciently This is in fact not an uncommon result of in monitoring, they need to have suf�cient institutional barriers between the central “skin in the game.� But, in view of risk aver- bank’s � nancial stability staff and the super- sion, large risk exposures can mean a high visors, even where both staffs are located in risk premium (see chapter 9). Thus, unless the same institution. supported in some fashion by the state (and perhaps even subsidized), market instruments that aim at risk discovery may simply be too Use of markets expensive to see the light of day, all the more A long-standing debate is carried on within so in shallower �nancial systems. An impor- the supervisory community regarding the tant research agenda for the region, therefore, efficacy of market-based financial indica- is to help design, introduce, and support such tors and the reliance on market discipline.8 instruments. Mandatory issue by financial Traditionally, supervisors have, for the institutions of subordinated debt or purchase most part, paid scant attention to � nancial of market-based insurance might be one such markets, except when market perceptions avenue to explore.12 SYSTEMIC SUPERVISION 225 Analysis and information that these reports are published may induce supervisors to exercise self-restraint for In addition to “skin in the game,� another fear of igniting a self-ful� lling run. Also, in key requirement for proper market disci- some countries central banks lack access to pline is analysis and information. Private detailed supervisory data on �nancial institu- market participants (rating agencies, audi- tions and thus are constrained to an overly tors, financial advisers, and so forth) can macro view. naturally provide much information on indi- vidual � nancial � rms. However, the quality and depth of these assessments is limited by New tools costs, confl icts of interest, and con�denti- Systemic supervision also will require devel- ality. Thus, some of the required informa- oping new tools to map the risk boundaries tion is in the nature of a public good that and the interconnectedness between � nan- supervisors must foster or provide. While cial institutions and between markets and total openness can be problematic (exami- institutions. nation report data are generally kept con�- dential), a good argument can be made that more information, provided it is presented Mapping the boundaries in a way that protects basic con�dentiality Having a better feel for the location of the regarding customers and business models, (systemic) cliffs is central to systemic super- can only be for the better.13 Indeed, this is vision. Since one wants to focus on tail already happening in LAC.14 risk, and tail events may not have occurred, But supervisors also need to provide anal- market data are unlikely to be suf�cient.17 ysis and information on the system, how it One alternative is stress testing, a technique is wired and interconnected, and what risks � rst widely disseminated through the Finan- may be brewing. This analysis and informa- cial Sector Assessment Programs (FSAPs).18 tion sharing is something that only supervi- In LAC, many countries already conduct sors can do. Markets will most certainly not such exercises on a regular basis.19 There do it on their own because it is too complex, are many difficulties inherent to systemic hence too costly and, given its public good stress testing, however. 20 First, financial nature, subject to too many collective action � rms’ natural incentive is to construct their frictions (free riding). As financial systems own tests so that the tests result in a need mature and become increasingly complex for less equity capital rather than more. Sec- and interconnected, such systemic analyses ond, models need to incorporate shocks, become increasingly valuable yet also increas- correlations, and joint default probabilities ingly costly. Supervisors need to scout ahead that are nowhere to be found in the data. and search for upcoming icebergs. When Third, stress tests must expand beyond capi- risks are detected, supervisors may need both tal adequacy, into liquidity, which is gener- to inform (so that market participants can ally where crises start. However, liquidity correct the course) and to act (thereby help- modeling is notoriously difficult. Fourth, ing resolve coordination failures).15 Gener- “connecting the dots,� identifying the ally, the former should facilitate the latter. potential fault lines, and looking both out- Financial stability reports, in LAC as in side the box and beyond the horizon are the other regions, have become the main vehicle most dif�cult, yet most crucial, elements of for regular reporting of the results of such good stress testing. Indeed, failure to iden- exercises.16 However, they have run into some tify the channels and links through which a practical dif�culties. In particular, there is a crisis could emerge and spread (rather than question as to whether they provide mean- simple compliance failures) was arguably ingful and authentically transparent and the most severe supervisory failure in the candid diagnostics of systemic risk. The fact United States. 226 SYSTEMIC SUPERVISION Therefore, before proper models or cri- organization (which requires full coop- sis scenarios can be designed, understand- eration with home supervisors) is therefore ing how the system is wired, and where the vital. Unfortunately, during the crisis some weakest connections are, is of utmost impor- home supervisors were unwilling to share tance. Failing to do that results in a real risk any information with host supervisors. of embarking on sophisticated but ultimately But balance sheet interconnectedness sterile number crunching. Supervisors need to across institutions with different ownership discuss the simulations with �nancial institu- also clearly matters. For example, given the tions and impose uniform stress parameters high economic concentration of the real sec- derived from well-formulated, macro-based tor in many LAC countries, � nancial � rms scenarios. In some cases, the latter may need often become heavily interconnected through to include such speci�cs as losses given default their borrowers. While market interconnect- and declines in collateral values. Although edness is less extensive in LAC than it was these discussions are likely to be difficult, in the United States, it is no less crucial, as they constitute a prime opportunity for col- illustrated in Mexico by the spillovers on the lective learning, not unlike those between whole commercial paper market of the �nan- the International Monetary Fund (IMF), the cial dif�culties incurred by a couple of small World Bank, and national authorities during Sofoles. How financial institutions fund FSAPs. Indeed, many LAC countries have themselves is particularly critical. Financial already embarked on this route.21 Much more � rms that fund themselves mainly through work clearly lies ahead. wholesale markets or their offshore parents are of course much more vulnerable. Under- standing cross-country contagion is also crit- Mapping interconnectedness ical. Indeed, empirical work has shown how One of the key lessons from the global cri- closely connected some Latin American coun- sis and the Lehman Brothers debacle is tries’ economies are to their neighbors (Sosa the extent and implications of financial 2010). Thus, mapping of interconnectedness institutions and markets’ interconnected- can be as much a statistical exercise based on ness. The latter can take place through analyzing price correlations as pure detective balance sheets, market prices and market work aimed at dissecting each of the links sentiment, or ownership. The last chan- described above.23 In LAC, several countries nel is particularly relevant in LAC in view have started to work on both sides of this of the prevalence of large, opaque � nancial problem. Again, however, there is much more groups with highly concentrated ownership. to be done.24 It is often unclear how the real and �nancial components of the group relate to each other, Getting there how reputational effects may transmit across the group, and how a crisis will affect the Implementing such changes will in turn parts and the whole, domestically as well as require shaping up the necessary skills, across borders. 22 In LAC, the prevalence of making sure supervisors have the necessary foreign banks and their subsidiaries brings powers, and putting in place proper orga- about another crucial source of cross-border nizational arrangements. Continued peer interconnectedness. Notwithstanding ring- reviews may also help. fencing capital through capital requirements on branches, there is always the risk of a loss Skills in con�dence in the host bank if the parent experiences difficulties. Also, the parent’s Boosting the skills required to address sys- ability to support its af�liate and sustain its temic risks involves a quantum leap, not a lending may become constrained. Under- marginal improvement. It will likely require standing the full dimensions of the parent’s an altogether different skill set from that SYSTEMIC SUPERVISION 227 traditionally required for checking compli- by adequate accountability. In this context, ance. The supervisors engaging in systemic it is worrisome that the use of automatic cor- supervision will need to become fluent in rective action triggers in LAC appears to be economics, statistics, and � nance and must in need of signi�cant strengthening.26 LAC’s venture into largely unmarked territory. remaining problems of insuf�cient indepen- Indeed, by some measure, they will need to dence and legal protection for supervisors engage in more research-like activities than (discussed in chapter 10) seem to add to the in sheer day-to-day operations. To avoid a challenge. So does civil (Roman) law, which repeat of the ultimately fruitless (or even is the norm in most of the region, as it puts downright deceptive) � xation on quantita- the burden of proof on supervisors rather tive modeling that prevailed among many than on the institutions they supervise. 27 financial institutions in the United States The fact that systemic risks cannot always prior to the crisis, supervisory agencies will be anticipated and are hard to catalogue need to ensure a proper balance between ex ante in legislation (to date, unlike with quantitative and qualitative approaches. Basel II, no international standards or lists Whether qualitatively or quantitatively of systemic risks exist) raises the bar even oriented, attracting, forming, and retaining further. Many LAC countries might have the quali�ed staff to conduct such analyses constitutional constraints to enacting legis- could be a severe problem, particularly where lation that facilitates true supervisory discre- private sector demand for similar skills is tion. However, in civil law frameworks, all intense and public sector pay is limited by the relevant conditions and criteria (the so- cross-government scales. Where supervisory called casuística) need to be spelled out ex agencies are unable to attract the staff, they ante, so that creative legal drafting will be may need to form cooperative arrangements needed to give supervisors suf�cient scope with central banks. Greater cooperation for discretion in determining the systemic between the supervisory and financial sta- risks to address and actions to take. bility staffs may be required even when they Interestingly, most supervisors believe that are both housed in the central bank. In all the powers they currently have are adequate cases, traditional supervisor-economist rival- to request that a �nancial institution increase ries will need to be put aside. That said, one its capital, provisions, or liquidity based on �nds some degree of comfort that supervisors a supervisory assessment of its exposure to across the regions are well aware of the chal- systemic risks. 28 Even more surprisingly, a lenges awaiting them. 25 Some countries are majority of respondents to the World Bank– already on their way in using advanced ana- Association of Supervisors of Banks of the lytical tools, particularly where the supervi- Americas (ASBA) survey do not see the civil sory function is part of the central bank (such code as an important hindrance to the use of as Brazil) or where the supervisory agency supervisory discretion. Rather, the agency’s has recruited economists to staff a macropru- legal mandate and the weak legal protection dential unit (such as Mexico). Many of the are considered to be the main obstacles. Be smaller countries have been slower to move. that as it may, a strengthening of supervisory powers appears to be needed in many coun- tries of the region. The institutional and legal Powers frameworks in place are not likely to make Systemic supervisors will need sufficient this any easier. powers to enforce systemic regulations, including the ability to order an institution Organizational arrangements to cease and desist from activities that pose a systemic risk. Of course the more supervi- An overbearing issue in implementing sors rely on discretion, rather than rules, the systemic supervision is how to organize more their powers will need to be balanced it. Because systemic supervision, by its 228 SYSTEMIC SUPERVISION construction, involves bringing together, supervision of multiproduct � nancial � rms. from a holistic perspective, all parts of the However, coordination arrangements for � nancial system, one can take the view that consolidated supervision of �nancial groups, all supervision should be consolidated into important as they are, do not necessarily take a single agency. If so, the need for better care of systemic supervision. coordination between monetary and pru- The challenge, therefore, is to set up a suit- dential management, as discussed in chap- able systemic oversight function endowed ter 11, naturally points toward the central with appropriate governance arrangements bank. Thus, after the crisis, some Euro- for coordination and decision making. Where pean countries have leaned toward bringing � nancial oversight and monetary policy are supervision back within the central bank’s distributed among different institutions, an fold. 29 Indeed, many central banks in the alternative to placing the systemic oversight world, including in most South American responsibility with the central bank is to countries facing the Atlantic Ocean, already make it reside with a committee or council house monetary and prudential oversight made up of the relevant authorities.32 Many policies in the central bank. However, where LAC countries are well on their way to estab- this is not the case, central bankers are often lishing (or have already established), in line reticent to assume a full-blown supervisory with international experience, such � nancial task that—at least at its micro level—is stability committees to monitor and address highly resource-intensive and may be viewed systemic risk. The region’s experience with as deviating from the pure macro-oriented inflation targeting, which rests on effective analytical tasks associated with monetary monetary policy governance arrangements management. The rise of an extremely (that is, independent central bank boards, powerful centralized agency that blends clear protocols for decision making and monetary, regulatory, and supervisory poli- communication with the public, and strong cies could also create a political backlash, accountability standards), should provide a feeding pressures for limiting the central valuable guide toward setting up appropriate bank’s independence, thereby undermining governance arrangements for systemic over- monetary policy. Thus, many countries are sight committees. To ensure the committees’ leaning toward a Financial Sector Authority effectiveness, transparency and accountabil- (FSA)-type consolidation of all supervisory ity will be of the essence, including a proper activities outside the central bank.30 framework and protocols for analysis, deci- In many LAC countries, as suggested by sion making, and reporting. The need for the World Bank–ASBA survey of supervisors, strong legal grounds and governance arrange- interagency cooperation seems to require ments for the systemic and macroprudential strengthening (see chapter 10). In some coun- oversight functions is even more crucial, tries, a further potential problem is the lack considering the potential costs of the asso- of well-defined mandates. In some cases, ciated decisions on � nancial intermediaries countries had formalized arrangements that (for example, spoiling the party. Regardless had grown out of previous experience with of the exact features of the organizational � nancial crises. 31 Other countries are con- arrangement chosen, information needs to templating the creation of a uni�ed �nancial flow,33 and staff must be encouraged to work supervisor, separate from the central bank, together across agencies.34 Given the natural following Colombia’s lead. Even when the tendencies of bureaucracies, this is easier said different parts of the financial sector are than done. supervised within the same entity, internal It is important to stress that the coordina- coordination problems may arise, however. tion function of a �nancial stability commit- Thus, both Mexico and Uruguay have recently tee, focused as it is on macroprudential and reviewed the organizational structure of their systemic oversight (and where the central bank main supervisory agency to facilitate the governor should be primus inter pares among SYSTEMIC SUPERVISION 229 committee members), is of a different nature assess the systemic importance of � nancial than other coordination functions pertain- institutions, markets, and instruments. Even- ing to �nancial oversight. An example of the tually, the international �nancial institutions latter is the coordination among those regula- and the Financial Stability Board are likely tory agencies needed to perform consolidated to develop standards of good practice in sys- supervision of �nancial conglomerates, where temic oversight, in much the same way such the primus inter pares should arguably be the standards already exist for other aspects of head of the supervisory agency that oversees � nancial oversight. Given the many dif�cult the largest member of the conglomerate and issues outlined in this chapter, peer reviews where central banks that do not house regu- based on such reports on the observance latory functions need not even be a member. of standards and codes (ROSCs) should be The failure to adequately separate different helpful, once they become available. But �nancial oversight coordination functions can given the experimental features of much end up overburdening the � nancial stability that needs to be done regarding systemic committees with extraneous responsibilities, supervision, countries may not necessar- thereby undermining their effectiveness. ily want to wait until formal ROSCs have In addition to home coordination, supervi- been established. Whether through formal sors also need to coordinate across borders. ROSC assessments or just exploratory con- In LAC, the importance of foreign banks sultative missions, the region should bene�t makes this even more of a priority. As dis- from such peer reviews. In addition to con- cussed in chapter 10, significant improve- tributing to the analytical process, ROSCs ments had already taken place before the can provide the authorities with some of the global crisis. However, the crisis has shown ammunition they may need to strengthen that the arrangements of the past were their powers and independence. generally not suf�cient to deal with the cri- sis. While there is no simple solution to this coordination challenge in a globalized world, The agenda ahead LAC might be better off seeking a regional Like the rest of the world, LAC needs to approach. This approach would help to tackle a work agenda for strengthening the ensure information sharing and coordina- supervision of systemic risks that is as rich tion among regional systemic supervisors and as it is complex. It will need to do so taking would provide a uni�ed front in dealing with into account the region’s speci� c features, the large developed countries with signi�cant relative weaknesses, and relative strengths. presence in Latin America. Alternative struc- Some of the region’s speci� c features that tures might include a college of supervisors, will probably color the way this agenda is a formal regional body (possibly under the implemented include the high concentration auspices of an international �nancial institu- of ownership, substantial conglomeration of tion), or an association of systemic supervi- �nancial sector services, and strong presence sors similar to the International Association of foreign banks. In addition, the region of Deposit Insurers. Wherever needed, laws continues to be exposed to large swings in that limit information sharing across borders capital flows, partly the result of remain- would need to be amended. ing interest rate differentials with respect to world rates; it is still—albeit decreasingly— dollarized; its oversight has been strength- Peers ened substantially and can now be viewed The April 2009 G -20 “Declaration on as generally hands-on and rather restrictive; Strengthening the Financial System� tasked many of its policy makers and supervisors the IMF, the Financial Stability Board, and are familiar with dealing with systemic cri- the Bank for International Settlements to ses; and its �nancial institutions are, by and develop guidelines on how countries can large, cautious and well protected. 230 SYSTEMIC SUPERVISION Indeed, the last three of the above features for an in-depth build up of systemically ori- can generally be put on the “comparative ented supervisory capacity. strengths� side of the ledger. These strengths 3. For example, this is the conclusion reached and the fact that the region has passed the test by Stallings and Studart (2003) in their study of Latin American countries after the tequila of the recent global crisis with mostly flying crisis. colors should not be viewed as a reason for 4. Granlund (2009) �nds that in the European complacency, however. As with other regions, Union countries he looked at, most of the LAC needs to prepare itself to address the supervisory changes were focused on market challenges of systemic supervision. In doing stability and on changes in the organizational so, it will need to address some institutional arrangements for supervision. weaknesses, such as limited supervisory inde- 5. The “rules versus discretion� debate origi- pendence and legal protection and a civil code nated in the literature relating to monetary that limits the scope for discretion. In addition policy and dates back to the 1950s and 1960s. to addressing these weaknesses, key items in See, for example, Friedman (1959). LAC’s agenda for reform will include (a) �nd- 6. By contrast, according to a civil-code style of administrative law, the private sector is per- ing the proper balance in setting prudential mitted to do whatever is not prohibited by requirements between financial stability laws and regulations. and �nancial development, (b) ensuring that 7. For example, in the securities industry, the supervisory agencies can attract and retain International Organization of Securities the human resources they will need for effec- Commissions (IOSCO) has focused its stan- tive systemic supervision, (c) �nding ways to dards on market conduct regulations and rely more on market signals and market dis- consumer protection, and to a much lesser cipline, (d) �nding institutional arrangements extent on prudential standards. Consequently, that promote interagency coordination and on-site inspection, to the extent it exists at adequately support systemic oversight, and securities regulators, tends to be compliance (e) �nding the most effective ways of coordi- focused. The recent experience in the United States, where securities �rms were allowed nating supervision across borders. to operate with virtually no capital but to engage in derivatives and structured products transactions with other �nancial institutions, Notes con�rms this vulnerability. 1. This chapter draws heavily on the paper 8. While economists tend to support greater “Systemic Supervision� by Katia D’Hulster reliance on capital market instruments, such and Steven A. Seelig, which is part of the as subordinated debt, many supervisors see Edited Volume that accompanies this LAC such instruments as dif�cult to implement flagship report. As discussed in chapters 6 and believe that supervisors have in any case and 10, past banking crises in Latin America a better grasp of bank risks. For a thorough originated for the most part with unsustain- review of the literature on this topic, see able macrodynamics accompanied by the col- Evanoff and Wall (2000). lapse of �xed exchange rate regimes. In many 9. The global crisis can be viewed as having, for cases, poor risk management, weak bank the most part, vindicated these views. Despite governance, and weak supervision worsened market signals of potential problems, market the vulnerability of banking sectors to these discipline alone did not curb the buildup of shocks. See Stallings and Studart (2003). risk in the system. Only when the crisis was a 2. As reported in chapter 10, LAC supervisors reality did market discipline become so severe are, by and large, well aware of the work that it forced regulatory and supervisory ahead. The majority of supervisors who responses. responded to the World Bank–Association of 10. Measures of expectations regarding indi- Supervisors of Banks of the Americas (ASBA) vidual �nancial intermediaries include the survey believe that recent global events will change in spreads on credit default swaps and require a rede�nition of the role and func- the changes in and volatility of stock prices tions of supervisors, with a consequent need (see Sharpe 1964). SYSTEMIC SUPERVISION 231 11. An early proposal for LAC supervisors to use shock. See Hirtle, Scheurman, and Stiroh interest rates paid by �nancial institutions (2009). Similar tests were conducted in in the interbank markets is found in Rojas- Europe by the Committee of European Suarez (2001). A recent body of work has Banking Supervisors (CEBS). In mid-2010. incorporated �nance theory, market data, and the European Union (EU) announced its a contingent claims approach to examining intention to conduct additional stress tests. �nancial stability risks, an approach that has 19. Twelve respondents to the World Bank–ASBA recently been applied in Chile. See Gray and survey of supervisors con�rmed that their Malone (2008) and Gray and Walsh (2008). agencies undertake systemic stress-testing 12. On subordinated debt issuance, the clas- exercises on a regular basis while nine respon- sic proposal is that of Calomiris (1999). As dents said they only perform this exercise discussed in chapter 9, the availability of occasionally. The remaining seven countries market-based insurance for �nancial inter- do not perform systemic stress tests but do mediaries to buy may require some kind of not rule them out for the future. When regu- private-public risk-sharing arrangement. lar stress tests are carried out, the common 13. Indeed, many supervisory agencies have practice appears to be to carry them out on a recently made great strides in overcoming three- to six-month cycle, but for the banking their natural fears of opening up. For example, sector only. Authorities that do not perform since the advent of risk-based deposit insur- regular stress tests indicate that they rely on ance premiums in the United States, analysts the FSAP stress tests. are readily able to estimate which banks are 20. For an early discussion of the limits and per- viewed as more risky by the Federal Deposit ils of stress testing in a systemic context, see Insurance Corporation. Given the implica- Haldane (2009a). tions of administrative law in LAC, however, 21. Five respondents to the World Bank–ASBA such disclosure may entail a legal liability for survey stated that they are considering a spe- supervisors. ci�c requirement for �nancial institutions to 14. For example, in Uruguay, the banking supervi- assess their exposure or contribution to sys- sor (part of the central bank) publishes detailed temic risk based on stress parameters provided �nancial information about each banking �rm by the supervisor. The institutions would then on a monthly basis with a degree of detail that be required to adjust their capital and liquid- allows analysts to fully understand the �nan- ity positions accordingly. Two countries stated cial condition of individual banks. In fact, that they already apply this procedure. using the central bank’s database, one �nan- 22. One such example of a complex �nancial cial reporter was able to predict when a small conglomerate that ran into �nancial dif�cul- cooperative bank might become insolvent. ties is CL Financial in Trinidad and Tobago. 15. The coordination failures that need to be pre- This �rm had subsidiaries in insurance, bank- vented are of the type famously alluded to by ing, real estate, and real sector commercial Citibank’s former CEO Charles Prince: “You activities in nine countries. See chapter 12 in need to keep dancing until the music stops.� this report and Powell (2010). 16. For example, �nancial stability reports are 23. On quantitative approaches to map intercon- issued by central banks in Brazil, Chile, nectedness, see, for example, IMF (2010). Colombia, and Mexico. The Peruvian supervi- 24. For example, about 60 percent of respon- sors also conduct a systemic risk assessment. dents to the World Bank–ASBA survey of 17. Stress testing was formally introduced into the supervisors reported that interconnectedness supervisory process with the adoption of Basel between �nancial institutions (in contrast to II. However, the absence of data spanning full control) is a key part of their inspection pro- business cycles generally led to misleading esti- cess. But only four countries have taken this mates of capital adequacy. Moreover, market practice a step further and reported on inter- data failed to capture the crisis covariances connectedness within the �nancial system. and correlations given interconnectedness. 25. Thus, when assessing the obstacles to conduct- 18. After the global crisis, stress tests were intro- ing systemic supervision, supervisors view the duced in the United States by the Federal development of proper staff skills as the most Reserve as a way to test whether the larg- important challenge. Although 11 respon- est banks could withstand another systemic dents to the survey rate their agency’s current 232 SYSTEMIC SUPERVISION systemic supervision capacity as good relative 31. For example, Uruguay has created a coun- to international standards, 5 rate it as fair, and cil made up of the FSA, the monetary policy 10 respondents think it needs strengthening. directorate (both in the central bank), and the 26. In the World Bank–ASBA survey, only one minister of economy and �nance to review supervisory authority states that it has a fully developments and coordinate any necessary operating system of automatic corrective measures. action triggers. Although another three Latin 32. In the United States, a Financial Stability American supervisors have such systems in Oversight Council, made up of all the major place, they admit the systems need strength- �nancial sector regulators, has been created. ening. Among other responsibilities, the council 27. Thus, a supervisory instruction may in sev- will determine which �nancial �rms are “too eral LAC countries be stopped by a success- big to fail.� Similarly, in December 2010, ful court appeal until supervisors identify and the European Union created the European prove noncompliance before a judge. Systemic Risk Board. In addition, a joint 28. In the World Bank–ASBA survey of supervi- committee of European supervisory authori- sors, only six agencies report that these pow- ties was established to ensure greater coop- ers are poor or nonexistent. eration among supervisors. 29. The United Kingdom and Germany have pro- 33. In any case, the central banks’ role as lender posed shifting supervision from the Financial of last resort requires that they have real-time Sector Authority (FSA) into the central bank. access to prudential supervisory information The French have moved toward creating an on all �nancial �rms, especially in a crisis FSA under the auspices of the central bank. environment. The Bank of England’s handling 30. The case of AIG is an often-cited example of of Northern Rock during the early stages of the failure of multiple supervisors, since its the U.K. crisis provides a useful illustration of insurance activities were under the jurisdic- the problems that can emerge when real-time tion of the state insurance supervisors, its access to information is not the case. On these securities activities were under the Securities issues, see Goodhart and Schoemaker (1995) and Exchange Commission, and, through its and Peek, Rosengren, and Tootell (1999). ownership of a savings and loan, it was also 34. For example, central bank staff should be regulated by the Of�ce of Thrift Supervision. encouraged to participate in on-site inspec- Thus, there have been calls for integrating the tions while supervisory agency staff should supervision of banking, securities, and insur- be encouraged in directly participating in the ance. See Garicano and Lastra (2010). drafting of �nancial stability reports. Summary of Policy Directions for the Road Ahead 14 T his chapter concludes this flagship two roots: the much-improved macroeco- report with a summary review of nomic policy frameworks and the enhanced the main themes of research and the quality of financial oversight. Regarding main policy directions coming out of the the macroeconomic underpinnings of � nan- report. The � rst section lays the groundwork cial stability, the main challenge is to build by briefly synthesizing where Latin America effective countercyclicality into �scal policy and the Caribbean (LAC) region currently while consolidating the hard-fought gains stands as regards both the stability and in monetary policy. As regards the finan- development of its financial systems. The cial oversight underpinnings, important next section reviews the broad set of issues gaps remain in traditional bank oversight, that are likely to require further work before particularly with respect to the indepen- a speci�c policy agenda can be formulated. dence of bank supervisors and their legal The two following sections summarize protection, the regulation of capital ade- the key policy directions that can already quacy, and the consolidated supervision of be fleshed out at this stage. The � rst of those � nancial conglomerates (both domestically sections deals with the bright side, the next and across borders). At the same time, there with the dark side. Beyond those discussions, is a substantial unevenness across LAC, with this report cannot (and certainly does not) the smaller and lower-income countries gen- provide a list of country-specific, detailed erally lagging behind. Finally, LAC, as the policy recommendations. Given its regional rest of the world, faces the big and complex perspectives and holistic approach, its aim challenge of developing a robust systemic is only to raise broad issues and suggest the oversight framework. Given LAC’s tumul- directions of policy response. tuous macro� nancial history, it has to some extent a leg ahead in such matters; however, much remains to be done. Where is LAC? LAC has also made substantial progress LAC has made substantial progress as as regards financial system development. regards � nancial system stability. This prog- First, the region experienced a general �nan- ress, which in part explains the region’s resil- cial deepening, as told in chapter 2’s broad ience to the recent global � nancial crisis, has developmental story, with capital markets 233 234 SUMMARY OF POLICY DIRECTIONS FOR THE ROAD AHEAD and institutional investors playing an increas- gap today and might also (although the story ingly important role, and new markets here is admittedly more murky) explain part and instruments springing up and making of the domestic equity turnover gap. The his- inroads. Consistent with this general deep- tory of low economic growth also appears to ening, the maturities of �xed-income instru- explain part of the banking gap as well as ments have lengthened considerably, yield the relatively high interest rates in many LAC curves have extended further into the future, countries—a major hindrance to monetary and the region has experienced a broad-based, policy. Yet the past is also a source of expe- albeit not yet complete, return to the local rience and strength for the future. In par- currency (both in banking and bonds). At the ticular, the hands-on oversight and prudent same time, the patterns of � nancial global- banking with large buffers, which contrib- ization have become safer, with lower debt uted to the resilience from the global crisis, liabilities and higher reserve assets. Finally, are explained largely by history. The systemic substantial progress has been made as regards experience and the region’s active experimen- �nancial inclusion, particularly in the LAC-7 tation with systemically oriented instruments countries, which in fact now appear to be and policies are also at least partly a legacy somewhat ahead of their peers. of the past. Yet large, measurable gaps remain. First, LAC also faces important challenges on the commercial banking sector underperforms two issues regarding which the region may in terms of both its size and its efficiency, not necessarily be performing worse than it with one probably having much to do with should, but that are nonetheless central to the the other. Second, a substantial increase in sound development of its � nancial systems consumer credit seems to have occurred, and those systems’ contribution to social though much of that has been at the expense and economic prosperity. The �rst such issue of other types of lending, including in the relates to the dif�culties of establishing deep mortgage market (where LAC lags the most) markets for long-term �nance (“going long�), and in small and medium enterprise (SME) the second to the thorny rationale for risk � nance. Third, the domestic equity market bearing by the state. also underperforms in trading and liquidity, As regards LAC’s capacity to go long, in if not in capitalization. Finally, the insurance spite of the strong development of (and high industry lags in the size of both premiums and fees charged by) its asset managers, the latter assets. These gaps matter because (a) they are continue to concentrate their portfolios in the likely to limit the growth potential as well as shorter-term and/or the more liquid securi- the access to � nance (however, � nding out ties. Moreover, they trade little. Although exactly to what extent will require further the annuities industry in some countries, research, as discussed below), and (b) by such as Chile, is a potential success story of limiting intertemporal consumption smooth- how to help channel demand toward the lon- ing (as in the case of housing � nance), they ger and the less liquid, dif�culties emerge at reduce welfare. But there is also substantial the interface between pensions and annuities unevenness across the region. Indeed, look- that most countries (to a greater or smaller ing on the brighter side, there are important extent) need to address. As regards housing success stories—such as banking, corporate �nance, the budding progress achieved before bonds, and insurance in Chile; equity and the global crisis was largely interrupted by mutual funds in Brazil; and public debt in the crisis, with substantial uncertainty now Brazil, Colombia, and Mexico—that provide looming regarding the best path to follow worthy examples to study and follow. ahead. LAC’s turbulent past is an important con- As regards the risk-taking role of the state, tributor to the mentioned gaps, but not all following wide swings over the past three to for the bad. Its history of � nancial instabil- four decades, the precrisis consensus view ity appears to explain much of the banking was to limit the risk-taking role of the state, SUMMARY OF POLICY DIRECTIONS FOR THE ROAD AHEAD 235 by privatizing public banks or restricting information (a public good), induce �nancial them to second-tier operations while improv- market players and users to internalize exter- ing their � nancial discipline. Yet the impor- nalities, and help steer markets in the right tant countercyclical role played by � rst-tier directions. At the same time, markets can public banks during the global crisis has (and need to) provide key signals to help guide called this view into question, raising the pre- states’ policy actions. Ex post, states have a mium on setting the rationale for risk absorp- unique coordinating and risk absorbing role tion by the state on a sound footing. to play. Yet this role naturally breeds moral hazard and free riding. How best to man- age these tensions and complexities in LAC’s The tough issues to think about regional context is also something LAC (as A first issue on which LAC will no doubt everybody else, for that matter) will need to need to reflect further is the role of the state continue reflecting upon. in � nancial development. The � nancial par- A third issue that will warrant revisit- adigms discussed in chapters 2 and 9 high- ing is the role of finance in growth. This light the comparative advantage of the state aim will � rst involve the question of how to (or markets) in solving collective (or agency) promote the bright side of �nancial develop- frictions. Yet the close interaction between ment without also promoting the dark side. the two classes of frictions complicates the The �nance paradigms highlight the ways in � ne allocation of roles at the same time as it which the two sides can interact. And they boosts the case for complementarity and pri- suggest that, unless the demons of the dark vate-public partnership. In addition, the role side are efficiently kept in check, at some of the state in taking risk must be reassessed point more �nance might be bad (rather than based on risk aversion and collective action good) for growth. In LAC, in view of the comparative advantages, rather than on large developmental gaps, one might take the agency frictions or externalities, as has com- view that the region is far from reaching any monly been the case. And states’ intervention such threshold, even if it exists. Taking this in finance needs to be justified, not based view too strongly would be unwise, how- so much on ultimate objectives, but based ever, given the growing interconnectedness instead on the best channels to reach these and globalization of LAC � nancial systems. objectives. Indeed, if the same objectives can Moreover, one can also argue that potential be met through less intrusive interventions— perils down the road should already guide particularly by promoting participation and current policies. A second, related issue in risk spreading by the private sector without LAC concerns the two-way direction of the state assuming the risk itself—it should causality between � nance and growth. The all be for the better. However, how best to fact that LAC’s � nancial development gaps translate these general principles into LAC’s in part reflect the mediocre growth of the day-to-day financial reality remains a big past implies that much of the � xing of the challenge and, to a large extent, a theme for regional underperformance in �nance needs further thinking. to take place outside �nance, particularly in A symmetric issue, the role of the state in the growth, productivity, and competitive- � nancial stability, also will deserve further ness arenas. The history of mediocre growth thought. As explained in chapter 2 and dis- also implies a need to focus more on �nancial cussed in chapters 11 to 13, the bright and policies that can help promote growth, as the dark sides of � nance interact in much more latter will in turn help resolve the region’s complex ways than previously believed. In �nancial development gaps. managing these interactions, states and mar- Exploring the border line and trade-offs kets need to interact closely and in ways that between �nancial stability and �nancial devel- are often tense and potentially conflictive. opment is a last broad generic issue worth Ex ante, states need to provide analysis and inscribing in LAC’s research agenda. While 236 SUMMARY OF POLICY DIRECTIONS FOR THE ROAD AHEAD much has been written on stability issues of some financial services—for example, after the global crisis, very little has as yet consumer lending or microlending—they been said on the links between stability and can severely hamper other services that are development. Indeed, despite the new features arguably more crucial to economic growth of the international financial architecture and social welfare, such as the � nancing of (including the establishment of the Financial SMEs, housing, and education and the avail- Stability Board and the G-20 process), it is ability of � nancial services (such as annui- still exclusively focused on � nancial stabil- ties) to ensure old-age income security. Given ity and is thus clearly lacking in its ability to that progress on these matters will not come appropriately tackle the issues in the interface easily, the reform agendas must be designed of �nancial development and �nancial stabil- and implementation efforts must be set in ity. Finding the right balance between these motion in earnest. two dimensions, which is a global challenge, Beyond that, a research agenda needs to acquires special tones in LAC. The cur- be developed to better understand the nature rent hands-on, silo-based, broad regulatory and implications of the banking gap. To what perimeter and innovation-cautious oversight extent and in what manner are SMEs truly have served the region well. However, some affected by lack of credit? To what extent is realignment may be needed as � nancial sys- the problem instead with the lack of bankable tems continue to mature and the intensity of projects? To what extent is lack of competi- cross-border competition increases. But the tion part of the problem? If so, what can be more room LAC opens for markets to play done about it? As research sheds further light and innovations to be introduced, the more on these questions, it shows that the policy it will need to rely on a well-targeted ex ante agenda would need to be focused on promot- internalization of systemic risks and an ex ing productivity-oriented credit (� rms, stu- post capacity to provide liquidity and absorb dents, infrastructure), which might include risks. The current developmental gaps are state interventions aimed at overcoming likely to complicate �nding the proper trade- coordination failures as well as interventions off, not the least because they might feed that resort to well-targeted and well-priced pressures against the regulatory tightening credit guarantees to foster longer-term associated with Basel III. investments (including asset-backed securi- ties or infrastructure bonds). Most impor- tantly, however, sustainability is the name of Policy directions: The bright side the game: a slower but more sustainable, less Dealing with the banking gap should be �scally risky catch-up is preferable to a more the first order of business as regards the ambitious program of �nancial sector expan- developmental policy agenda. To that end, sion that ends badly. a premium should be put on continuing to As regards the equity gap, a strengthen- strengthen the contractual and institutional ing of the contractual environment will also environment, which is responsible for at help, as shown in chapter 7. But again, more least part of this gap. As discussed in chap- research is clearly needed to assess impacts ter 2, LAC has had an easier time reducing and uncover possible solutions. As regards agency frictions, based on information- impacts, the links between the lack of stock intensive technological innovations that can trading and the ef�ciency of stock price dis- be readily imported, than it has had in mak- covery need to be ascertained. And research ing the necessary improvements in enforce- is also needed to assess how the lower liquid- ment, creditor rights, or property rights, a ity of the stocks of smaller �rms affects their taller order that includes the highly complex price. As regards solutions, while the region’s task of strengthening the judiciary. Although atypically low turnover relative to the bench- such contractual and legal weaknesses may mark cannot be explained by size (because not significantly hinder the development the benchmarking carried out in chapter 3 SUMMARY OF POLICY DIRECTIONS FOR THE ROAD AHEAD 237 was all controlled for size), clearly size mat- nudge de� ned-contribution funds into mim- ters immensely when it comes to stock market icking the investment behavior of defined- development policy. With the clear exception bene�t funds could be necessary to lengthen of Brazil, this is indeed the major challenge their portfolios. In some cases, pension fund for LAC. Although regional integration of regulations may need to be revised to encour- stock exchanges might help overcome the age investments in long instruments, such major size constraints, it does not necessar- as infrastructure bonds, possibly with some ily solve the constraints associated with the partial public guarantees (see below). Clearly, small size of stock issues. Furthermore, addi- however, there is a �ne line not to be crossed tional research is needed to ascertain whether between internalizing the positive externali- regional integration of stock markets can ties of long-term finance and undermining achieve any special benefits that could not pension funds’ fiduciary responsibility by be effectively achieved through global inte- obliging them to invest in the political pet gration. There is also a need to identify the projects of the day. In view of consumers’ governance frameworks that are appropriate and workers’ bounded rationality, regula- to the larger as well as the smaller stock mar- tions that, by default, channel their savings kets. While further improvements in market into investment portfolios that are the most infrastructure are of course welcome, they appropriate for them should also generally be will probably only help at the margin. In desirable. However, the scope of state inter- the end, it might be the case that more can vention again clearly needs to be limited. A be done through venture capital funds (that proper balance must be found between pro- is, through relationship-based, non-liquid- tecting those consumers who are clearly not equity �nance) than through traditional mar- equipped to manage their portfolios and ket-based equity �nance. If so, the emphasis encouraging those who are to do so, thereby should be put on ways to promote the growth enhancing market discipline. of such funds. As regards housing finance, the tricky With respect to the goal of lengthening balance to be ascertained here is that �nancial contracts, the general strengthening between monitoring, skin-in-the-game, and of the contractual and institutional environ- risk spreading, which underlies much of the ment should again help. But, in addition, spe- current debate about covered bonds versus ci�c efforts should be geared at producing the securitization. The failure of the originate- basic information required for market com- to- distribute model of U.S. housing � nance pletion. These efforts include (a) promoting has put much of the spotlight on ensuring credible, well-accepted inflation indexes that suf�cient skin in the game and mistrusting can be used to denominate longer-term con- rating agencies (the key delegated monitor in tracts; (b) enhancing transparency, including the system). As a result, the scope for risk through more standardized information and spreading may be reduced, hindering the analysis on asset managers’ performance, in development of a deep market for mort- a way that does not exacerbate their focus on gage-backed securities and widening the short-term returns; and (c) further lengthen- scope for the development of covered bonds ing yield curves on public bonds and boosting (where skin in the game is assured, because their liquidity, so as to provide the necessary banks and their depositors keep the risk). signals for the pricing of the longer private While current efforts in many countries to instruments. strengthen the legal framework for covered There might also be room for strengthen- bonds are of course welcome, a key question ing regulations that encourage longer-term is therefore whether renewed efforts should investing. For life insurance companies, pru- be jointly pursued to reactivate the market dential regulation that encourages a match- for securitized instruments. In any event, ing of maturities may suffice. For pension the development of “mutualized� second- funds, life-cycle funds or regulations that tier � nance facilities to help spread risk and 238 SUMMARY OF POLICY DIRECTIONS FOR THE ROAD AHEAD boost economies of scale deserves further review them carefully and analyze the extent attention, particularly for the underdevel- to which they are exportable to the LAC envi- oped mortgage markets. The necessary role ronment and what can be done to facilitate of the state in limiting tail risks (at least as their introduction. Where states can play a long as the monolines are not back in busi- useful role in directly assuming risk through ness) and ensuring liquidity of last resort partial credit guarantee (PCG) programs, will also need to be revisited. care must be taken to follow four basic More generally, to facilitate participation principles: (a) program objectives, mandates, and risk spreading in a sound manner, a �rst and reporting and disclosure requirements avenue is to socialize some of the informa- need to be refocused around risk- aversion tion gathering and dissemination (that is, and more explicitly linked to the agency or to make it a public good) or to facilitate the collective action frictions with which risk use of common platforms. In some cases, aversion interacts; (b) to ensure �nancial via- the public provision of market indexes—for bility, guarantees should be adequately priced example, housing prices by city or rainfall to reflect expected losses and other costs of statistics by region—might suffice to help running the PCG programs; (c) where PCG markets develop. In other cases, the unbun- programs are self-sustaining, they should dling of collection and back-of�ce functions eventually be divested to the private sector; from portfolio management (as is currently and (d) where there are hidden risks (fat-tailed attempted in several countries for pension or systemic) that free markets cannot handle, funds) might be needed to produce the econo- these risks should be explicitly recognized mies of scale needed to raise affordability and also priced in to the extent possible. Auc- and, hence, to promote participation. tioning the guarantees can help facilitate risk Another possible avenue to reduce partici- discovery. Where information is available, pation costs (and boost returns to scale) is to public guarantees can be speci�cally targeted mandate participation. For example, combin- just outside the private risk frontier to avoid ing the existing second-pillar (mandatory) competition and to enhance complementarity pension systems in the region with policies with private ventures. to promote the development of a vibrant The risks taken by public banks when annuities industry can go a long way toward they move further away from the private promoting sustainable long-term finance. risk frontier can be bounded, and public However, the relatively complex legal, regula- governance protected, in a variety of ways. tory, and informational problems that stand Earmarked capital brought in by the state in the way of developing the market for life to cover specific risks can help state banks insurance products (including annuities) will assume risk in a responsible, bounded man- need to be removed. To further facilitate the ner while protecting the banks’ own capital link between pensions and annuities, it may from depletion. Public-private partnerships also be helpful to introduce well-designed in which the state assumes most (but not all) state guarantees on annuities. In addition, the of the risk at a fair price can help facilitate composition of asset holdings in the worker’s price and risk discovery. Recurrent strategic pension fund upon retirement should be reviews can help strengthen the governance made as similar as possible to that of the life of second-tier banks engaging in higher-risk insurance company from whom the worker activities. Although it is helpful for second- will buy his or her annuity. tier development banks to be supervised by As regards risk taking and spreading, a the bank supervisors, their supervision needs �rst avenue is to promote private sector par- to be based on different criteria and methods ticipation in guarantee schemes where risk than those used for regular banks, to appro- is mutualized, based on peer pressure. Such priately acknowledge that development banks’ schemes are thriving in Europe and other activities should complement, rather than parts of the world. The LAC region should substitute, private �nancial service providers. SUMMARY OF POLICY DIRECTIONS FOR THE ROAD AHEAD 239 Finally, as regards the countercyclical role served the region well thus far. But again, to of public banks, their increased lending in limit regulatory arbitrage and potential dis- downturns can play a useful role; this applies economies of scope, governments may need all the more so in LAC because it can take to make changes as �nancial systems mature. pressure off central banks that, in view of Should the silo approach be retained, efforts the political economy risks of reforming their will be needed to improve the oversight of the charters, would have a hard time revising large and complex � nancial conglomerates their role as lenders of last resort (more on that this approach unavoidably promotes, this below). However, it does not necessarily which currently incurs major weaknesses. follow that public banks must play an even Improvements in the oversight of conglom- greater role in normal times so as to be able erates will in turn need to be paired with a to expand lending in the bad times. Instead, revisiting and, possibly, a major overhaul of public banks that do not compete with pri- the regulatory and resolution framework for vate banks (especially those that are strictly financial conglomerates as well as the sys- second-tier) can more easily obtain from temically important �nancial entities (SIFIs). private banks the information they need to The improvements (as yet largely untested) safely raise the coverage of the guarantees that have already been introduced across during downturns. the region as regards the resolution of indi- vidual financial institutions will now need to be extended to the resolution of � nancial Policy directions: The dark side groups and SIFIs, including across borders. On the dark side of � nance, much will need Although SIFIs will undoubtedly require to be done on the regulatory front to deal tighter oversight, the region should avoid the adequately with the growing interconnect- U.S. example of formally anointing them as edness of � nancial markets and institutions. such. Instead, the intensity of supervision and The starting point should be a revisiting of tightness of regulation should be adjusted the outer perimeter of regulation. LAC’s very continuously (without sharp boundaries) broad regulatory reach has served the region according to criteria that apply to everyone. well thus far, although it has tended to put At the same time, the region will need to a heavy burden on the already stretched revamp its liquidity regulations according to human and � nancial resources of regulatory a more systemic perspective, to a large extent agencies. And LAC is likely to face additional following the emerging guidelines provided challenges as financial systems mature. by Basel III. It will also need to review its sys- Hence, policies that reduce the burden of temic liquidity access policies to better meet oversight by delegating the oversight of the the challenges of the future. The systemic unregulated to the regulated might become linkages increasingly taking place through desirable. At the same time, LAC needs to funding markets highlight the possible need find effective ways to oversee the smaller for lender of last resort (LLR) facilities to � nancial institutions, such as the credit co- support critically important markets, and not ops. One way is to leave them mostly unreg- just individual liquidity-distressed � nancial ulated but to strictly limit their funding to institutions. Although LAC central banks the contributions of their members. Another are understandably reluctant to move toward (not necessarily mutually exclusive) way to more flexible liquidity support arrangements, oversee them is to use auxiliary supervision possible avenues to manage inherent conflicts (for instance, to rely on co-op federations). might include the temporary relaxation of In that case, however, ways will need to be counterparty and collateral eligibility require- found to address the potential pitfalls of ments, the strengthening of governance such delegated oversight. arrangements, the creation of private liquid- As regards the inner perimeter, LAC’s pre- ity pools and insurance, and improved coun- vailing silo-based regulatory culture has also tercyclical operation of development banks. 240 SUMMARY OF POLICY DIRECTIONS FOR THE ROAD AHEAD Dealing with � nancial system dynamics complicate policy choices, as they make the will be another major component of LAC’s task of differentiating between healthy catch- systemic oversight reform. LAC will need ups and unsustainable booms even more to set its macroprudential policy objec- dif�cult. tives across a menu of progressively more Thus, putting in place a proper institu- ambitious goals, ranging from simply cor- tional setting should be the �rst priority. The recting the distortions brought about by region’s experience with inflation targeting traditional prudential norms to the most should provide a valuable guide toward ambitious objective of dampening exces- implementing the financial stability com- sive fluctuations, and passing through the mittees that have already been introduced in intermediate goal of simply making � nan- several countries of the region. Ensuring the cial systems more resilient to fluctuations. effectiveness, transparency, and accountabil- The goals and design of macroprudential ity of these committees will be of the essence. tools and policies will also need to refl ect In fact, de�ning who bears the �nal authority the fact that LAC’s financial cycles have and responsibility for systemically oriented been more frequent and pronounced and prudential decisions is even more crucial than have ended badly more often than in other for monetary policy in view of the potential regions. The region’s current exposure to a costs (spoiling the party). potentially lethal mix of capital inflows and LAC’s capacity to effectively implement commodity price booms further raises the the use of discretion in � nancial oversight premium on quickly establishing or consoli- will also largely be conditioned on put- dating its macroprudential capacity. On the ting in place an effective framework for brighter side, however, by limiting the sub- systemic supervision. Key components of stitutability between currencies, the floating such a framework will include the seamless exchange rate regimes that now prevail in intermingling of bottom-up microsystemic LAC should help limit regulatory arbitrage supervision with top-down macropruden- across borders. This should enhance the tial supervision, an effective mix of market scope for a more active macroprudential signals and publicly provided analysis and home policy, even when the latter is asyn- information, and the development of sophis- chronous with that of rest of the world. ticated tools to map boundaries and inter- All in all, macroprudential policy should connectedness. All of the above, including clearly not be viewed as a silver bullet. While the design of instruments to elicit market it can assist monetary policy, particularly by signals, is highly complex and will require smoothing out the potential conflicts between much work and attention. At the same time, monetary and exchange rate policies, it forming and retaining the quali�ed staff for should be viewed as a complement (not a sub- designing the tools and implementing the stitute) for monetary (or �scal) policies. reforms could be a severe problem. Where In putting together proper macropru- supervisory agencies are unable to attract dential tools, a number of design issues will the staff, they may need to form cooperative need to be addressed, many of which are arrangements with central banks. In fact, highly technical and go much beyond the greater cooperation between the supervisory scope of this report. One key architectural and � nancial stability staffs are likely to be issue deserves special mention, however. required even when they are both housed in While rules-based countercyclical prudential the central bank. This cooperation should norms in principle would be more congenial be based on appropriate decision-making to LAC’s relatively weak institutions, a sig- and interagency coordinating arrangements, ni�cant scope for discretion will unavoidably including allowing information to flow and need to be built in, given the major uncertain- encouraging staff to work together across ties surrounding the impact of such policies. agencies A strengthening of supervisory The region’s credit lags are likely to further powers will be also needed across much of SUMMARY OF POLICY DIRECTIONS FOR THE ROAD AHEAD 241 the region, all the more so where problems of supervisors, a formal regional body (pos- of supervisory legal protection and indepen- sibly under the auspices of an international dence are already present. �nancial institution), or an association of sys- The improvements in domestic coordi- temic supervisors similar to the International nation will need to be matched by similar Association of Deposit Insurers. Wherever improvements across borders. In fact, the they are in force, laws that limit informa- importance of foreign banks in LAC further tion sharing across borders will need to be puts a premium on strengthening such cross- amended. border supervisory coordination. A regional Finally, the region should continue to approach could help ensure information shar- benefit from international peer reviews ing and coordination among regional sys- that, in addition to contributing to the ana- temic supervisors and provide a uni�ed front lytical process, can provide the ammunition in dealing with the large developed countries. the authorities may need to strengthen their Alternative structures might include a college resources, powers, and independence. References Adrian, T., and H. Shin. 2010. “Liquidity and básico para la Asamblea General de ALIDE Leverage.� Journal of Financial Intermedia- 39, Curaçao, Antillas Holandesas. tion 19 (3): 418–37. Allen, F., and D. Gale. 2000. Comparing Finan- Adrian, T., and H. Shin. 2011. “Financial Inter- cial Systems. Cambridge, MA: MIT Press. mediaries and Monetary Economics.� In Angeletos, G., G. Lorenzoni, and A. Pavan. 2010. Handbook of Monetary Economics, ed. B. M. “Beauty Contests and Irrational Exuberance: Friedman and M. Woodford, vol. 3, 601–46. A Neoclassical Approach.� NBER Working San Diego, CA: Elsevier. Paper 15883, National Bureau of Economic Aghion, P., and P. Bolton. 1997. “A Theory of Research, Cambridge, MA. Trickle-Down Growth and Development.� Angelini, P., L. Clerc, V. Curdia, L. Gambacorta, Review of Economic Studies 64: 151–72. A. Gerali, A. Locarno, R. Motto, W. Roeger, S. Aizenman, J. Forthcoming. “The Impossible van den Heuvel, and J. Vlcek. 2011. “Basel III: Trinity.� In The Encyclopedia of Financial Long-Term Impact on Economic Performance Globalization, ed. G. Caprio. San Diego, CA: and Fluctuations.� Document de Travail 323, Elsevier. Banque de France, Paris. Aizenman, J., and G. K. Pasricha. 2010. “Deter- Anginer, D., A. de la Torre, and A. Ize. 2011. minants of Financial Stress and Recovery “R isk-Absor pt ion by t he St ate: W hen during the Great Recession.� Working paper, Is It Good Public Policy?� World Bank, Santa Cruz Institute for International Econom- Washington, DC. ics, University of California Santa Cruz, CA Antolin, P., R. Hinz, H. Rudolph, and J. Yermo. Akerlof, G. 1970. “The Market for ‘Lemons’: 2010. Evaluating the Financial Performance of Quality Uncertainty and the Market Mecha- Pension Funds. Washington, DC: World Bank. nism.� Quarterly Journal of Economics 84 (3): Anzoategui, D., M. S. Martínez Pería, and R. R. 488–500. Rocha. 2010. “Bank Competition in the Mid- Alexander, G., C. Eun, and S. Janakiramanan. dle East and Northern Africa Region.� Review 1987. “Asset Pricing and Dual Listing on For- of Middle East Economics and Finance 6 (2). eign Capital Markets: A Note.� Journal of Arcand, J. L., E. Berkes, and U. Panizza. 2011. Finance 42 (1): 151–58. “Too Much Finance?� Unpublished manu- ALIDE (Latin American Association of Develop- script, Graduate Institute, Geneva. ment Finance Institutions). 2009. “El Papel Arizala, F., E. Cavallo, and A. Galindo. 2009. Contracíclico de la Banca de Desarrollo Frente “Financial Development and TFP Growth: a la Crisis Económica Internacional.� Estudio Cross-Country and Industry-Level Evidence.� 243 244 REFERENCES RES Working Paper 4630, Inter-American Buffer Proposal.� Consultative document 172, Development Bank, Washington, DC. Bank for International Settlements, Basel, Armendáriz de Aghion, B., and J. Morduch. 2005. Switzerland. The Economics of Micro� nance. Cambridge, ———. 2010b. “Report and Recommendations of MA: MIT Press. the Cross-Border Bank Resolution Group.� BIS Arping, S., G. Loranth, and A. Morrison. 2010. Papers No. 162, Bank for International Settle- “Public Initiatives to Support Entrepreneurs: ments, Basel, Switzerland. Credit Guarantees versus Co-Funding.� Jour- ———. 2010c. “Proposal to Ensure the Loss nal of Financial Stability 6: 26–35. Absorbency of Regulatory Capital at the Arrow, K. J., and R. C. Lind. 1970. “Uncertainty Point of Non-Viability.� BIS Papers No. 174, and the Evaluation of Public Investment Deci- Bank for International Settlements, Basel, sions.� American Economics Review 60 (3): Switzerland. 364–78. ———. 2010d. “Basel III, A Global Regula- Arvai, Z., R. Rocha, and Y. Saadani. 2011. “A tory Framework for More Resilient Banks Review of Credit Guarantee Schemes in the and Banking Systems.� BIS Papers No. 189, Middle East and North Africa Region.� Policy Bank for International Settlements, Basel, Research Working Paper 5612, World Bank, Switzerland. Washington, DC. Beck, T., 2011. “Bank Resolution: A Concep- Ashbaugh- Skaife, H ., D. W. Collins, and tual Framework.� In Financial Regulation at R. LaFond, 2006. “The Effects of Corporate the Crossroads: Implications for Supervision, Governance on Firms’ Credit Ratings.� Jour- Institutional Design and Trade, ed. P. Deli- nal of Accounting and Economics 42 (1–2): matsis and N. Herger, 53–72. Alphen aan den 203–43. Rijn, Netherlands: Kluwer Law International. Ashcraft, A., and T. Schuermann. 2008. “7 Beck, T., and A. de la Torre. 2006. “The Basic Deadly Frictions in Subprime Mortgage Secu- Analytics of Access to Financial Services.� ritization.� PRMIA Philippines Credit Forum, Policy Research Working Paper 4026, World Asian Institute of Management, Manila. Bank, Washington, DC. Athreya, K., X. S. Tam, and E. R. Young. 2010. Beck, T., A. Demirgüç-Kunt, and R. Levine. 2007. “Loan Guarantee Programs for Unsecured “Finance, Inequality and the Poor.� Journal of Consumer Credit Markets.� Working paper, Economic Growth 12 (1): 27–49. University of Cambridge, Cambridge, U.K. Beck, T., A. Demirgüç-Kunt, and M. S. Martínez Banerjee, A., E. Duflo, R. Glennerster, and Pería. 2008. “Banking Services for Everyone? C. Kinnan. 2009. “The Miracle of Microfi- Barriers to Bank Access and Use around the nance? Evidence from a Randomized Evalua- World.� World Bank Economic Review 22 (3): tion.� Department of Economics, MIT, Abdul 397–430. Latif Jameel Poverty Action Lab, Cambridge, ———. 2011. “Bank Financing for SMEs: Evi- MA. dence Across Countries and Bank-Ownership Banerjee, A., and A. Newman. 1993. “Occupa- Types.� Journal of Financial Services Research tional Choice and the Process of Development,� 39 (1): 35–54. Journal of Political Economy 101: 274–98. Beck, T., L. F. Klapper, and J. C. Mendoza. 2010. Barajas, A., G. Dell’Ariccia, and A. Levchenko. “The Typology of Partial Credit Guarantee 2009. “Credit Booms: The Good, the Bad, and Funds around the World.� Journal of Financial the Ugly.� IMF Research Bulletin 10 (2). Stability 6 (1): 10–25. Barth, J., G. Caprio, and R. Levine. 2001. “Bank- Beck, T., and R. Levine. 2005. “Legal Institu- ing Systems around the Globe: Do Regulation tions and Financial Development.� In Hand- and Ownership Affect Performance and Sta- book of New Institutional Economics, ed. bility?� In Prudential Regulation and Super- C. Ménard and M. M. Shirley, 251–78. New vision: Why Is It Important and What Are York: Springer. the Issues? ed. F. Mishkin. Cambridge, MA: Beck, T., R. Levine, and N. Loayza. 2000. National Bureau of Economic Research. “Finance and the Sources of Growth.� Journal ———. 2008. “Bank Regulations Are Changing: of Financial Economics 58 (1–2): 261–300. For Better or Worse?� Policy Research Working Benartzi, S., and R. H. Thaler. 2007. “Heuris- Paper 4646, World Bank, Washington, DC. tics and Biases in Retirement Savings Behav- BCBS (Basel Committee on Banking Supervi- ior.� Journal of Economic Perspectives 21 (3): sion). 2010a. “Counter- Cyclical Capital 81–104. REFERENCES 245 Benavente, J. M., A. Galetovic, and R. Sanhueza. Link in the Transmission Mechanism?� BIS 2006. “Fogape: An Economic Analysis.� Working Papers 268, Bank for International Unpublished paper, World Bank, Washington, Investments, Basel, Switzerland. DC Brizzi, A. 2001. “Rural Development and Agricul- Berger, A. N., L. F. Klapper, and G. F. Udell. ture.� In Mexico: A Comprehensive Develop- 2001. “The Ability of Banks to Lend to Infor- ment Agenda for the New Era, ed. M. Guigale, mationally Opaque Small Businesses.� Journal O. Lafourcarde, and V. H. Nguyen, 319–46. of Banking and Finance 25 (12): 2127–67. Washington, DC: World Bank. Bernanke, B., and M. Gertler. 1990. “Financial Broner, F., T. Didier, A. Erce, and S. Schmukler. Fragility and Economic Performance.� Quar- 2010. “Gross Capital Flows: Dynamics and terly Journal of Economics 105 (1): 87–114. Crises.� Working Paper 1039, Bank of Spain, ———. 1995. “Inside the Black Box: The Credit Madrid. Channel of Monetary Policy Transmission.� Broner, F., G. Lorenzoni, and S. Schmukler. Journal of Economic Perspectives 9: 27–48. Forthcoming. “Why Do Emerging Economies Berstein, S., and C. Cabrita. 2010. “Los Determi- Borrow Short Term?� Journal of the European nantes de la Elección de AFP en Chile: Nueva Economic Association. Evidencia a Partir de Datos Individuales.� Brunnermeier, M., A. Crockett, C. Goodhart, A. Estudios de Economía 34 (1): 53–72. Persaud, and H. Shin. 2009. “The Fundamen- Bhattacharya, U., 1997. “Communication Costs, tal Principles of Financial Regulation.� Geneva Information Acquisition, and Voting Decisions Report on the World Economy 11, Geneva: in Proxy Contests.� Review of Financial Stud- International Center for Monetary and Bank- ies 10 (4): 1065–97. ing Studies; London: Centre for Economic Bhojraj, S., and P. Sengupta. 2003. “Effect of Policy Research. Corporate Governance on Bond Ratings and Brunnermeier, M. K., and L. H. Pedersen. 2009. Yields: The Role of Institutional Investors and “Market Liquidity and Funding Liquid- Outside Directors.� Journal of Business 76 (3): ity.� Review of Financial Studies 22 (6): 455–76. 2201–38. Birdsall, N., A. de la Torre, and F. Valencia. 2010. Burgess, R., and R. Pande. 2005. “Do Rural “The Washington Consensus: Assessing a Banks Matter? Evidence from Indian Social Damaged Brand.� Policy Research Working Banking Experiment.� American Economic Paper Series 5316, World Bank, Washington, Review 95: 780–95. DC. Caballero, R. J., and A. Krishnamurthy. 2008. Blake, D., A. Cairns, and K. Dowd. 2009. “Collective Risk Management in a Flight to “Designing a De�ned Contribution Plan: What Quality Episode.� Journal of Finance 63 (5): to Learn from Aircraft Designers.� Financial 2195–2230. Analysts Journal 65 (1): 37–42. Caballero, R. J., and P. Kurlat. 2009. “The ‘Sur- Blake, D., B. N. Lehmann, and A. Timmermann. prising’ Origin and Nature of Financial Crises: 2002. “Performance Clustering and Incentives A Macroeconomic Policy Proposal.� Prepared in the UK Pension Fund Industry.� Journal of for the Symposium on Financial Stability and Asset Management 3: 173–94. Macroeconomic Policy, Jackson Hole, WY, Bordo, M. D., and O. Jeanne. 2002a. “Boom- August 20–22. Busts in Asset Prices, Economic Instability and Calderón, C., and L. Servén. 2011. “Macro- Monetary Policy.� CEPR Discussion Papers Prudential Policies over the Cycle in Latin 3398, Centre for Economic Policy Research, America.� LAC Finance Development Flag- London. ship, Of�ce of the Chief Economist for Latin ———. 2002b. “Monetary Policy and Asset America and the Caribbean, World Bank, Prices: Does ‘Benign Neglect’ Make Sense?� Washington, DC. International Finance 5 (2): 139–64. Calderón-Colín, R., E. E. Domínguez, and Borio, C., C. Fur� ne, and P. Lowe. 2001. “Pro- M. J. Schwartz. 2008. “Consumer Confu- cyclicality of the Financial System and Finan- sion: The Choice of AFORE in Mexico.� IMF cial Stability: Issues and Policy Options.� BIS Working Paper 08/177, International Mon- Papers No. 1, Bank for International Invest- etary Fund, Washington, DC. ments, Basel, Switzerland. Calomiris, C. 1999. “Building an Incentive- Borio, C., and H. Zhu. 2008. “Capital Regulation, Compatible Safety Net.� Journal of Banking Risk-Taking and Monetary Policy: A Missing and Finance 23, 1499–1519. 246 REFERENCES Calomiris, C. W., and C. P. Himmelberg. 1994. Speeches, Bank for International Investments, “Directed Credit Programs for Agriculture Basel, Switzerland. and Industry: Arguments from Theory and Cheikhrouhou, H., R. Rocha, H. Rudolph, and Fact.� In Proceedings of the 1993 World Bank C. Thorburn. 2006. “Financial Sector Dimen- Annual Conference on Development Econom- sions of the Colombian Pension System.� Policy ics, 113–37. World Bank Economic Review Research Working Paper 5775, World Bank, (supplement), March 1. Washington, DC. Calomiris, C. W., and C. M. Kahn. 1989. “A Chinn, M., and H. Ito. 2008. “A New Measure of Theoretical Framework for Analyzing Self- Financial Openness.� Journal of Comparative Regulation of Banks.� Working paper, Uni- Policy Analysis 10 (3): 309–22. versity of Illinois, Urbana. Christiano, L., C. L. Itut, R. Motto, and M. ———. 1991. “The Role of Demandable Debt Rostagno. 2010. “Monetary Policy and Stock in Structuring Optimal Banking Arrange- Market Booms.� NBER Working Paper 16402, ments.� American Economic Review 81 (3): National Bureau of Economic Research, 497–513. Cambridge, MA. Calvo, G., A. Izquierdo, and E. Talvi. 2003. “Sud- Claessens, S., G. Dell’Ariccia, D. Igan, and den Stops, the Real Exchange Rate, and Fiscal L. Laeven. 2010. “Lessons and Policy Implica- Sustainability: Argentina’s Lessons.� NBER tions from the Global Financial Crisis.� IMF Working Paper 9828, National Bureau of Eco- Working Paper 10/44, International Monetary nomic Research, Cambridge, MA. Fund, Washington, DC. Calvo, G., and C. M. Reinhart. 2002. “Fear of Claessens, S., D. Klingebiel, and S. Schmukler. Floating.� Quarterly Journal of Economics 2006. “Stock Market Development and Inter- 107 (2): 379–408. nationalization: Do Economic Fundamentals Campbell, J., and L. Viceira. 2002. Strategic Spur Both Similarly?� Journal of Empirical Asset Allocation: Portfolio Choice for Long- Finance 13 (3): 316–50. Term Investor. Oxford, U.K.: Oxford Univer- Claessens, S., M. A. Kose, and M. E. Terrones. sity Press. 2011a. “How Do Business and Financial Caprio, G., J. Barth, R. Levine, and D. Nolle. Cycles Interact?� IMF Working Paper 11/88, 2007. “Comparative International Character- International Monetary Fund, Washington, istics of Banking.� Bank of England Confer- DC. ence Proceedings 2008, London. ———. Forthcoming. “Financial Cycles: What? Caprio, G., and P. Honohan. 2001. Finance for How? When?� In NBER International Semi- Growth: Policy Choices in a Volatile World. nar on Macroeconomics 2010, ed. L. Reichlin Washington, DC: World Bank; New York: and K. West, 303–43. National Bureau of Eco- Oxford University Press. nomic Research. National Bureau of Economic Cárdenas, M., and E. Levy-Yeyati. 2011. “Latin Research, Cambridge, MA. America Economic Perspectives: Shifting Claessens, S., R. Lee, and J. Zechner. 2003. Gears in an Age of Heightened Expectations.� “The Future of Stock Exchanges in European Washington, DC: Brookings Institution. Union Accession Countries.� Corporation of Carvalho, F. A., and C. F. Azevedo. 2008. “The London/Centre for Economic Policy Research, Incidence of Reserve Requirements in Brazil: London. Do Bank Stockholders Share the Burden?� Claessens, S., and N. Van Horen. 2008. “Loca- Journal of Applied Economics 11 (1): 61–90. tion Decisions of Foreign Banks and Institu- Castaneda, P., and H. Rudolph. 2011. “Upgrad- tional Competitive Advantage.� DNB Work- ing Investment Regulations in Second Pillar ing Paper 172, Netherlands Central Bank, Pension Systems: A Proposal for Colombia.� Amsterdam. Policy Research Working Paper 5775, World ———. 2009. “Being a Foreigner among Domes- Bank, Washington, DC. tic Banks: Asset or Liability?� MPRA Paper Cerda, R. 2006. “Movilidad en al Cartera de 13467, Universit y Librar y of Mu nich, Cotizantes por AFP: La importancia de Ser Germany. Primero en Rentabilidad.� PUC Working Clarke, G., R. Cull, and M. S. Martínez Pería. Paper 309, Pontificia Universidad Católica, 2006. “Foreign Bank Participation and Access Santiago. to Credit across Firms in Developing Coun- Chan, N. T. L. 2011. “Universal Banking— tries.� Journal of Comparative Economics 34 Hong Kong’s Perspective.� Central Banker’s (4): 774–95. REFERENCES 247 Columba, F., L. Gambacorta, and P. E. Mistrulli. De Gregorio, J., and P. Guidotti. 1995. “Financial 2009. “The Effects of Mutual Guarantee Development and Economic Growth.� World Consortia on the Quality of Bank Lending.� Development 23 (3): 433–48. MPRA Paper 17052, University Library of Deidda, L ., and B. Fattouh. 20 02 . “Non- Munich, Germany. Linearity between Finance and Growth.� Corbo, V., and K. Schmidt-Hebbel. 2003. “Mac- Economics Letters 74: 339–45. roeconomic Effects of Pension Reform in de la Torre, A., C. Calderón, T. Didier, J. Mes- Chile.� In FIAP: Pension Reforms: Results and sina, and S. Schmukler. 2010a. “From Global Challenges, International Federation of Pen- Collapse to Recovery: Economic Adjustment sion Funds Administrator, Santiago, Chile. and Growth Prospects in Latin America and ———. 2010. “The International Crisis and Latin the Caribbean.� World Bank Annual Meetings America: Growth Effects and Development Report, Washington, DC. Strategies.� CASE Network Studies and Anal- de la Torre, A., C. Calderón, T. Didier, T. Kouame, yses No. 429, Universidad Católica de Chile, M. I. Reyes, and S. Schmukler. 2011. “LAC Santiago de Chile. Success Put to the Test.� World Bank Annual Corporación Andina de Fomento. 2011. Servi- Meetings Report, Spring. Washington, DC. cios Financieros para el Desarrollo: Promovi- de la Torre, A., C. Calderón, T. Didier, E. Levy- endo el Acceso en América Latina. Reporte de Yeyati, and S. Schmukler. 2010b. “The New Economía y Desarrollo, Corporación Andina Face of Latin America and the Caribbean: de Fomento, Bogotá. Globalized, Resilient, Dynamic.� World Bank Cortes, M., M. Dijkman, and E. Gutierrez. Annual Meetings Report, Washington, DC. 2011. “Micro-Systemic Regulation: A LAC de la Torre, A., E. Feyen, and A. Ize. 2011. “Bench- Perspective.� LAC Finance Development Flag- marking LAC’s Financial Development,� LAC ship, Of�ce of the Chief Economist for Latin Finance Development Flagship, Of�ce of the America and the Caribbean, World Bank, Chief Economist for Latin America and the Washington, DC. Caribbean, World Bank, Washington, DC. Cull, R., and M. S. Martínez Pería. 2010. “For- de la Torre, A., J. C. Gozzi, and S. L. Schmukler, eign Bank Participation in Developing Coun- 2006. “Financial Development in Latin tries: What Do We Know about the Drivers America: Big Emerging Issues, Limited Policy and Consequences of This Phenomenon?� Answers,� Policy Research Working Paper Policy Research Working Paper Series 5398, 3969, World Bank, Washington, DC. World Bank, Washington, DC. ———. 2007a. “Stock Market Development Dahlquist, M., L. Pinkowitz, R. Stulz, and under Globalization: Whither the Gains from R. Williamson. 2003. “Corporate Governance Reforms?� Journal of Banking and Finance 31 and the Home Bias.� Journal of Financial and (6): 1731–54. Quantitative Analysis 38 (1): 87–110. ———. 2007b. Innovative Experiences in Access Dahlquist, M., and G. Robertsson. 2001. “Direct to Finance: Market Friendly Roles for the Foreign Ownership, Institutional Investors, Visible Hand? Brookings Institution and the and Firm Characteristics.� Journal of Finan- World Bank, Washington, DC. cial Economics 59: 413–40. de la Torre, A., E. Feyen, and A. Ize. 2011. “Bench- Debrun, X., and R. Kapoor. 2010. “Fiscal Policy marking LAC’s Financial Development,� LAC and Macroeconomic Stability: Automatic Sta- Finance Development Flagship, Of�ce of the bilizers Work, Always and Everywhere.� IMF Chief Economist for Latin America and the Working Paper 10/111, International Mon- Caribbean, World Bank, Washington, DC. etary Fund, Washington, DC. ———. 2007c. “Financial Development: Emerg- de Carvalho, A. G., and G. Pennacchi. 2011. ing and Maturing Policy Issues.� World Bank “Can a Stock Exchange Improve Corporate Research Observer 22 (1): 67–102. Behavior? Evidence from Firms’ Migration to de la Torre, A., and A. Ize. 2010. “Regulatory Premium Listings in Brazil.� Journal of Cor- Reform: Integrating Paradigms.� International porate Finance, 1–21. Elsevier B.V. Finance 13 (1): 109–39. de Grauwe, P. 2009. “The Crisis as a Paradigm ———. 2011. “Containing Systemic R isk: Shift.� In Aftershocks: Economic Crisis and Paradigm-Based Perspectives on Regulatory Institutional Choice, ed. A. Hemerijck, B. Reform.� Economía 11 (1): 25–64. Knapen, and E. van Doorne, 82–90. Amster- de la Torre, A., M. S. Martínez Pería, and S. dam: Amsterdam University Press. Schmukler. 2010. “Bank Involvement with 248 REFERENCES SMEs: Beyond Relationship Lending.� Journal ———. 2011b. “Issues on SME Financing in of Banking and Finance 34 (9): 2280–93. Chile.� Technical note prepared for the de la Torre, A., and S. Schmukler. 2004. “Cop- 2010/2011 Chilean FSA P, World Bank, ing with Risks Through Mismatches: Domes- Washington, DC. tic and International Financial Contracts for Didier, T., C. Hevia, and S. Schmukler. 2011. Emerging Economies.� International Finance “How Resilient Were Emerging Economies to 7 (3): 349–90. the Global Crisis?� Policy Research Working ———. 2007. Emerging Capital Markets and Paper 5637, World Bank, Washington, DC. Globalization: The Latin American Experi- Didier, T., R. Rigobon, and S. Schmukler. 2011. ence. Washington, DC: World Bank. “Unexploited Gains from International Delis, M., and G. Kouretas. 2011. “Interest Rates Diversification: Patterns of Portfolio Hold- and Bank Risk-Taking.� Journal of Banking ings around the World.� NBER Working and Finance 35: 840–55. Paper 16629, National Bureau of Economic Dell’Ariccia, G., and R. Marquez. 2008. “Can Research, Cambridge, MA. Cost Increases Increase Competition? Asym- Didier, T., and S. Schmukler. 2011a. “Financial metric Information and Equilibrium Prices.� Development in Latin America: Stylized Facts RAND Journal of Economics 39 (1): 144–62. and the Road Ahead.� LAC Finance Develop- de Meza, D., and D. C. Webb. 1987. “Too Much ment Flagship, Of�ce of the Chief Economist Investment: A Problem of Asymmetric Infor- for Latin America and the Caribbean, World mation.� Quarterly Journal of Economics 102 Bank, Washington, DC. (2): 281–92. ———. 2011b. “Financial Globalization: Some ———. 1999. “Wealth, Enterprise and Credit Pol- Basic Indicators for Latin America and the icy.� Economic Journal 109 (455): 153–63. Caribbean,� LAC Finance Development Flag- Demirgüç-Kunt, A., E. Feyen, and R. Levine. ship, Of�ce of the Chief Economist for Latin 2011. “The Evolving Roles of Banks and Mar- America and the Caribbean, World Bank, kets in Economic Development.� Working Washington, DC. Paper, World Bank, Washington, DC. Dimson, E., P. Marsh, and M. Staunton. 2006. Demirgüç-Kunt, A., and R. Levine. 1999. “Bank- “The Worldwide Equity Premium: A Smaller Based and Market-Based Financial Sys- Puzzle.� Paper prepared for the 33rd European tems: Cross-Country Comparisons.� Policy Finance Association Annual Meeting, Zurich, Research Working Paper 2143, World Bank, August 23−26. Washington, DC. Djankov, S., R. La Porta, F. Lopez-de-Silanes, and Demirgüç-Kunt, A., and L. Servén. 2010. “Are A. Shleifer. 2008. “The Law and Economics of All the Sacred Cows Dead? Implications of the Self-Dealing.� Journal of Financial Economics Financial Crisis for Macro and Financial Poli- 88 (3): 430–65. cies.� World Bank Research Observer 25(1): Domonski, D. 2005. “Foreign Banks in Emerging 91–124. Market Economies: Changing Players, Chang- de Nicolo, G. D., P. Honohan, and A. Ize. 2005. ing Issues.� BIS Quarterly Review, December. “Dollarization of Bank Deposits: Causes Domowitz, I., J. Glen, and A. Madhavan. 1998. and Consequences.� Journal of Banking and “International Cross-Listing and Order Flow Finance 29 (7): 1697–1727. Migration: Evidence from an Emerging Mar- Dewatripont, M., and J. Tirole. 1994. “A Theory ket.� Journal of Finance 53 (6): 2001–27. of Debt and Equity: Diversity of Securities and Drehmann, M., C. Borio, L. Gambacorta, G. Manager-Shareholder Congruence.� Quarterly Jiminez, and C. Trucharte. 2010. “Counter- Journal of Economics 109 (4): 1027–54. cyclical Capital Buffers: Exploring Options.� Diamond, D. W., and P. H. Dybvig. 1983. “Bank BIS Working Paper 317, Bank for International Runs, Deposit Insurance, and Liquidity.� Jour- Investments, Basel, Switzerland. nal of Political Economy 91: 401–19. Dupas, P., and J. Robinson. 2009. “Savings Con- Diamond, D., and R. Rajan. 2000. “A Theory straints and Microenterprise Development: of Bank Capital.� Journal of Finance 55: Evidence from a Field Experiment in Kenya.� 2431–65. NBER Working Paper 14693, National Bureau Didier, T. 2011a. “Information Asymmetries of Economic Research, Cambridge, MA. and Institutional Investor Mandates.� Policy Easterly, W. 1993. “How Much Do Distortions Research Paper 5586, World Bank, Wash- Affect Growth?� Journal of Monetary Eco- ington, DC. nomics 32: 187–212. REFERENCES 249 Edison, H., and F. Warnock. 2004. “U.S. Inves- for Assessing Country Vulnerability? Evi- tors’ Emerging Market Equity Portfolios: A dence from the 2008– 09 Global Crisis.� Security-Level Analysis.� Review of Econom- N BER Working Paper 16047, National ics and Statistics 86 (3): 691–704. Bureau of Economic Research, Cambridge, Eichengreen, B., and R. Hausmann. 1999. MA. “Exchange Rates and Financial Fragility.� In Friedman, M. 1959. A Program for Monetary Proceedings, 329–68. Kansas City, MO: Fed- Stability. New York, NY: Fordham University eral Reserve Bank of Kansas City. Press. ———. 2002. “How to Eliminate Original Finan- FSB (Financial Stability Board). 2010. “Reduc- cial Sin.� Financial Times, November 22. ing the Moral Hazard Posed by Systemi- Eichengreen, B., R. Hausmann, and U. Panizza. cally Important Financial Institutions.� FSB 2005. “The Mystery of Original Sin,� in Recommendations and Time Lines, October Other People’s Money: Debt Denomination 20. http://www.financialstabilityboard.org/ and Financial Instability in Emerging Mar- publications/r_101111a.pdf. ket Economies, ed. B. Eichengreen, and R. Galati, G., and R. Moessner. 2011. “Macro- Hausmann. Chicago: University of Chicago prudential Policy: A Literature Review.� BIS Press. Working Paper 337, Bank for International Epstein, L. G. 1999. “A De�nition of Uncertainty Settlements, Basel, Switzerland. Aversion.� Review of Economic Studies 66 (3): Gale, W. G. 1991. “Economic Effects of Fed- 579–608. eral Credit Programs.� American Economic Errunza, V. 2001. “Foreign Portfolio Equity Invest- Review 81 (1): 133–52. ments, Financial Liberalization, and Economic Galindo, A., and L. Rojas-Suarez. 2011. “Pro- Development.� Review of International Eco- visioning Requirements in Latin America: nomics 9 (4): 703–26. Where Does the Region Stand?� IDB Policy Escriva, J. L., E. Fuentes, and A. Garcia-Herrero. Brief 119, Inter-American Development Bank, 2010. “A Balance and Projections of the Expe- Washington, DC. rience in Infrastructure of Pension Funds in Galor, O., and J. Zeira. 1993. “Income Distribu- Latin America.� Research Department Work- tion and Macroeconomics.� Review of Eco- ing Paper 10/02, Banco Bilbao Vizcaya Argen- nomic Studies 60 (1): 35–52. taria, Vizcaya, Spain. Garicano, L., and R. M. Lastra. 2010. “Towards Evanoff, D. D., and L. D. Wall. 2000. “Subordi- a New Architecture for Financial Stability: nated Debt and Bank Capital Reform.� Work- Seven Principles.� Journal of International ing Paper Series 24, Federal Reserve Bank of Economic Law 13 (3): 597–621. Atlanta, GA. Geanakoplos, J. 2009. “The Leverage Cycle.� Eyzaguirre, N., R. Valdes, S. Phillips, and L. Discussion Paper 1715, Cowles Foundation for Cubeddu. 2011. “Regional Economic Out- Research in Economics, Yale University, New look: Western Hemisphere—Watching Out for Haven, CT. Overheating.� Washington, DC: International Gelos, G. 2009. “Banking Spreads in Latin Amer- Monetary Fund. ica.� Economic Inquiry 47 (4): 796–814. Ferguson, N. 2008. The Ascent of Money: A Gennaioli, N., A. Shleifer, and R. W. Vishny. Financial History of the World. New York: 2010. “Neglected Risks, Financial Innova- Penguin Press. tion, and Financial Fragility.� NBER Working Fernández de Lis, S., and A. García-Herrero. Paper 16068, National Bureau of Economic 2010. “Dynamic Provisioning: Some Lessons Research, Cambridge, MA. from Existing Experiences.� ADBI Working Giannetti, M., and S. Ongena. 2009. “Financial Paper Series No. 218, Asian Development Bank Integration and Firm Performance: Evidence Institute, Tokyo. from Foreign Bank Entry in Emerging Mar- Feyen, E., R. Lester, and R. Rocha. 2011. “What kets.� Review of Finance 13: 181–223. Drives the Development of the Insurance Sec- Gill, I., T. Packard, and J. Yermo. 2004. “Keeping tor: An Empirical Analysis Based on a Panel of the Promise of Social Security in Latin Amer- Developed and Developing Economies,� World ica.� Stanford, CA: Stanford University Press; Bank Policy Research Working Paper 5572, Washington, DC: World Bank. World Bank, Washington, DC. Glaser E., S. Johnson, and A. Shleifer. 2001. Frankel, J. A., and G. Saravelos. 2010. “Are “Coase versus the Coasians.� Quarterly Jour- Leading Indicators of Financial Crises Useful nal of Economics 116 (3): 853–99. 250 REFERENCES Goldsmith, R. 1969. Financial Structure and Gray, D., and J. P. Walsh. 2008. “Factor Model Development. New Haven, CT: Yale Univer- for Stress-testing with a Contingent Claims sity Press. Model of the Chilean Banking System.� IMF Goodhart, C. 2010a. “How Should We Regulate Working Paper 10/124, International Mon- Bank Capital and Financial Products? What etary Fund, Washington, DC. Role for ‘Living Wills’?� Revista de Economia Greenwald, B., and J. Stiglitz. 1989. “Toward a Institucional 12: 85–109. Theory of Rigidities.� American Economic ———. 2010b. “Is a Less Pro-Cyclical Financial Review 79 (2): 364–69. System an Achievable Goal?� National Insti- Haines, G., J. Madill, and A. Riding. 2007. tute Economic Review 211: 81–90. “Incrementality of SME Loan Guarantees.� ———. 2011. “The Emerging New Architecture Small Business Economics 29 (1): 47–61. of Financial Regulation.� CFS Working Paper Halac, M., and S. L. Schmukler. 2004. “Distribu- 2011/12, Center of Financial Studies, Goethe- tional Effects of Crises: The Financial Chan- Universität, Frankfurt. nel.� Economía 5(1); 1–67. Goodhart, C., and A. Persaud. 2008. “How Haldane, A. 2009a. “Rethinking the Financial to Avoid the Next Crash.� Financial Times, Network.� Speech delivered at the Financial January 30. Student Association, Amsterdam, April. Goodhart, C., and D. Schoemaker. 1995. “Insti- ———. 2009b. “Why Banks Failed the Stress tutional Separation between Supervisory and Test.� Speech presented at the Marcus-Evans Monetary Agencies.� In The Central Bank and Conference on Stress-Testing, London, Febru- the Financial System, ed. C. Goodhart, 333– ary 9–10. 413. Cambridge, MA: MIT Press. Haldane, A. 2010a. “What Is the Contribution Gordy, M., and B. Howells. 2006. “Procyclical- of the Financial System: Miracle or Mirage?� ity in Basel II: Can We Treat the Disease With- In The Future of Finance and the Theory that out Killing the Patient?� Journal of Financial Underpins It, 87–120. London: London School Intermediation 15: 395–417. of Economics Press. Gorton, G., and A. Metrick. 2010. “Securitized ———. 2010b. “The $100 Billion Question.� Banking and the Run on Repo.� Yale ICF Bank of England, London, March. Working Paper 09-14, International Center for Hanson, S., A. Kashyap, and J. Stein. 2010. “A Finance, Yale University, New Haven, CT. Macroprudential Approach to Financial Regu- Gourinchas, P. O., N. Govillot, and H. Rey. 2010. lation.� Journal of Economic Perspectives 25 “Exorbitant Privilege and Exorbitant Duty.� (1): 3–28. IMES Discussion Paper Series 10-E-20, Insti- Harding, D., and A. R. Pagan. 2002a. “A Com- tute for Monetary and Economic Studies, Bank parison of Two Business Cycle Dating Meth- of Japan, Tokyo. ods.� Journal of Economic Dynamics and Gourinchas, P. O., and H. Rey. 2007. “Interna- Control 27: 1681–90. tional Financial Adjustment.� Journal of Polit- ———. 2002b. “Dissecting the Cycle: A Method- ical Economy 115 (4): 665–703. ological Investigation.� Journal of Monetary Gourinchas, P. O., H. Rey, and K. Truempler. Economics 29: 365–81. 2011. “The Financial Crisis and the Geogra- Hargis, K. 2000. “International Cross-Listing phy of Wealth Transfers.� NBER Working and Stock Market Development in Emerging Paper 17353, National Bureau of Economic Economies.� International Review of Eco- Research, Cambridge, MA. nomics and Finance 9 (2): 101–22. Gozzi, J. C., R. Levine, and S. Schmukler. 2010. Hellwig, M. 2010. “Capital Regulation after “Patterns of International Capital Raisings.� the Crisis: Business as Usual?� Max Planck Journal of International Economics 8 (1): Institute for Research on Collective Goods, 45–57. Bonn. Granlund, P. 2009. “Supervisory Approaches and Heysen, S., and M. Auqui. 2011. “Recent Financial Market Development: Some Cor- Trends in Banking Supervision in LAC.� LAC relation Based Evidence.� Journal of Banking Finance Development Flagship, Of�ce of the Regulation 11: 6–30. Chief Economist for Latin America and the Gray, D., and S. Malone. 2008. Macro� nancial Caribbean, World Bank, Washington, DC. Risk Analysis. Wiley Finance Series. West Sus- Hirtle, B., T. Scheurman, and K. Stiroh. 2009. sex, England: John Wiley and Sons. “Macroprudential Supervision of Financial REFERENCES 251 Institutions: Lessons from the SCAP.� Staff Ioannidou, V., S. Ongena, and J. L. Peydró. 2009. Report 409, Federal Reserve Bank of New “Monetary Policy, Risk-Taking and Pricing: York, NY. Evidence from a Quasi-Natural Experiment.� Holland, A., and C. Mulder. 2006. “Can Indexed Discussion Paper 2009-31S, Tilburg Univer- Debt Absolve Original Sin? The Role of sity, Center for Economic Research, Tilburg, Inflation-Indexed Debt in Developing Local Netherlands. Currency Markets.� In Financial Dollariza- Izquierdo, A., and E. Talvi. 2011. One Region, tion: The Policy Agenda, ed. A. Armas, A. Two Speeds? Challenges of the New Economic Ize, and E. Levy-Yeyati. Palgrave Macmillan, Order for Latin America and the Caribbean. New York, NY. Washington, DC: Inter-American Develop- Holmstrom, B., and P. Milgrom. 1991. “Multi- ment Bank. task Principal-Agent Analyses: Incentive Con- Jaffee, D. M., and T. Russell. 1976. “Imperfect tracts, Asset Ownership, and Job Design.� Information, Uncertainty, and Credit Ration- Journal of Law, Economics, and Organiza- ing.� Quarterly Journal of Economics 90 (4): tion 7: 24–52. 651–66. Holmstrom, B., and J. Tirole. 1998. “Private and Jeanne, O., and A. Korinek. 2011. “Managing Public Supply of Liquidity.� NBER Working Credit Booms and Busts: A Pigouvian Taxa- Paper 5817, National Bureau of Economic tion Approach.� NBER Working Paper 16377, Research, Cambridge, MA. National Bureau of Economic Research, Honohan, P. 2008. “Partial Credit Guarantees: Cambridge, MA. Principles and Practice.� Prepared for the Con- Jenkins, H. 2000. “Commercial Bank Behavior in ference on Partial Credit Guarantees, Wash- Micro and Small Enterprise Finance.� Develop- ington, DC, March 13–14. ment Discussion Paper 741, Harvard Institute Huang, R., and L. Ratnovski. 2011. “The Dark for International Development, Cambridge, Side of Bank Wholesale Funding.� Journal of MA. Financial Intermediation 20 (2): 248–63. Jiménez, G., S. Ongena, J. L. Peydró, and J. IDB (Inter-American Development Bank). 2005. Saurina. 2008. “Hazardous Times for Mon- Unlocking Credit: The Quest for Deep and etary Policy: What Do Twenty-Three Million Stable Bank Lending. Economic and Social Bank Loans Say about the Effects of Monetary Progress Report. Washington, DC: IDB. Policy on Credit Risk-Taking?� Working Paper ———. 2007. Living with Debt: How to Limit the 0833, Bank of Spain, Madrid. Risks of Sovereign Finance. Washington, DC: Joh, S. W. 2003. “Corporate Governance and IDB. Firm Pro�tability: Evidence from Korean Firms Ilyina, A., A. Guscina, and H. Kamil. 2010. “Does before the Economic Crisis.� Journal of Finan- Procyclical Fiscal Policy Reinforce Incentives cial Economics 68 (2): 287–322. to Dollarize Sovereign Debt?� IMF Working Joint Forum. 2010. “Review of the Differentiated Paper 10/168, International Monetary Fund, Nature and Scope of Financial Regulation: Washington, DC. Key Issues and Recommendations.� Bank for IMF (International Monetary Fund). 2010. International Settlements, Basel, Switzerland. “Meeting New Challenges to Stability and Jorda, O., M. Schularick, and A. M. Taylor. 2010. Building a Safer System.� Global Financial Sta- “Financial Crises, Credit Booms, and Exter- bility Report, IMF, Washington, DC. nal Imbalances: 140 Years of Lessons.� NBER Impavido, G., E. Lasagabaster, and M. Garcia- Working Paper 16567, National Bureau of Huitron. 2010. “Competition and Asset Economic Research, Cambridge, MA. Allocation Challenges for Mandatory De�ned Kacperczyk, M., C. Sialm, and L . Zheng. Contributions Pensions: New Policy Direc- 20 08. “Unobserved Actions of Mutual tions.� World Bank, Washington, DC. Funds.� Review of Financial Studies 21 (6): Innes, R. 1991. “Investment and Government 2379–2416. Intervention in Credit Markets When There Is Kang, J. K., and R. Stulz. 1997. “Why Is There Asymmetric Information.� Journal of Public a Home Bias? An Analysis of Foreign Portfo- Economics 46 (3): 347–81. lio Equity Ownership in Japan.� Journal of ———. 1992. “Adverse Selection, Investment, and Financial Economics 46 (1): 3–28. Pro�t Taxation.� European Economic Review Karlan, D., and J. Zinman. 2010. “Expand- 36 (7): 1427–52. ing Credit Access: Using Randomized Supply 252 REFERENCES Decisions to Estimate Impacts.� Review of ———. 2007. “The External Wealth of Nations Financial Studies 23 (1): 433–64. Mark II: Revised and Extended Estimates of Kashyap, A., and J. Stein. 2004. “Cyclical Implica- Foreign Assets and Liabilities, 1970–2004.� tions of the Basel II Capital Standards.� Federal Journal of International Economics 73 (2): Reserve Bank of Chicago. Economic Perspec- 223–50. tives 28: 18–31. Lanyi, A., and R. Saracoglu. 1983. “The Impor- Kay, J. 2010. “Should We Have Narrow Bank- tance of Interest Rates in Developing Econo- ing� In The Future of Finance and the Theory mies.� Finance and Development 20 (2): that Underpins It, ed. A. Turner et al., 215–34. 20–23. London: London School of Economics. La Porta, R., F. Lopez-de-Silanes, and A. Shleifer. Keynes, J. M. 1936. The General Theory of 2002. “Government Ownership of Banks.� Employment, Interest and Money. New York: Journal of Finance 57: 265–301. Macmillan; Cambridge University Press. La Porta, R., F. Lopez-de-Silanes, A. Shleifer, Kindleberger, C. P. 1989. Manias, Panics, and and R. Vishny. 1997. “Legal Determinants of Crashes: A History of Financial Crises, rev. ed. External Finance.� Journal of Finance 52 (3): New York: Basic Books. 1131–50. King, R. G., and R. Levine. 1993. “Finance Larraín, C, and J. Quiroz. 2006. “Estudio para el and Growth: Schumpeter Might Be Right.� Fondo de Garantía de Pequeños Empresarios.� Quarterly Journal of Economics 108 (3): Banco Estado, Santiago, Chile. 717–37. Lee, R. M. G. 1999. What Is an Exchange? The Kiyotaki, N., and J. Moore. 1997. “Credit Cycles.� Automation, Management and Regulation of Journal of Political Economy 105: 211–48. Financial Markets. Oxford, U.K.: Oxford Uni- Klapper, L., 2006. “The Role of Factoring for versity Press. Small and Medium Enterprises.� Journal of Lerner, J., and P. Tufano. 2011. “The Conse- Banking and Finance 30 (11): 3111–30. quences of Financial Innovation: A Counter- Klapper, L., and I. Love. 2004. “Corporate Gov- factual Research Agenda.� NBER Working ernance, Investor Protection, and Performance Paper 16780, National Bureau of Economic in Emerging Markets.� Journal of Corporate Research, Cambridge, MA. Finance 10 (5): 703–28. Levine, R., and S. Schmukler. 2006. “Internation- Klapper, L., and R. Zaidi. 2005. “A Survey of alization and Stock Market Liquidity.� Review Government Regulation Intervention in of Finance 10 (1): 153–87. Financial Markets.� Prepared for the World ———. 2007. “Migration, Liquidity Spillovers, Development Report 2005, World Bank, and Trade Diversion: The Effects of Inter- Washington, DC. nationalization on Stock Market Activity.� Klein, M. 1996. “Risk, Taxpayers, and the Role Journal of Banking and Finance 31 (6): of Government in Project Finance.� Policy 1595–1612. Research Working Paper 1688, World Bank, Levy-Yeyati, E., 2011a. “The Currency Issue: A Washington, DC. Note on Globalization, Offshoring and Dol- Kroszner, R., and R. Shiller. 2011. Reforming larization in Latin America,� LAC Finance US Financial Markets: Reflections Before and Development Flagship, Office of the Chief Beyond Dodd-Frank. Cambridge, MA: MIT Economist for Latin America and the Carib- Press. bean, World Bank, Washington, DC. Lacker, J. M. 1994. “Does Adverse Selection Jus- ———. 2011b. “Financial Globalization in Latin tify Government Intervention in Loan Mar- America: Myth, Reality, and Policy Matters.� kets?� Federal Reserve Bank of Richmond. LAC Finance Development Flagship, Of�ce of Economic Quarterly 80 (1): 61–95. the Chief Economist for Latin America and the Laeven, L., and F. Valencia. 2008. “Systemic Caribbean, World Bank, Washington, DC. Banking Crises: A New Database.� IMF Work- Li, W. 1998. “Government Loan, Guarantee, ing Paper 08/224, International Monetary and Grant Programs: An Evaluation.� Federal Fund, Washington, DC. Reserve Bank of Richmond. Economic Quar- Lane, P. R., and G. M. Milesi-Ferretti. 2001. terly 84 (4): 25–51. “The External Wealth of Nations: Measures Licht, A. N. 1998. “Stock Market Integration in of Foreign Assets and Liabilities for Industrial Europe.� CAER II Discussion Paper No. 15, and Developing Countries.� Journal of Inter- Harvard Institute for International Develop- national Economics 55 (2): 263–94. ment, Cambridge, MA. REFERENCES 253 Lo, A. W. 2009. “Regulatory Reform in the Wake Minsky, H.P. 1975. John Maynard Keynes. New of the Financial Crisis of 2007–2008.� Journal York: Columbia University Press. of Financial Economic Policy 1 (1): 4–43. Mishkin, F. 2011. “Monetary Policy Strategy: Lombra, R., and M. Wasylenko. 1984. “The Sub- Lessons from the Crisis.� NBER Working sidization of Small Businesses through Federal Paper 16755, National Bureau of Economic Credit Programs: Analytical Foundations.� Research, Cambridge, MA. Journal of Economics and Business 36 (2): Montoro, C., and R. Moreno. 2011. “The Use of 263–74. Reserve Requirements as a Policy Instrument Lorenzoni, G. 2008. “Inef�cient Credit Booms.� in Latin America.� BIS Quarterly Review, Review of Economic Studies 75 (3): 809–33. March, 53–65. Mankiw, N.G. 1986. “The Allocation of Credit Montoro, C., and C. Tovar. 2010. “Macropru- and Financial Collapse.� Quarterly Journal of dential Tools: Assessing the Implications of Economics 101 (3): 455–70. Reserve Requirements in a DGSE Model.� Martínez Pería, M. S. 2011. “Financial Inclusion Draft, Bank for International Settlements. in Latin America.� LAC Finance Develop- Morrison & Foerster. 2011. “ICB Interim Report ment Flagship, Of�ce of the Chief Economist on UK Banking Reform.� News Bulletin, for Latin America and the Caribbean, World April 29. Bank, Washington, DC. North, D. C. 1990. Institutions, Institutional McA ndrews, J., and C . Stefanadis. 20 02 . Change, and Economic Performance. Cam- “The Consolidation of European Stock bridge, U.K.: Cambridge University Press. Exchanges.� Current Issues in Economics Olivares, J. A. 2005. “Investment Behavior of the and Finance 8 (6). Chilean Pension Funds.� European Conference McKinnon, R. I. 1973. “Money and Capital in Paper 360419, European Financial Manage- Economic Development.� Brookings Institu- ment Association Conference (EFMA), Siena, tion, Washington, DC. Italy. June 8–11. McMillan, M., and D. Rodrik 2011. “Global- Opazo, L., C. Raddatz, and S. Schmukler. 2009. ization, Structural Change, and Productivity “The Long and the Short of Emerging Market Growth.� Debt.� Policy Research Working Paper 5056, NBER Working Paper 17143, National Bureau of World Bank, Washington, DC. Economic Research, Cambridge, MA. Ordover, J., and A. Weiss. 1981. “Information Merton, R. C. 1969. “Lifetime Portfolio Selec- and the Law: Evaluating Legal Restrictions on tion under Uncertainty: The Continuous-Time Competitive Contracts.� American Economic Case.� Review of Economics and Statistics 51 Review 71 (2): 399–404. (3): 247–57. Ostry, J., A. Ghosh, K. Habermeier, L. Laeven, Merton, R., and Z. Bodie. 2004. “The Design M. Chamon, M. Qureshi, and A. Kokenyne. of Financial Systems: Towards A Synthesis 2011. “Managing Capital Inflows: What of Function and Structure.� NBER Working Tools to Use?� IMF Staff Discussion Note, Paper 10620, National Bureau of Economic International Monetary Fund, Washington, Research, Cambridge, MA. DC. Mesquita, M., and M. Torós. 2010. “Brazil and Panzar, J. C., and J. N. Rosse. 1987. “Testing for the 2008 Panic.� In “The Global Crisis and ‘Monopoly’ Equilibrium.� Journal of Indus- Financial Intermediation in Emerging Market trial Economics 35 (4): 443–56. Economies,� BIS Paper 54, Bank for Interna- Park, Y. 2011. “The Role of Macroprudential tional Settlements, Basel, Switzerland. Policy for Financial Stability in East Asia’s Mian, A. 2003. “Foreign, Private Domestic, Emerging Economies.� ADBI Working Paper And Government Banks: New Evidence from 284, Asian Development Bank Institute, Emerging Markets.� Working paper, Graduate Tokyo. School of Business, University of Chicago, IL. Peek, J., E. S.Rosengren, and G. M. B. Tootell. ———. 2006. “Distance Constraints: The Limits 1999. “Is Bank Supervision Central to Central of Foreign Lending in Poor Countries.� Jour- Banking?� Quarterly Journal of Economics nal of Finance 61 (3): 1465–1505. 114 (2): 629–53. Micco, A., and U. Panizza. 2005. “Public Banks in Penner, R. G, and W. L. Silber. 1973. “The Inter- Latin America.� Paper prepared for the confer- action between Federal Credit Programs and ence on Public Banks in Latin America: Myth the Impact on the Allocation of Credit.� Amer- and Reality, Washington, DC. February 25. ican Economic Review 63 (5): 838–52. 254 REFERENCES Perotti, E., and J. Suárez. 2011. “A Pigouvian Repullo, R., and J. Suarez. 2009. “The Procycli- Approach to Liquidity Regulation.� CEPR cal Effects of Basel II.� CEPR Discussion Paper Discussion Paper 8271, Centre for Economic 6862, Centre for Economic Policy Research, Policy Research, London. London. Porzecanski, A. C. 2009. “Latin America: The Rioja, F., and N. Valev. 2004. “Does One Size Missing Financial Crisis.� Washington Of�ce Fit All? A Reexamination of the Finance and Studies and Perspectives Series No. 6. Eco- Growth Relationship.� Journal of Develop- nomic Commission for Latin America, United ment Economics 74 (2): 429–47. Nations Economic Commission for Latin Robinson, M. S. 2001. The Micro Finance Rev- America, New York. olution: Sustainable Finance for the Poor. Powell, A. 2010. “LAC Central Banks ahead Washington, DC: World Bank. of the Curve? Financial Sector Regulation Rocha, R., D. Vittas, and H. P. Rudolph. 2011. and Supervision.� Presentation at the Inter- Annuities and Other Retirement Products: American Development Bank meeting in Designing the Payout Phase. Washington, DC: Lacea, Medellin, Colombia, November. World Bank. Raddatz, C. 2011. “Institutional Investors and Rocha, R., and C. Thorburn. 2006. Developing Agency Issues in Latin American Financial Annuities Markets: The Experience of Chile. Markets: Issues and Policy Options.� LAC Washington, DC: World Bank. Finance Development Flagship, Of�ce of the Rojas-Suarez, L. 2001. “Rating Banks in Emerg- Chief Economist for Latin America and the ing Markets: What Credit Rating Agencies Caribbean, World Bank, Washington, DC. Should Learn from Financial Indicators.� Raddatz, C. and S. Schmukler. 2008. “Pension Working Paper 01-6, Institute for International Funds and Capital Market Development: How Economics, Washington, DC. Much Bang for the Buck?� Policy Research Roubini, N., and X. Sala-i-Martin. 1992. “Finan- Working Paper 4787, World Bank, Washing- cial Repression and Economic Growth.� Jour- ton, DC. nal of Development Economics 39 (1): 5–30. ———. 2011. “Deconstructing Herding: Evidence Rudolph, H. 2009. “State Financial Institutions: from Pension Fund Investment Behavior.� Mandates, Governance and Beyond.� World Policy Research Working Paper 5700, World Bank Working Paper 5141, World Bank, Bank, Washington, DC. Washington, DC. Raith, M. G., T. Staak, and C. Starke. 2006. “The Samuelson, P.A. 1969. “Lifetime Portfolio Selec- Goal Achievement of Federal Lending Pro- tion by Dynamic Stochastic Programming.� grams.� Working Paper Series 06019, Faculty Review of Economics and Statistics 51 (3): of Economics and Management, Magdeburg, 239–46. Germany. Santana, M. H. 2008. “The Novo Mercado.� Rajan, R. 2010. Fault Lines: How Hidden Frac- In Focus 5—Novo Mercado and Its Follow- tures Still Threaten the World Economy. Princ- ers: Case Studies in Corporate Governance eton, NJ: Princeton University Press. Reform, ed. M. H. Santana, M. Ararat, P. Rajan, R., and L. Zingales. 2001. “Financial Alexandru, B. B. Yurtoglu, and M. R. Cunha, Systems, Industrial Structure, and Growth.� 2–39. Washington, DC : International Finance Oxford Review of E conomic Policy 17: Corporation. 467–82. Saurina, J. 2011. “Working Macroprudential ———. 2003. Saving Capitalism from the Capi- Tools.� In Macroprudential Regulatory Poli- talists: Unleashing the Power of Financial cies: The New Road to Financial Stability, pro- Markets to Create Wealth and Spread Oppor- ceedings of the Thirteenth Annual International tunity. New York: Crown Business Division of Banking Conference, Chicago, IL, September Random House. 23–24, 2010. Repullo, R., and J. Saurina. 2011. “The Counter- Schularick, M., and A. M. Taylor 2009. “Credit cyclical Capital Buffer of Basel III: A Critical Booms Gone Bust: Monetary Policy, Lever- Assessment.� Discussion Paper 8304, Centre age Cycles and Financial Crises, 1870–2008.� for Economic Policy Research, London. NBER Working Paper 15512, National Bureau Repullo, R., J. Saurina, and J. Trucharte. 2010. of Economic Research, Cambridge, MA. “Mitigating the Pro-Cyclicality of Basel II.� Seelig, S. A., and A. Novoa. 2009. “Governance CEPR Discussion Paper 7382, Centre for Eco- Practices at Financial Regulatory and Supervi- nomic Policy Research, London. sory Agencies.� IMF Working Paper 09/135, REFERENCES 255 International Monetary Fund, Washington, Stallings, B., and R. Studart. 2003. “Financial DC. Regulation and Supervision in Emerging Mar- Sengupta, R., and C. P. Aubuchon. 2008. “The kets: The Experience of Latin America since Micro Finance Revolution: An Overview.� the Tequila Crisis.� Economic Commission for Federal Bank of St. Louis Review January/ Latin America and the Caribbean (CEPAL), February. Santiago, Chile. Sharpe, W. 1964. “Capital Asset Prices: A Theory ———. 2006. Finance for Development: Latin of Market Equilibrium under Conditions of America in Comparative Perspective. Wash- Risk.� Journal of Finance 19: 425–42. ington, DC: Brookings Institution; Santi- Shaw, E. 1973. Financial Deepening in Economic ago, Chile: Economic Commission for Latin Development. Oxford, U.K.: Oxford Univer- America and the Caribbean (Spanish). sity Press. Stiglitz, J. E., J. Vallejo, and Y. C. Park. 1993. Shiller, R. 2006. Irrational Exuberance. Princ- “The Role of the State in Financial Markets.� eton, NJ: Princeton University Press. In “Annual Conference on Development Eco- ———. 2008. Subprime Solution: How Today’s nomics� (supplement). World Bank Research Global Financial Crisis Happened and What Observer 19: 1–61. to Do about It. Princeton, NJ: Princeton Uni- Stiglitz, J. E., and A. Weiss. 1981. “Credit Ration- versity Press. ing in Markets with Imperfect Information.� Shin, H. S. 2010. Risk and Liquidity. Clarendon American Economic Review 71 (3): 393–410. Lectures in Finance. Oxford, U.K.: Oxford Stulz, R. 1999. “Globalization, Corporate Finance University Press. and the Cost of Capital.� Journal of Applied Shin, H., and Y. Shin. 2011. “Procyclicality Corporate Finance 12 (3): 8–25. and Monetary Aggregates.� NBER Working Svensson, L. 2010. “Inflation Targeting.� NBER Paper 16836, National Bureau of Economic Working Paper 16654, National Bureau of Research, Cambridge, MA. Economic Research, Cambridge, MA. Shleifer, A., and R. W. Vishny. 1997a. “A Survey Taylor, M. 1995. “Twin Peaks: A Regulatory of Corporate Governance.� Journal of Finance Structure for the New Century.� Centre for the 52 (2): 737–83. Study of Financial Innovation, London. ———. 1997b. “The Limits of Arbitrage.� Jour- ———. 2009. “The Road from “Twin Peaks’ and nal of Finance 52 (1): 35–55. the Way Back.� Connecticut Insurance Law ———. 2011. “Fire Sales in Finance and Macro- Journal 16 (1). economics.� Journal of Economic Perspectives Terrier, G., R. Valdés, C. E. Tovar, J. Chan-Lau, 25 (1): 29–48. C. Fernández-Valdovinos, M. García Escrib- Smith, B. D., and M. J., Stutzer. 1989. “Credit ano, C. Medeiros, M. K.. Tang, M. Vera Mar- Rationing and Government Loan Programs: A tin, and C. Walker. 2011. “Policy Instruments Welfare Analysis.� Real Estate Economics 17 to Lean against the Wind in Latin America� (2): 177–93. IMF Working Paper 11/159, International Sosa, S. 2010. “The Influence of ‘Big Brothers’: Monetary Fund, Washington, DC. How Important are Regional Factors for Uru- Thompson, J. 2009. “On Systemically Important guay?� IMF Working Paper 10/60, Interna- Financial Institutions and Progressive Systemic tional Monetary Fund, Washington, DC. Risk Migration.� Policy Discussion Paper No. Spence, M. 1973. “Job Market Signaling.� Quar- 27, Federal Reserve Bank of Cleveland. terly Journal of Economics 87(3): 355–74. Turner, A. 2010. “What Do Banks Do? Why Do Squam Lake Working Group on Financial Regula- Credit Booms and Busts Occur and What Can tion. 2009. “Improving Resolution Options for Public Policy Do about It?� In The Future of Systemically Relevant Financial Institutions.� Finance and the Theory that Underpins It. Working Paper, Center for Economic Studies, Center for Economic Performance. London: Council on Foreign Relations, Washington, London School of Economics. DC. Vargas, H., C. Varela, Y. Betancourt, and N. Srinivas, P. S., E. Whitehouse, and J. Yermo. Rodriguez. 2010. “Effects of Reserve Require- 2000. “Regulating Private Pension Funds’ ments in an Infl ation Targeting Regime: The Structure, Performance, and Investments: Case of Colombia.� Borradores de Economía Cross-Country Evidence.� Social Protec- 587, Banco de la República. tion Discussion Paper 23302, World Bank, Viceira, L. 2010. “Application of Advances in Washington, DC. Financial Theory and Evidence to Pension Fund 256 REFERENCES Design in Developing Economies.� In Evalu- World Bank. 1989. World Development Report ating the Financial Performance of Pension Washington, DC: World Bank. Funds, ed. R. Hinz, H. Rudolph, P. Antolin, ———. 20 01. Fin ance for Growth: Policy and J. Yermo. Washington, DC: World Bank. Choices in a Volatile World. Washington, Viñals, J. 2011. “The Do’s and Don’t’s of Macro- DC: World Bank; New York: Oxford Univer- prudential Policy.� Paper presented at the sity Press. European Commission and ECB Conference ———. 2005. Economic Growth in the 1990s: on Financial Integration and Stability, Brus- Learning from a Decade of Reform. World sels, May 2. Bank: Washington, DC. Weitzman, M. L. 1974. “Prices vs. Quantities.� ———. 2006. Emerging Capital Markets and Review of Economic Studies 41: 477–91. Globalization: The Latin American Experi- Wezel, T. 2010. “Dynamic Loan Loss Provisions in ence. Washington, DC: World Bank. Uruguay.� IMF Working Paper 10/125, Inter- ———. 2007. Finance for All? Policies and Pit- national Monetary Fund, Washington, DC. falls in Expanding Access. Washington, DC: Williamson, S. D. 1994. “Do Informational Fric- World Bank. tions Justify Federal Credit Programs?� Journal Yunus, M. 2003. Banker to the Poor: Micro- of Money, Credit and Banking 26 (3): 523–44. Lending and the Battle Against World Pov- Woodford, M. 2011. “Inflation Targeting and erty. New York: Public Affairs. Financial Stability.� Keynote address at the Zervos, S. 2004. “The Transaction Costs of Pri- conference “The Future of Monetary Policy,� mary Issuance: The Case of Brazil, Chile and Einaudi Institute for Economics and Finance, Mexico.� Policy Research Working Paper Rome, September 2010 (Revised August 2011). 3424, World Bank, Washington, DC. ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. The • 14 trees Office of the Publisher has chosen to print • 5 million BTUs of Financial Development in Latin America total energy and the Caribbean: The Road Ahead on recy- • 6,140 lbs. of net cled paper with 50 percent postconsumer fiber greenhouse gases in accordance with the recommended stand- • 12,504 gallons of ards for paper usage set by the Green Press waste water Initiative, a nonprofit program supporting • 389 lbs. of publishers in using fiber that is not sourced solid waste from endangered forests. For more informa- tion, visit www.greenpressinitiative.org. D uring the 1980s and 1990s, financial sectors were the Achilles’ heel of economic development in Latin America and the Caribbean (LAC). Since then, these sectors have grown and deepened, becoming more integrated and competitive, with new actors, markets, and instruments springing up and finan- cial inclusion broadening. To crown these achievements, the region’s financial systems were left largely unscathed by the global financial crisis of 2008–09. Now that the successes of LAC’s macro- financial stability are widely recognized and tested, it is high time for an in-depth stocktaking of what remains to be done. Financial Development in Latin America and the Caribbean: The Road Ahead provides both a stocktaking and a forward-looking assessment of the region’s financial development. Rather than going into detail about sector-specific issues, the report focuses on the main architectural issues, overall perspectives, and intercon- nections. The report’s value added thus hinges on its holistic view of the development process, its broad coverage of the financial services industry beyond banking, its emphasis on benchmarking, its systemic perspective, and its explicit effort to incorporate the lessons from the recent global financial crisis. Financial Develop- ment in Latin America and the Caribbean: The Road Ahead builds on and complements several overview studies on financial devel- opment in both LAC countries and the developing world that were published in the past decade. It will be of interest to policy makers and financial analysts interested in improving the financial sector in the LAC region. ISBN 978-0-8213-8847-1 SKU 18847