Maintaining Stability in a Changing Financial System

Foreword

In recent decades, financial systems have transformed from traditional bank-based structures of financial intermediation to market-based systems. These dramatic changes offer investors more investment options and the ability to manage their risk exposures more effectively. The recent financial crisis, however, represents the first real stress test of the stability and integrity of this new financial structure. The past year has revealed some critical strengths and weaknesses of the current global financial system. Central banks around the world have used a variety of tools to limit the spillover of this financial disruption to the broader global economy. To better understand the current financial crisis, and the implications for monetary and regulatory policy, the Federal Reserve Bank of Kansas City sponsored a symposium, “Maintaining Stability in a Changing Financial System,” at Jackson Hole, Wyo., from Aug. 21-23, 2008. The symposium allowed central bankers, financial market participants, and academics to discuss the current financial crisis and its implications for the overall macro economy.

vii

08 Book.indb 7

2/13/09 3:58:09 PM

viii

We greatly appreciate the efforts of the authors, discussants, panelists, and participants for their contribution to the symposium. Special thanks go to Federal Reserve Bank of Kansas City staff members who helped plan and arrange the symposium.

Thomas M. Hoenig President and Chief Executive Officer Federal Reserve Bank of Kansas City

08 Book.indb 8

2/13/09 3:58:09 PM

The Contributors Franklin Allen, Professor, The Wharton School, University of Pennsylvania Mr. Allen is the Nippon Life Professor of Finance and Economics at the Wharton School of the University of Pennsylvania, as well as the co-director of the Wharton Financial Institutions Center. From 2001 to 2004, he was an adjunct professor of finance at New York University. He has held visiting appointments at the University of Oxford, University of Tokyo, University of Frankfurt, and Princeton University. He has been a visiting scholar at the Federal Reserve Banks of New York and Philadelphia. He holds a directorship position with the Glenmede Fund and the Glenmede Portfolios. Mr. Allen has numerous teaching awards and honors. He has held numerous positions in academic associations and editorial positions. His areas of interest include corporate finance, asset pricing, comparing financial systems, and financial crises. He has written several books on finance and financial crises and has published extensively in professional journals. Ben S. Bernanke, Chairman, Board of Governors of the Federal Reserve System Mr. Bernanke took office in February 2006 as chairman and a member of the Board of Governors of the Federal Reserve System. He also serves as chairman of the Federal Open Market Committee. From June 2005 to January 2006, he was chairman of the President’s Council of Economic Advisers. He also was a visiting scholar at the ix

08 Book.indb 9

2/13/09 3:58:10 PM

x

The Contributors

Federal Reserve Banks of Philadelphia, Boston, and New York, and a member of the Academic Advisory Panel at the Federal Reserve Bank of New York. Previously, Mr. Bernanke was the Class of 1926 Professor of Economics and Public Affairs at Princeton University. He also was the Howard Harrison and Gabrielle Snyder Beck Professor of Economics and Public Affairs and chair of the Economics Department at Princeton. Before arriving at Princeton, he was an associate professor of economics from 1983 to 1985 and an assistant professor of economics from 1979 to 1983 at the Graduate School of Business at Stanford University. He has published many articles on a wide variety of economic issues and is the author of several scholarly books and two textbooks. He has held a John Simon Guggenheim Memorial Foundation fellowship and an Alfred P. Sloan fellowship, and he is a fellow of the Econometric Society and of the American Academy of Arts and Sciences. He served as the director of the Monetary Economics Program of the National Bureau of Economic Research (NBER) and as a member of NBER’s Business Cycle Dating Committee. In 2001, he was appointed the editor of the American Economic Review. Alan S. Blinder, Professor, Princeton University Mr. Blinder is the Gordon S. Rentschler Memorial Professor of Economics and Public Affairs at Princeton University and co-director of Princeton’s Center for Economic Policy Studies, which he founded in 1990. He is also vice chairman of the Promontory Interfinancial Network. From 1994 until 1996, Mr. Blinder served as vice chairman of the Board of Governors of the Federal Reserve System. Prior to that position, he served on President Clinton’s Council of Economic Advisers. He is the author or co-author of 17 books, numerous articles, and a column in The New York Times Sunday business section. He appears frequently on PBS, CNBC, CNN, Bloomberg TV, and elsewhere. He is a member of the Bretton Woods Commitee, the Bellagio Group, the Council on Foreign Relations, and a former governor of the American Stock Exchange. He also serves on academic advisory panels for the Federal Reserve Bank of New York, the Bank for International Settlements,

08 Book.indb 10

2/13/09 3:58:10 PM

The Contributors

xi

the FDIC Center for Financial Research, and the Hamilton Project. He has been elected to the American Philosophical Society and the American Academy of Arts and Sciences. Michael Bordo, Professor, Rutgers University Mr. Bordo is a professor in the Economics Department at Rutgers University. He previously was the Pitt Professor of American History and Institutions at Cambridge University, where he was also a fellow of Kings College. Other teaching positions include the University of South Carolina and Carleton University in Ottawa, Canada. He has been a visiting professor at Harvard University, a visiting fellow at Princeton University, and a visiting scholar at the Federal Reserve Board of Governors and several Reserve Banks, among other organizations. Mr. Bordo is a research associate of the National Bureau of Economic Research and editor and board member for various organizations and publications. He has published extensively. Willem H. Buiter, Professor, European Institute, London School of Economics and Political Science Mr. Buiter is a professor of European Political Economy at the London School of Economics and Political Science, as well as a professor of economics at the University of Amsterdam in the Netherlands. He previously taught at the University of Cambridge, where he was a fellow of Trinity College from 1994 to 2000. He has taught at Yale University and Princeton University, among others. From 2000 until 2005, Mr. Buiter was chief economist and special counsellor to the president of the European Bank for Reconstruction and Development. He was an external member of the Bank of England’s Monetary Policy Committee from 1997 until 2000. He has been adviser to and consultant for the IMF, the World Bank, the InterAmerican Development Bank, the European Commission, and a number of national governments and private financial enterprises. Mr. Buiter has written several books on international macroeconomics and has published extensively in professional journals.

08 Book.indb 11

2/13/09 3:58:10 PM

xii

The Contributors

Charles W. Calomiris, Professor, Graduate School of Business, Columbia University Mr. Calomiris is the Henry Kaufman Professor of Financial Institutions at the Columbia University Graduate School of Business and a professor at Columbia’s School of International and Public Affairs. He also is the academic director of the Chazen Institute of International Business and of the Center for International Business Economics and Research at Columbia. Additionally, he co-directs the Project on Financial Deregulation and is the Arthur Burns Scholar in International Economics, both at the American Enterprise Institute. Mr. Calomiris is a research associate of the National Bureau of Economic Research. He was a member of the Shadow Financial Regulatory Committee and a senior fellow at the Council on Foreign Relations. He serves or has served as a consultant or visiting scholar to numerous organizations and governments. His research spans several areas, including banking, corporate finance, financial history, and monetary economics. He is or has been a member of the editorial boards of numerous academic journals and has been widely published. Elena Carletti, Professor, Center for Financial Studies, University of Frankfurt Ms. Carletti is an associate professor in the Department of Finance at the University of Frankfurt. Her previous positions include assistant professor at the University of Mannheim and tutorial fellow in finance at the London School of Economics. From 1997 to 1998, she was an economist with the Italian Antitrust Authority in Rome. Ms. Carletti’s research interests include banking, financial stability, industrial organization, and competition policy. Currently, she is working on the relationship between competition and stability in banking, bank consolidation, and the implications of the structure of policy institutions for the efficiency and stability of credit markets. Her work has been published in books, conference volumes, and professional journals.

08 Book.indb 12

2/13/09 3:58:10 PM

The Contributors

xiii

Mario Draghi, Governor, Bank of Italy Mr. Draghi is the governor of the Bank of Italy. He is a member of the Governing and General Councils of the European Central Bank and a member of the board of directors of the Bank for International Settlements. He also serves on the boards of governors of the International Bank for Reconstruction and Development and the Asian Development Bank. In 2006, he was elected chairman of the Financial Stability Forum. Prior to his position at the Bank of Italy, Mr. Draghi’s professional experience includes work at Goldman Sachs International, the Italian Treasury, the European Economic and Financial Committee, and the World Bank. He is also on the Board of Trustees of the Princeton Institute for Advanced Study and the Brookings Institution. He was an Institute of Politics fellow at the Kennedy School of Government at Harvard University. He has authored and edited publications on macroeconomics and financial issues. Martin Feldstein, Professor, Harvard University, and President Emeritus, National Bureau of Economic Research Mr. Feldstein is the George F. Baker Professor of Economics at Harvard University and former president and chief executive officer of the National Bureau of Economic Research. From 1982 through 1984, he was chairman of the Council of Economic Advisers and President Reagan’s chief economic adviser. He served as president of the American Economic Association in 2004. In 2006, President Bush appointed him to be a member of the President’s Foreign Intelligence Advisory Board. Mr. Feldstein is a member of the American Philosophical Society, a corresponding fellow of the British Academy, a fellow of the Econometric Society, and a fellow of the National Association of Business Economics. He is also a member of the Trilateral Commission, the Council on Foreign Relations, the Group of 30, and the American Academy of Arts and Sciences.

08 Book.indb 13

2/13/09 3:58:10 PM

xiv

The Contributors

Mr. Feldstein has received honorary doctorates from several universities and is an honorary fellow of Nuffield College, Oxford. In 1977, he received the John Bates Clark Medal of the American Economic Association. He is a director of two corporations (American International Group and Eli Lilly), an economic adviser to several businesses and government organizations, and a regular contributor to The Wall Street Journal. Stanley Fischer, Governor, Bank of Israel Mr. Fischer has been the governor of the Bank of Israel since 2005. From 2002 through 2005, he was the vice chairman of Citigroup, where he also was the head of the Public Sector Group, the chairman of the Country Risk Committee, and the president of Citigroup International. He was the first deputy managing director of the International Monetary Fund (IMF) from 1994 through 2001. Before he joined the IMF, he was the Killian Professor and the head of the Department of Economics at the Massachusetts Institute of Technology. From 1988 to 1990, he was the vice president of Development Economics and the chief economist at the World Bank. Mr. Fischer is a fellow of the Econometric Society and the American Academy of Arts and Sciences; a member of the Council on Foreign Relations, the Group of 30, and the Trilateral Commission; a John Simon Guggenheim Memorial Foundation fellow; and a research associate of the National Bureau of Economic Research (NBER). He has served on the boards of the Institute for International Economics, Women’s World Banking, and the International Crisis Group. He is the author or editor of several macroeconomics books and has published extensively in professional journals. From 1986 to 1994, he was the editor of the NBER Macroeconomics Annual. Peter R. Fisher, Managing Director and Co-Head of Fixed Income, BlackRock Mr. Fisher has served as the managing director and co-head of fixed income at BlackRock investment management firm since 2004. Prior to joining BlackRock, he worked from 2001 to 2003 as under secretary of the Treasury, where he had primary responsibility for U.S.

08 Book.indb 14

2/13/09 3:58:10 PM

The Contributors

xv

debt management. While at the Treasury, he took the lead on several critical initiatives, including a broad restructuring of Treasury debt management practices and the reopening of financial markets following 9/11. He also played a key role as the Treasury Board representative to the Pension Benefit Guaranty Corporation. Mr. Fisher worked for the Federal Reserve Bank of New York for 15 years, where he most recently was responsible for the Federal Reserve’s open market and foreign exchange operations. He also worked for the Bank for International Settlements in Basel, Switzerland. Gary B. Gorton, Professor, School of Management, Yale University Mr. Gorton is professor of finance at the Yale School of Management. He is a research associate of the National Bureau of Economic Research. Mr. Gorton was the Robert Morris Professor of Banking and Finance at the Wharton School of the University of Pennsylvania, as well as a professor of economics in the College of Arts and Sciences. He has served as a member of Moody’s Investors Services Academic Advisory Panel, a director of the research program on banks and the economy for the Federal Deposit Insurance Corporation, and a senior economist with the Federal Reserve Bank of Philadelphia. In 1994, he was the Houblon-Norman Fellow at the Bank of England. Mr. Gorton’s research expertise includes the role of stock markets and banks, arbitrage pricing, loan sales, and bank regulation. He has published in numerous journals, is currently an editor of the Review of Economic Studies, and is on the editorial board of many other journals. He is a member of the American Finance Association, the American Economic Association, and the Econometric Society. He also works as a consultant and adviser to AIG Financial Products. Bengt Holmström, Professor, Massachusetts Institute of Technology Mr. Holmström is the Paul A. Samuelson Professor of Economics at the Massachusetts Institute of Technology (MIT), where he also was head of the Economics Department. He has a joint appointment with MIT’s Sloan School of Management. He is a fellow of the American Academy of Arts and Sciences, the Econometric Society,

08 Book.indb 15

2/13/09 3:58:10 PM

xvi

The Contributors

and the European Corporate Governance Institute. Additionally, he is a research associate of the National Bureau of Economic Research and a member of the executive committee for the Center for Economic Policy Research. Mr. Holmström is also an elected foreign member of several organizations, including the Royal Swedish Academy of Sciences and the Finnish Academy of Sciences and Letters. He has served as an associate professor at the Kellogg Graduate School of Management at Northwestern University and a professor at Yale University’s School of Management. He has held numerous other memberships, as well as editorial, professional, and board positions. Anil K. Kashyap, Professor, Graduate School of Business, University of Chicago Mr. Kashyap is the Edward Eagle Brown Professor of Economics and Finance at the University of Chicago’s School of Business. He also serves as one of the faculty co-directors of the Initiative on Global Markets. His research interests include the Japanese financial system, monetary policy, banking, and the sources of business cycles. His book, “Corporate Financing and Governance in Japan: The Road to the Future,” (with Takeo Hoshi) was selected for the 45th Nikkei Prize for Excellent Books in Economic Science. He was the senior Houblon-Norman fellow at the Bank of England in 2001. Mr. Kashyap serves on the editorial boards of many academic journals. He is a member of the University of Chicago’s Center for East Asian Studies, a consultant for the Federal Reserve Bank of Chicago’s Research Department, and a research associate of the National Bureau of Economic Research, in addition to working with many other national and international organizations. Previously, he was an economist for the Board of Governors of the Federal Reserve System. John Lipsky, First Deputy Managing Director, International Monetary Fund Mr. Lipsky has been first deputy managing director of the International Monetary Fund (IMF) since 2006. Previously, he was vice chairman of the JPMorgan Investment Bank, where, among other work, he published independent research on the principal forces

08 Book.indb 16

2/13/09 3:58:10 PM

The Contributors

xvii

shaping global financial markets. He has served as chief economist of JPMorgan, as well as chief economist and director of research at Chase Manhattan Bank. He was chief economist of Salomon Brothers Inc. In the years prior, he was based in London and directed Salomon Brothers’ European Economic and Market Analysis Group. Mr. Lipsky serves on the board of directors of the National Bureau of Economic Research. Prior to his current position with the IMF, he served as a director of several corporations and non-profit organizations. Raghuram G. Rajan, Professor, Graduate School of Business, University of Chicago Mr. Rajan is the Eric J. Gleacher Distinguished Service Professor of Finance at the Graduate School of Business at the University of Chicago. He served as the chief economist at the International Monetary Fund from 2003 to 2006. He was a visiting professor at the Kellogg School of Management at Northwestern University, the Massachusetts Institute of Technology Economics Department, the Sloan School of Management, and the Stockholm School of Economics. He also has worked as a consultant for the Indian Finance Ministry, World Bank, Federal Reserve Board of Governors, Swedish Parliamentary Commission, and various financial institutions. Mr. Rajan’s work has earned him a number of awards. He received the inaugural Fischer Black Prize in 2003, which is awarded by the American Finance Association for the person under 40 who has contributed the most to the theory and practice of finance. Three times he has won the Brattle Prize for a distinguished paper in the Journal of Finance. Jean-Charles Rochet, Professor, University of Toulouse Mr. Rochet is a professor of mathematics and economics at the University of Toulouse. He also serves as the research director at the Institute of Industrial Economics (IDEI). He has been a fellow of the Econometric Society since 1995 and of the European Economic Association since 2004, where he also served as a council member. Previously, he taught at ENSAE and Polytechnique in Paris and was a visiting professor at the London School of Economics from 2001

08 Book.indb 17

2/13/09 3:58:10 PM

xviii

The Contributors

to 2002. He is currently the co-editor of Annals of Finance and has served as associate editor of Econometrica and several other journals. Additionally, he is a member of the advisory board for the Journal of Financial Stability. Mr. Rochet’s research interests include two-sided markets, industrial organization of the banking sector, banking regulation, and dynamic security design. He is the author of more than 50 articles published in international scientific journals and has written six books. Hyun Song Shin, Professor, Princeton University Mr. Shin is the Hughes-Rogers Professor of Economics at Princeton University. From 2000 through 2005, he was a professor of finance at the London School of Economics. He was a university lecturer at Oxford University and a fellow at Nuffield College in Oxford from 1996 through 2000, and he held tutorial and research fellowships in the years prior. In 2002, he was co-director of the Regulation and Financial Stability program for Financial Markets Group, LSE. Mr. Shin has been a fellow for many organizations, including the British Academy in 2005, the Econometric Society in 2004, and the European Economic Association, also in 2004. He serves as associate editor or on the editorial board for many academic journals. Jeremy C. Stein, Professor, Harvard University Mr. Stein is the Moise Y. Safra Professor of Economics at Harvard University. He also serves as the 2008 president of the American Finance Association. From 1990 through 2000, he was on the finance faculty of the Sloan School of Management at the Massachusetts Institute of Technology, most recently as the J.C. Penney Professor of Management. From 1987 through 1990, he was an assistant professor of finance at the Harvard Business School. Mr. Stein’s research areas include behavioral finance and stock market efficiency, financial intermediation, and monetary policy. He is a co-editor of the Journal of Economic Perspectives and a research associate of the National Bureau of Economic Research. He has served on many editorial boards, including the American Economic Review,

08 Book.indb 18

2/13/09 3:58:10 PM

The Contributors

xix

the Quarterly Journal of Economics, the Journal of Finance, and the Journal of Financial Economics, as well as on the board of directors of the American Finance Association. Yutaka Yamaguchi, Former Deputy Governor, Bank of Japan Mr. Yamaguchi is the former deputy governor of the Bank of Japan, a position he assumed in 1998 after 34 years of continuous service. Other positions he held at the Bank of Japan include chief manager of the Policy Planning Department, deputy director of the Information System Services Department, and adviser for the Management and Budget Control Department. In 1989, Mr. Yamaguchi became the Bank of Japan’s chief representative in the Americas in New York. He served as director of the Research and Statistics Department from 1991 to 1992, director of the Policy Planning Department from 1992 to 1996, and executive director from 1996 to 1998. Mr. Yamaguchi currently advises Japanese and international firms on economic and policy developments.

08 Book.indb 19

2/13/09 3:58:10 PM

Maintaining Stability in a Changing Financial System— An Introduction to the Bank’s 2008 Economic Symposium Gordon H. Sellon, Jr. and Brent Bundick

The recent turmoil in global financial markets, associated with the U.S. subprime mortgage crisis, has raised important questions about the stability and integrity of the new international financial structure. These events represent the first real stress test of the world’s financial system since the late 1990s when the Asian financial crisis, Russian debt default, and Long-Term Capital Management (LTCM) shocked the system. Over the past year, credit quality problems in a relatively small part of the U.S. mortgage market have disrupted financial markets around the world and caused large financial losses at banks and other financial institutions. In response, many central banks have provided emergency liquidity to financial markets and institutions, and some have eased monetary policy to limit the spillover of this financial disruption to the broader economy. To better understand the ongoing financial crisis and the implications for monetary and regulatory policy, the Federal Reserve Bank of Kansas City sponsored the symposium, “Maintaining Stability in a Changing Financial System,” at Jackson Hole, Wyo., on Aug. 21-23, 2008. The symposium brought together monetary policy makers, financial market experts, and academic economists to discuss the current financial crisis and its implications for the broader economy. This introduction describes the context in which the crisis developed xxi

08 Book.indb 21

2/13/09 3:58:10 PM

xxii

Gordon H. Sellon, Jr. and Brent Bundick

and highlights three major themes discussed at the symposium: the nature and origins of the crisis, the efficacy and appropriateness of central banks’ response to the crisis, and implications for future financial system regulation. Institutional Context for the Crisis Recent decades have seen the transformation of financial systems around the world. Against a backdrop of falling costs for gathering and processing information, the development of sophisticated financial modeling techniques, and rising competitive pressures, many countries are moving from a traditional bank-based system of financial intermediation to a more market-based system. A central development in this process is the spread of securitization, which allows loans that were once held on bank balance sheets to be repackaged into securities that can be sold to investors around the world. Another key development is the growth of credit risk transfer, which allows lenders to shift default risk to other parties even when they keep loans on their books. These changes have a number of potential benefits. For example, borrowers may have greater access to credit markets at lower cost, and investors may have more investment options and more ability to manage risk exposures. However, the potential impact on the overall stability of the financial system under these new developments is less clear. Some argue that these changes can promote financial stability by transferring risks more widely to those most able and willing to bear them. Indeed, until the onset of the recent subprime mortgage crisis, financial markets experienced a long period of relative calm. This reduced financial market volatility, coupled with low credit risk spreads, lent support to this view. Others have been more skeptical and note that these new financial developments may require a learning period and stress testing before investors and regulators fully appreciate the risks involved with these new products and practices. In addition, some have suggested the financial stability of recent years may reflect the very benign macroeconomic environment in which the new system has developed.

08 Book.indb 22

2/13/09 3:58:10 PM

Introduction

xxiii

Governments have struggled to keep pace with the evolving landscape. Bank regulation, under the international Basel I and II accords, has attempted to adapt capital measures to address both traditional and new risks to the banking system. But many financial developments in recent years have occurred outside of the banking system. These institutions and markets are generally not subject to the same prudential supervision and regulation as the banking system and have not had direct access to the public safety net in the form of deposit insurance and access to central bank liquidity facilities. Instead, to function effectively, these markets and institutions have relied heavily on asset-backed collateralized lending, third-party credit risk insurance, and credit ratings. Two other important trends in recent years have been the globalization of finance and the growth of large, complex financial institutions. The globalization of finance is the product of the ongoing liberalization of domestic financial markets by governments around the world. This process has great benefits to the extent that funds flow where returns are greatest and investors around the world can more effectively diversify their financial assets. Potential externalities also exist, however, if financial difficulties in one country can spread rapidly to other countries. Moreover, there are significant regulatory issues associated with financial institutions operating across many countries. In recent years, significant consolidation both within the banking sector and across different types of financial organizations has resulted in a number of very large financial institutions both in the U.S. and abroad. Several factors have driven this consolidation, including the falling costs of information technology, the removal of regulatory barriers, and competitive pressures. While consolidation may offer greater efficiency, which can benefit households and businesses, serious concerns exist about whether these large, complex organizations can be effectively managed, especially in terms of their exposure to a broad range of risks. Consolidation also raises concerns that some institutions may have become too big or too important to fail because their failure could cause contagion across financial markets and institutions, threatening the stability of the broader financial system. The moral hazard implications of this development could weaken market

08 Book.indb 23

2/13/09 3:58:10 PM

xxiv

Gordon H. Sellon, Jr. and Brent Bundick

discipline and corporate governance and place smaller institutions at a competitive disadvantage. Nature and Origins of the Crisis Although subprime mortgage lending clearly played a key role in the crisis, symposium participants identified a number of fundamental economic factors that contributed to the development and spread of the crisis. These factors include: poor incentive structures in a variety of contracts, information problems, weaknesses in supervision and regulation, design features of subprime securitizations, the use of market-value accounting, and features of the macroeconomic environment that contributed to increased financial leverage and increased risk-taking. In the opening paper, Charles Calomiris and discussant Michael Bordo viewed the current crisis through the historical lens of past crises. According to Calomiris and Bordo, the current crisis has both old and new elements when compared to previous financial crises. Real estate crises are not uncommon historically, and Calomiris noted that the roots of many earlier crises can be traced to a combination of accommodative monetary policy, rapid growth of untested financial instruments, and government subsidization of risk-taking. In the current crisis, the subprime lending boom has elements of all three factors, but Calomiris suggested that they are not really sufficient to explain the crisis. Rather, he emphasized agency or incentive problems in asset management. He argued that the focus on causes of the subprime lending debacle should not be on the individuals and institutions who originated the loans, or on the sponsors of structured assets and credit rating agencies, but rather, on the financial institutions and institutional investors who purchased the securities. According to Calomiris, these institutions knowingly allowed asset managers to underprice the risks of subprime loans and securities backed by these loans. Calomiris cited a number of factors for this mispricing of risk, including: industry compensation practices for asset managers, regulatory policies that led to a relaxation of ratings practices, and changes in bank capital standards that led banks to move the riskier portions of asset securitizations to off-balance-sheet entities.

08 Book.indb 24

2/13/09 3:58:10 PM

Introduction

xxv

Bordo also noted that the current crisis has some similar features to the U.S. banking crises in the pre-Federal Reserve/National Banking era. In particular, the banking panic of 1907 originated in a new type of financial intermediary, trust companies, who operated outside of the existing private safety net provided by the New York Clearing House. The second paper, by Gary Gorton with discussion by Bengt Holmström, focused on asset securitization and problems in the design of mortgage securities backed by subprime loans. Asset securitization has been an important part of financial markets for more than 25 years. Thus, it is important to understand why problems arose in subprime mortgages and assets backed by subprime mortgages when similar problems had not emerged previously in assets backed by conventional home mortgages, auto and credit card loans, and even commercial real estate loans. Gorton argued that the problems in subprime mortgage securitization can be traced to unique features of the securitization process for these loans. Specifically, subprime securitizations were based on adjustable-rate loans that were forced to be refinanced over a short time horizon. As long as housing prices increased, the lower income borrowers who received these loans could use their increased home equity to refinance into more conventional loans. Thus, the viability of these loans depended heavily on rising house prices. Furthermore, the layers of securities that were backed by these loans depended on a dynamic form of credit enhancement that made sense only when subprime borrowers could continue to refinance their loans via rising home prices. In addition, Gorton stressed that the extremely complex securitization structures that were built on the subprime loans resulted in a significant loss of information, so that the ultimate holders of the securities could not determine the credit quality of the underlying subprime loans. According to Gorton, when house prices declined and subprime loans began to default, investors had difficulty accurately valuing their mortgage-backed securities, a contributing factor to the loss of liquidity in short-term money markets. In discussing Gorton’s paper, Holmström focused on two issues: the role that subprime mortgage-backed securities played in the collapse of liquidity in short-term money markets and the

08 Book.indb 25

2/13/09 3:58:10 PM

xxvi

Gordon H. Sellon, Jr. and Brent Bundick

economic forces behind the growth in subprime lending. According to Holmström, a distinguishing feature of markets and institutions that provide liquidity is that they are not very information-intensive. To function effectively, these markets rely on trust and stability of asset values rather than a detailed analysis of underlying credit quality. Consequently, the design flaws in subprime mortgage securities identified by Gorton made them especially unsuited to serve as the basis for short-term asset-backed lending. Furthermore, said Holmström, the nature of liquidity-providing markets and institutions suggests that “marking to market” may not be an appropriate accounting framework for these institutions. As to why the subprime market developed and grew to such great size, Holmström suggested that its growth was driven mainly by the demand for assets resulting from inflows of savings from emerging market economies. The liquidity crisis, associated with the collapse of subprime lending, was examined in more depth in a paper presented by Franklin Allen and Elena Carletti and discussed by Peter Fisher. As noted by many symposium participants, one of the most surprising aspects of the current crisis has been the severe disruption to the liquidity of the financial system. Markets relying on asset-based borrowing have collapsed, the interbank lending market has been disrupted, and prices of many classes of securities have fallen below fundamental values. Allen and Carletti viewed these developments from the perspective of economic models of financial intermediation. They noted that liquidity provision can be inefficient in models with incomplete financial markets. These models can also generate both dysfunctional asset pricing, where asset values are determined by “cash-in-the-market” pricing, rather than by fundamental factors, and by contagion across financial institutions. This market failure also suggests a possible role for central banks to provide liquidity in crisis situations. Their analysis identified a need for a better understanding of how markets and institutions provide liquidity and whether mechanisms can be developed to make liquidity provision more efficient. They also recommended that market values should be supplemented by model-based and historical-cost valuations in a financial crisis when liquidity is scarce and asset prices do not reflect fundamental values.

08 Book.indb 26

2/13/09 3:58:10 PM

Introduction

xxvii

In his discussion, Peter Fisher argued that a clearer definition of liquidity is needed to understand the liquidity problems that developed during the crisis. Instead of viewing liquidity as a stock variable that could be drawn down or disappear in a crisis, he suggested focusing on liquidity as a behavior reflecting the willingness and ability of lenders to lend. Fisher also proposed looking closer at the whole mechanism of asset-based finance, which has come to dominate short-term lending in many markets. In his view, the shift to asset-based finance, where lenders look only at collateral values and not at the underlying sources of repayment, may impart a dangerous procyclicality to the financial system. While much of the symposium discussion focused on factors specific to subprime lending and market liquidity that may have caused or intensified the crisis, the paper by Tobias Adrian and Hyun Song Shin examined whether broader financial trends and the macroeconomic environment contributed to the crisis. Adrian and Shin noted that there was a clear trend for mortgage loans and mortgage-backed securities to be held outside the banking system in market-based financial intermediaries, such as investment banks. They also observed that these institutions exhibit procyclical leverage, which can lead to a pronounced cycle in asset prices. In this sense, the evolution of the financial system from a bank-based to a more market-based system may have increased financial fragility. In addition, Adrian and Shin suggested that monetary policy may have accentuated the cyclicality of asset prices in two ways. First, short-term interest rates determine the cost of leverage; so an accommodative monetary policy may lead to a more rapid growth of financial intermediary balance sheets and rise in asset prices. Second, greater monetary transparency may cause financial markets to underprice risk by reducing the uncertainty of future policy actions. In his discussion of the Adrian and Shin paper, John Lipsky indicated that the broad trends they found in U.S. financial markets can also be seen in other countries. Furthermore, according to International Monetary Fund (IMF) research, the trend away from bank-based finance to more arms-length finance appears to be associated with financial instability being transmitted more readily

08 Book.indb 27

2/13/09 3:58:11 PM

xxviii

Gordon H. Sellon, Jr. and Brent Bundick

across markets. In addition, he noted that economic downturns following episodes of financial stress appear to be more severe the larger the preceding rise in house prices and credit growth and the more firms and households had previously relied on external sources of finance. Central Bank Responses to the Financial Crisis A second major theme of the symposium examined the responses of central banks to the crisis. While many central banks altered lending and liquidity mechanisms as the crisis developed, the Federal Reserve, unlike most other central banks, also eased policy significantly. These policy actions by central banks elicited sharply different views from symposium participants on a number of key issues. First, should central banks ease policy in response to a financial crisis—or, more generally, what is the role of a financial stability mandate for central banks? Second, can—and should—central banks separate monetary policy from liquidity provision? And third, should central banks respond more symmetrically to asset-price or credit bubbles? In his paper, Willem Buiter provided a pointed critique of central bank responses to the financial crisis. Focusing on the Federal Reserve, European Central Bank, and Bank of England, Buiter was highly critical of their behavior and especially of some of the actions taken by the Federal Reserve. In particular, Buiter thought the Federal Reserve’s aggressive easing of policy was inappropriate and inconsistent with maintaining price stability. While he was more favorable to the liquidity responses of the three banks, he suggested that all three were unprepared for the crisis and had failed to understand the changing structure of the financial system and its need for nontraditional liquidity facilities. The discussants of Buiter’s paper, Alan Blinder and Yutaka Yamaguchi, gave the central banks higher marks. Both thought that the Federal Reserve had acted appropriately in easing monetary policy in an environment of great uncertainty to prevent the likelihood of financial distress spilling over to the broader economy. In noting the similarities of the current crisis to the recent Japanese experience, Yamaguchi also suggested it was time to reconsider whether central banks should move away from the conventional

08 Book.indb 28

2/13/09 3:58:11 PM

Introduction

xxix

wisdom that they should only clean up after a financial crisis but not actively resist the buildup of credit bubbles. Three other issues raised in earlier papers and discussions are also relevant to the issues covered in this session. While some participants supported Buiter’s position that a central bank should rely on liquidity measures rather than monetary policy in response to a crisis, Michael Bordo suggested an opposing view. Bordo noted that targeted liquidity measures may have the undesirable feature of putting credit allocation decisions in the hands of the central bank, which could undermine its independence. The paper by Adrian and Shin also questioned whether central bank lending and monetary policy actions could really be separated. In their view, a central bank’s shortterm interest rate target affects the cost of financial system leverage. This suggests that central banks can contribute to credit and assetprice excesses if they maintain a low policy rate for too long. However, this view also suggests that cutting the target rate in a crisis may help in unwinding leverage, thus lessening the severity of a crisis. Finally, in his luncheon remarks, Mario Draghi questioned whether central banks should have a formal financial stability mandate and whether this mandate could come into conflict with its mandate for price stability. Future Financial System Regulation and Supervision The third major theme of the symposium focused on prospective changes to the system of financial system supervision and regulation. In his opening remarks, Federal Reserve Chairman Ben Bernanke emphasized the importance of strengthening the financial system to reduce the frequency and severity of future crises while, at the same time, mitigating the potential moral hazard problems created by government intervention in financial crises. One way of strengthening the financial system, said Bernanke, is to strengthen the financial infrastructure to make it better able to function effectively in periods of stress. Possible measures to accomplish this end include: improving methods for clearing and settling trades for credit default swaps and other over-the-counter (OTC) derivatives; enhancing the resilience of markets for triparty repo and

08 Book.indb 29

2/13/09 3:58:11 PM

xxx

Gordon H. Sellon, Jr. and Brent Bundick

reducing the use of this market for overnight financing of less-liquid forms of collateral; providing the Federal Reserve with explicit oversight authority for systemically important payments systems; and developing procedures for prompt resolution of financial institutions whose failure poses systemic risks to the financial system. A second approach to strengthening the financial system discussed by Bernanke is to develop a systemwide approach to financial supervision and regulation. As compared to more traditional supervision and regulation, which focuses on activities and risks at individual institutions, a systemwide or “macroprudential” approach focuses on common patterns in risk profiles across institutions and sectors and also looks at the interconnections among institutions and markets. This approach also focuses on whether the existing structure of financial supervision and regulation may impart procyclical behavior to credit extension and asset prices over the business cycle and whether a redesign of this structure may reduce this procyclicality. Several symposium participants also noted that “macroprudential” supervision and regulation, which has been advocated by the Bank for International Settlements (BIS) for many years, may be an alternative to using monetary policy to prevent the formation of credit and asset-price bubbles. In his luncheon address, Mario Draghi also focused on the need to revamp financial supervision and regulation to dampen the cyclicality of credit, asset prices, and risk-taking. According to Draghi, progress is needed in three key areas: improving incentives for risk management and control, improving the resiliency of the system to shocks through a stronger financial infrastructure and shock absorbers, and developing measures for dampening the cyclicality of risk-taking. With regard to these objectives, Draghi suggested a number of actions. To improve incentives, he advocated implementing Basel II, strengthened to take account of new risk exposures and with improved liquidity procedures, and methods to enhance transparency and valuation practices. With regard to strengthening the resilience of the system to shocks, he identified infrastructure improvements similar to those discussed by Bernanke as well as better national and cross-border resolution procedures for systemically important

08 Book.indb 30

2/13/09 3:58:11 PM

Introduction

xxxi

institutions. For individual institutions, he emphasized a renewed effort to establish capital and liquidity buffers that would enable institutions to withstand external shocks without impeding efficiency or encouraging regulatory arbitrage. With regard to reducing the cyclicality of risk-taking, Draghi supported efforts to examine how capital requirements might be used to dampen procyclicality. However, he stressed the need for any changes in the structure of capital requirements to be both transparent and consistent across countries. The role of capital requirements and their potential use in reducing the incidence and severity of financial crises was examined in more detail in the symposium paper presented by Anil Kashyap, Raghuram Rajan, and Jeremy Stein and discussed by Jean-Charles Rochet. According to Kashyap, Rajan, and Stein, the appropriate design of capital requirements requires an understanding of the incentives driving financial institution behavior and the implications of this behavior for financial stability. They noted that use of leverage may be optimal from the standpoint of individual institutions in addressing corporate governance issues, but an unwinding of leverage can also serve as a mechanism for propagating problems at individual institutions across the system in times of crisis. They also examined how capital requirements might be designed to deal with these issues and suggested that a form of capital insurance may be superior to both fixed capital requirements and capital requirements that vary over the cycle. Under their proposal, financial institutions would purchase an insurance policy that would automatically provide more capital when the financial system as a whole is under stress. Such an approach would generally be cheaper to financial institutions because they would not have to hold excess capital in good times but would be able to automatically replenish capital in bad times without facing the difficulty of raising new capital from markets in periods of stress. In discussing the Kashyap, Rajan, and Stein paper, Jean-Charles Rochet agreed that existing capital regulations were in need of reform to better incorporate systemic risks and reduce procyclicality in credit extension and risk-taking. While he found the Kashyap, Rajan, and Stein capital insurance plan interesting, Rochet identified some potential problems that might reduce its effectiveness.

08 Book.indb 31

2/13/09 3:58:11 PM

xxxii

Gordon H. Sellon, Jr. and Brent Bundick

According to Rochet, rather than relying on private insurance contracts, it might be preferable to have the government provide the insurance. Rochet also suggested that existing capital requirements under Basel II are flawed because they rely on a value-at-risk (VaR) methodology that assumes a fixed probability of default. Such an approach, according to Rochet, does not force financial institutions to properly take account of the systemic risks that they might impose on the financial system. In addition, Rochet advocated that institutions with access to central bank liquidity facilities be required to satisfy more stringent requirements for capital, liquidity, and risk management. He also recommended that a central clearing platform be used for OTC financial contracts and that such a framework could also be used to improve the credit-rating process by removing the direct linkages between security issuers and the credit rating agencies. Postscript This year’s symposium provided considerable insight into the ongoing subprime mortgage crisis and its effects on financial markets and institutions. At the conclusion of the symposium, participants were hopeful that financial stress would begin to abate so that the process of financial reconstruction could begin. However, in mid-September, financial conditions worsened significantly in the United States and abroad, and there were increasing signs of weaker economic activity around the world. In response, many central banks and governments responded aggressively with additional liquidity measures, monetary policy easing, and other wide-ranging policy actions. The unprecedented scope of these actions underscores the enormity and complexity of the issues discussed at this year’s symposium and suggests that these topics will be discussed at future policy conferences for many years to come.

Author’s Note: Gordon H. Sellon, Jr. is a Senior Vice President and Director of Research at the Federal Reserve Bank of Kansas City. Brent Bundick is an Assistant Economist at the bank.

08 Book.indb 32

2/13/09 3:58:11 PM

Opening Remarks Ben S. Bernanke

In choosing the topic for this year’s symposium—maintaining stability in a changing financial system—the Federal Reserve Bank of Kansas City staff is, once again, right on target. Although we have seen improved functioning in some markets, the financial storm that reached gale force some weeks before our last meeting here in Jackson Hole has not yet subsided, and its effects on the broader economy are becoming apparent in the form of softening economic activity and rising unemployment. Add to this mix a jump in inflation, in part the product of a global commodity boom, and the result has been one of the most challenging economic and policy environments in memory. The Federal Reserve’s response to this crisis has consisted of three key elements. First, we eased monetary policy substantially, particularly after indications of economic weakness proliferated around the turn of the year. In easing rapidly and proactively, we sought to offset, at least in part, the tightening of credit conditions associated with the crisis and thus to mitigate the effects on the broader economy. By cushioning the first-round economic impact of the financial stress, we hoped also to minimize the risks of a so-called adverse feedback loop in which economic weakness exacerbates financial stress, which, in turn, further damages economic prospects.

1

08 Book.indb 1

2/13/09 3:58:11 PM

2

Ben S. Bernanke

In view of the weakening outlook and the downside risks to growth, the Federal Open Market Committee (FOMC) has maintained a relatively low target for the federal funds rate despite an increase in inflationary pressures. This strategy has been conditioned on our expectation that the prices of oil and other commodities would ultimately stabilize, in part as the result of slowing global growth, and that this outcome, together with well-anchored inflation expectations and increased slack in resource utilization, would foster a return to price stability in the medium run. In this regard, the recent decline in commodity prices, as well as the increased stability of the dollar, has been encouraging. If not reversed, these developments, together with a pace of growth that is likely to fall short of potential for a time, should lead inflation to moderate later this year and next year. Nevertheless, the inflation outlook remains highly uncertain, not least because of the difficulty of predicting the future course of commodity prices, and we will continue to monitor inflation and inflation expectations closely. The FOMC is committed to achieving medium-term price stability and will act as necessary to attain that objective. The second element of our response has been to offer liquidity support to the financial markets through a variety of collateralized lending programs. I have discussed these lending facilities and their rationale in some detail on other occasions.1 Briefly, these programs are intended to mitigate what have been, at times, very severe strains in short-term funding markets and, by providing an additional source of financing, to allow banks and other financial institutions to deleverage in a more orderly manner. We have recently extended our special programs for primary dealers beyond the end of the year, based on our assessment that financial conditions remain unusual and exigent. We will continue to review all of our liquidity facilities to determine if they are having their intended effects or require modification. The third element of our strategy encompasses a range of activities and initiatives undertaken in our role as financial regulator and supervisor, some of which I will describe in more detail later in my remarks. Briefly, these activities include cooperating with other regulators to monitor the health of individual financial institutions; working with the private sector to reduce risks in some key markets;

08 Book.indb 2

2/13/09 3:58:11 PM

Opening Remarks

3

developing new regulations, including new rules to govern mortgage and credit card lending; taking an active part in domestic and international efforts to draw out the lessons of the recent experience; and applying those lessons in our supervisory practices. Closely related to this third group of activities is a critical question that we as a country now face: how to strengthen our financial system, including our system of financial regulation and supervision, to reduce the frequency and severity of bouts of financial instability in the future. In this regard, some particularly thorny issues are raised by the existence of financial institutions that may be perceived as “too big to fail” and the moral hazard issues that may arise when governments intervene in a financial crisis. As you know, in March the Federal Reserve acted to prevent the default of the investment bank Bear Stearns. For reasons that I will discuss shortly, those actions were necessary and justified under the circumstances that prevailed at that time. However, those events also have consequences that must be addressed. In particular, if no countervailing actions are taken, what would be perceived as an implicit expansion of the safety net could exacerbate the problem of “too big to fail,” possibly resulting in excessive risk-taking and yet greater systemic risk in the future. Mitigating that problem is one of the design challenges that we face as we consider the future evolution of our system. As both the nation’s central bank and a financial regulator, the Federal Reserve must be well prepared to make constructive contributions to the coming national debate on the future of the financial system and financial regulation. Accordingly, we have set up a number of internal working groups, consisting of governors, Reserve Bank presidents and staff, to study these and related issues. That work is ongoing, and I do not want to prejudge the outcomes. However, in the remainder of my remarks today I will raise, in a preliminary way, what I see as some promising approaches for reducing systemic risk. I will begin by discussing steps that are already under way to strengthen the financial infrastructure in a manner that should increase the resilience of our financial system. I will then turn to a discussion of regulatory and supervisory practice, with particular attention to whether a more comprehensive, system-wide perspective in financial

08 Book.indb 3

2/13/09 3:58:11 PM

4

Ben S. Bernanke

supervision is warranted. For the most part, I will leave for another occasion the issues of broader structural and statutory change, such as those raised by the Treasury’s blueprint for regulatory reform.2 I.

Strengthening the Financial Infrastructure

An effective means of increasing the resilience of the financial system is to strengthen its infrastructure. For my purposes today, I want to construe “financial infrastructure” very broadly, to include not only the “hardware” components of that infrastructure—the physical systems on which market participants rely for the quick and accurate execution, clearing and settlement of transactions—but also the associated “software,” including the statutory, regulatory and contractual frameworks and the business practices that govern the actions and obligations of market participants on both sides of each transaction. Of course, a robust financial infrastructure has many benefits even in normal times, including lower transactions costs and greater market liquidity. In periods of extreme stress, however, the quality of the financial infrastructure may prove critical. For example, it greatly affects the ability of market participants to quickly determine their own positions and exposures, including exposures to key counterparties, and to adjust their positions as necessary. When positions and exposures cannot be determined rapidly­­—as was the case, for example, when program trades overwhelmed the system during the 1987 stock market crash—potential outcomes include highly risk-averse behavior by market participants, sharp declines in market liquidity and high volatility in asset prices. The financial infrastructure also has important effects on how market participants respond to perceived changes in counterparty risk. For example, during a period of heightened stress, participants may be willing to provide liquidity to a market if a strong central counterparty is present, but not otherwise. Considerations of this type were very much in our minds during the Bear Stearns episode in March. The collapse of Bear Stearns was triggered by a run of its creditors and customers, analogous to the run of depositors on a commercial bank. This run was surprising, however, in that Bear Stearns’ borrowings were largely secured—that is, its lenders held collateral to ensure repayment even if the company

08 Book.indb 4

2/13/09 3:58:11 PM

Opening Remarks

5

itself failed. However, the illiquidity of markets in mid-March was so severe that creditors lost confidence that they could recoup their loans by selling the collateral. Many short-term lenders declined to renew their loans, driving Bear to the brink of default. Although not an extraordinarily large company by many metrics, Bear Stearns was deeply involved in a number of critical markets, including (as I have noted) markets for short-term secured funding as well as those for over-the-counter (OTC) derivatives. One of our concerns was that the infrastructures of those markets and the riskand liquidity-management practices of market participants would not be adequate to deal in an orderly way with the collapse of a major counterparty. With financial conditions already quite fragile, the sudden, unanticipated failure of Bear Stearns would have led to a sharp unwinding of positions in those markets that could have severely shaken the confidence of market participants. The company’s failure could also have cast doubt on the financial conditions of some of Bear Stearns’ many counterparties or of companies with similar businesses and funding practices, impairing the ability of those firms to meet their funding needs or to carry out normal transactions. As more firms lost access to funding, the vicious circle of forced selling, increased volatility and higher haircuts and margin calls that was already well advanced at the time would likely have intensified. The broader economy could hardly have remained immune from such severe financial disruptions. Largely because of these concerns, the Federal Reserve took actions that facilitated the purchase of Bear Stearns and the assumption of Bear’s financial obligations by JPMorgan Chase & Co. This experience has led me to believe that one of the best ways to protect the financial system against future systemic shocks, including the possible failure of a major counterparty, is by strengthening the financial infrastructure, including both the “hardware” and the “software” components. The Federal Reserve, in collaboration with the private sector and other regulators, is intensively engaged in such efforts. For example, since September 2005, the Federal Reserve Bank of New York has been leading a joint public-private initiative to improve arrangements for clearing and settling trades in credit

08 Book.indb 5

2/13/09 3:58:11 PM

6

Ben S. Bernanke

default swaps and other OTC derivatives. These efforts include gaining commitments from private-sector participants to automate and standardize the clearing and settlement process, encouraging improved netting and cash settlement arrangements, and supporting the development of a central counterparty for credit default swaps. More generally, although customized derivatives contracts between sophisticated counterparties will continue to be appropriate in many situations, on the margin it appears that a migration of derivatives trading toward more-standardized instruments and the increased use of well-managed central counterparties, either linked to or independent of exchanges, could have a systemic benefit. The Federal Reserve and other authorities also are focusing on enhancing the resilience of the markets for triparty repurchase agreements (repos). In the triparty repo market, primary dealers and other large banks and broker-dealers obtain very large amounts of secured financing from money funds and other short-term, risk-averse investors. We are encouraging firms to improve their management of liquidity risk and to reduce over time their reliance on triparty repos for overnight financing of less-liquid forms of collateral. In the longer term, we need to ensure that there are robust contingency plans for managing, in an orderly manner, the default of a major participant. We should also explore possible means of reducing this market’s dependence on large amounts of intraday credit from the banks that facilitate the settlement of triparty repos. The attainment of these objectives might be facilitated by the introduction of a central counterparty but may also be achievable under the current framework for clearing and settlement. Of course, like other central banks, the Federal Reserve continues to monitor systemically important payment and settlement systems and to compare their performance with international standards for reliability, efficiency and safety. Unlike most other central banks, however, the Federal Reserve does not have general statutory authority to oversee these systems. Instead, we rely on a patchwork of authorities, largely derived from our role as a banking supervisor, as well as on moral suasion, to help ensure that the various payment and settlement systems have the necessary procedures and controls

08 Book.indb 6

2/13/09 3:58:11 PM

Opening Remarks

7

in place to manage the risks they face. As part of any larger reform, the Congress should consider granting the Federal Reserve explicit oversight authority for systemically important payment and settlement systems. Yet another key component of the software of the financial infrastructure is the set of rules and procedures used to resolve claims on a market participant that has defaulted on its obligations. In the overwhelming majority of cases, the bankruptcy laws and contractual agreements serve this function well. However, in the rare circumstances in which the impending or actual failure of an institution imposes substantial systemic risks, the standard procedures for resolving institutions may be inadequate. In the Bear Stearns case, the government’s response was severely complicated by the lack of a clear statutory framework for dealing with such a situation. As I have suggested on other occasions, the Congress may wish to consider whether such a framework should be set up for a defined set of nonbank institutions.3 A possible approach would be to give an agency—the Treasury seems an appropriate choice—the responsibility and the resources, under carefully specified conditions and in consultation with the appropriate supervisors, to intervene in cases in which an impending default by a major nonbank financial institution is judged to carry significant systemic risks. The implementation of such a resolution scheme does raise a number of complex issues, however, and further study will be needed to develop specific, workable proposals. A stronger infrastructure would help to reduce systemic risk. Importantly, as my FOMC colleague Gary Stern has pointed out, it would also mitigate moral hazard and the problem of “too big to fail” by reducing the range of circumstances in which systemic stability concerns might be expected by markets to prompt government intervention.4 A statutory resolution regime for nonbanks, besides reducing uncertainty, would also limit moral hazard by allowing the government to resolve failing firms in a way that is orderly but also wipes out equity holders and haircuts some creditors, analogous to what happens when a commercial bank fails.

08 Book.indb 7

2/13/09 3:58:11 PM

8

II.

Ben S. Bernanke

A Systemwide Approach to Supervisory Oversight

The regulation and supervisory oversight of financial institutions is another critical tool for limiting systemic risk. In general, effective government oversight of individual institutions increases financial resilience and reduces moral hazard by attempting to ensure that all financial firms with access to some sort of federal safety net—including those that creditors may believe are too big to fail—maintain adequate buffers of capital and liquidity and develop comprehensive approaches to risk and liquidity management. Importantly, a welldesigned supervisory regime complements rather than supplants market discipline. Indeed, regulation can serve to strengthen market discipline, for example, by mandating a transparent disclosure regime for financial firms. Going forward, a critical question for regulators and supervisors is what their appropriate “field of vision” should be. Under our current system of safety-and-soundness regulation, supervisors often focus on the financial conditions of individual institutions in isolation. An alternative approach, which has been called system-wide or macroprudential oversight, would broaden the mandate of regulators and supervisors to encompass consideration of potential systemic risks and weaknesses as well. At least informally, financial regulation and supervision in the United States already include some macroprudential elements. As one illustration, many of the supervisory guidances issued by federal bank regulators have been motivated, at least in part, by concerns that a particular industry trend posed risks to the stability of the banking system as a whole, not just to individual institutions. For example, following lengthy comment periods, in 2006, the federal banking supervisors issued formal guidance on underwriting and managing the risks of nontraditional mortgages, such as interest-only and negative amortization mortgages, as well as guidance warning banks against excessive concentrations in commercial real estate lending. These guidances likely would not have been issued if the federal regulators had viewed the issues they addressed as being isolated to a few banks. The regulators were concerned not only about individual banks but also about the systemic risks associated with excessive industrywide

08 Book.indb 8

2/13/09 3:58:11 PM

Opening Remarks

9

concentrations (of commercial real estate or nontraditional mortgages) or an industrywide pattern of certain practices (for example, in underwriting exotic mortgages). Note that, in warning against excessive concentrations or common exposures across the banking system, regulators need not make a judgment about whether a particular asset class is mispriced—although rapid changes in asset prices or risk premiums may increase the level of concern. Rather, their task is to determine the risks imposed on the system as a whole if common exposures significantly increase the correlation of returns across institutions. The development of supervisory guidances is a process that often involves soliciting comments from the industry and the public and, where applicable, developing a consensus among the banking regulators. In that respect, the process is not always as nimble as we might like. For that reason, less-formal processes may sometimes be more effective and timely. As a case in point, the Federal Reserve—in close cooperation with other domestic and foreign regulators—regularly conducts so-called horizontal reviews of large financial institutions, focused on specific issues and practices. Recent reviews have considered topics such as leveraged loans, enterprise-wide risk management and liquidity practices. The lessons learned from these reviews are shared with both the institutions participating in these reviews as well as other institutions for which the information might be beneficial. Like supervisory guidance, these reviews help increase the safety and soundness of individual institutions, but they may also identify common weaknesses and risks that may have implications for broader systemic stability. In my view, making the systemic risk rationale for guidances and reviews more explicit is certainly feasible and would be a useful step toward a more systemic orientation for financial regulation and supervision. A system-wide focus for financial regulation would also increase attention to how the incentives and constraints created by regulations affect behavior, especially risk-taking, through the credit cycle. During a period of economic weakness, for example, a prudential supervisor concerned only with the safety and soundness of a particular institution will tend to push for very conservative lending policies. In contrast, the macroprudential supervisor would recognize that, for

08 Book.indb 9

2/13/09 3:58:11 PM

10

Ben S. Bernanke

the system as a whole, excessively conservative lending policies could prove counterproductive if they contribute to a weaker economic and credit environment. Similarly, risk concentrations that might be acceptable at a single institution in a period of economic expansion could be dangerous if they existed at a large number of institutions simultaneously. I do not have the time today to do justice to the question of the procyclicality of, say, capital regulations and accounting rules. This topic has received a great deal of attention elsewhere and has also engaged the attention of regulators; in particular, the framers of the Basel II capital accord have made significant efforts to measure regulatory capital needs “through the cycle” to mitigate procyclicality. However, as we consider ways to strengthen the system for the future in light of what we have learned over the past year, we should critically examine capital regulations, provisioning policies and other rules applied to financial institutions to determine whether, collectively, they increase the procyclicality of credit extension beyond the point that is best for the system as a whole. A yet more ambitious approach to macroprudential regulation would involve an attempt by regulators to develop a more fully integrated overview of the entire financial system. In principle, such an approach would appear well-justified, as our financial system has become less bank-centered and because activities or risk-taking not permitted to regulated institutions have a way of migrating to other financial firms or markets. Some caution is in order, however, as this more comprehensive approach would be technically demanding and possibly very costly both for the regulators and the firms they supervise. It would likely require at least periodic surveillance and information-gathering from a wide range of nonbank institutions. Increased coordination would be required among the private- and public-sector supervisors of exchanges and other financial markets to keep up-to-date with evolving practices and products and to try to identify those that may pose risks outside the purview of each individual regulator. International regulatory coordination, already quite extensive, would need to be expanded further.

08 Book.indb 10

2/13/09 3:58:11 PM

Opening Remarks

11

One might imagine also conducting formal stress tests, not at the firm level as occurs now, but for a range of firms and markets simultaneously. Doing so might reveal important interactions that are missed by stress tests at the level of the individual firm. For example, such an exercise might suggest that a sharp change in asset prices would not only affect the value of a particular firm’s holdings but also impair liquidity in key markets, with adverse consequences for the ability of the firm to adjust its risk positions or obtain funding. System-wide stress tests might also highlight common exposures and “crowded trades” that would not be visible in tests confined to one firm. Again, however, we should not underestimate the technical and information requirements of conducting such exercises effectively. Financial markets move swiftly, firms’ holdings and exposures change every day, and financial transactions do not respect national boundaries. Thus, the information requirements for conducting truly comprehensive macroprudential surveillance could be daunting indeed. Macroprudential supervision also presents communication issues. For example, the expectations of the public and of financial market participants would have to be managed carefully, as such an approach would never eliminate financial crises entirely. Indeed, an expectation by financial market participants that financial crises will never occur would create its own form of moral hazard and encourage behavior that would make financial crises more, rather than less, likely. With all these caveats, I believe that an increased focus on systemwide risks by regulators and supervisors is inevitable and desirable. However, as we proceed in that direction, we would be wise to maintain a realistic appreciation of the difficulties of comprehensive oversight in a financial system as large, diverse, and globalized as ours. III.

Conclusion

Although we at the Federal Reserve remain focused on addressing the current risks to economic and financial stability, we have also begun thinking about the lessons for the future. I have discussed today two strategies for reducing systemic risk: strengthening the financial infrastructure, broadly construed, and increasing the system-wide focus of financial regulation and supervision. Work on the financial

08 Book.indb 11

2/13/09 3:58:11 PM

12

Ben S. Bernanke

infrastructure is already well under way, and I expect further progress as the public and private sectors cooperate to address common concerns. The adoption of a regulatory and supervisory approach with a heavier macroprudential focus has a strong rationale, but we should be careful about over-promising, as we are still rather far from having the capacity to implement such an approach in a thoroughgoing way. The Federal Reserve will continue to work with the Congress, other regulators and the private sector to explore this and other strategies to increase financial stability. When we last met here in Jackson Hole, the nature of the financial crisis and its implications for the economy were just coming into view. A year later, many challenges remain. I look forward to the insights into this experience that will be provided by the papers at this conference.

08 Book.indb 12

2/13/09 3:58:12 PM

Opening Remarks

13

Endnotes See, for example, Ben S. Bernanke (2008), “Liquidity Provision by the Federal Reserve,” speech delivered (via satellite) at the Federal Reserve Bank of Atlanta Financial Markets Conference, Sea Island, Ga., May 13. 1

See Department of the Treasury (March 2008), Blueprint for a Modernized Financial Regulatory Structure. 2

3 Ben S. Bernanke (2008), “Financial Regulation and Financial Stability,” speech delivered at the Federal Deposit Insurance Corporation’s Forum on Mortgage Lending for Low and Moderate Income Households, Arlington, Va., July 8.

See, for example, Gary H. Stern and Ron J. Feldman (2004), Too Big to Fail: The Hazards of Bank Bailouts (Washington: Brookings Institution Press). 4

08 Book.indb 13

2/13/09 3:58:12 PM

Preface Martin Feldstein

As Ben’s remarks indicate, this year’s discussion is a very natural follow-on to the themes we discussed here last year. In last year’s meeting, we focused on housing, its role in the business cycle, and importantly, its impact on financial institutions. The decline in house prices continues to be central to our economic problems.  What started as a subprime issue is now spreading to house and asset classes more generally.  We are in the midst of a financial crisis caused by the correction of a serious mispricing of all risks and by the collapse of house prices that had expanded during a serious bubble. The financial crisis is getting worse because of the downward spiral of house prices. That decline is being driven by the increasing number of houses with substantial negative equity—that is, with substantial mortgage debt in excess of house values. Negative equity and defaults are such an important problem because mortgages in the United States, unlike most other countries, are generally no-recourse loans. If a homeowner defaults, creditors can take the house, but they cannot take other property or attach income to make up for any unpaid balances. Even in those states of the United States where mortgages are not norecourse loans, creditors generally do not pursue the assets and incomes of individuals who default. 15

08 Book.indb 15

2/13/09 3:58:12 PM

16

Martin Feldstein

We can’t be sure how much further house prices are going to fall.  Experts say that another 15 percent decline is needed to get back to equilibrium level. But there is nothing to stop the decline at that point.  As homeowners with large, negative equity default, the foreclosed houses will contribute to the excess supply that drives prices down further. And the lower prices will then lead to more negative equity and, therefore, more defaults and foreclosures. It is simply not clear—at least not clear to me—what will stop this self-reinforcing process. Declining house prices are key to the financial crisis because mortgage-backed securities and the derivatives associated with them are the primary assets that are weakening financial institutions. Until house prices stabilize, the mortgage-backed securities cannot be valued with any confidence. The uncertain values of mortgage-backed securities, and the associated derivatives, mean that the financial institutions cannot have confidence in the liquidity, or even the solvency of counterparties, or indeed even in the value of their own capital. Without such confidence, credit will not flow, and economic activity is inevitably limited. This shortage of credit is exacerbated by the need of financial institutions to deleverage. And, since raising capital is both difficult and expensive, that deleveraging is happening by lending less. The macroeconomic weakness in the United States now goes beyond this decreased supply of credit. Falling house prices decrease household wealth and therefore consumer spending. Household wealth has fallen by more than $4 trillion since the peak in house prices two years. Falling employment is reducing wage incomes, and higher prices of food and energy are reducing real incomes further. The declining economic activity in other parts of the world will also reduce the growth of U.S. exports. The Federal Reserve has, in my judgment, responded appropriately this year by reducing the federal funds interest rate sharply to 2 percent and creating a variety of new credit facilities. The low short-term interest rate helped by making the dollar more competitive. But otherwise, monetary policy is, in my judgment, generally not having the kind of traction that it did traditionally because of the condition of the housing sector and because of the dysfunctional character of credit markets.

08 Book.indb 16

2/13/09 3:58:12 PM

Preface

17

So the Congress and the administration responded to this situation earlier this year by enacting the $100 billion tax rebate in an attempt to stimulate consumer spending. Those of us who supported that policy generally knew that past experience and economic theory both implied that such one-time fiscal transfers would have little effect. But we hoped this time might be different. Our support for the fiscal rebates was, in the words of Samuel Johnson, “a triumph of hope over experience.” But our hopes turned out to be frustrated. The data are now in, and the rebates did very little to stimulate spending. More than 80 percent of the rebate dollars were saved or used to pay down debt. Very little was added to consumer spending. So that is where we are today: in the middle of a financial crisis with the economy sliding into recession, with monetary policy at maximum easing and fiscal transfers impotent. The agenda of this meeting focuses on what can be done to achieve and to maintain stability. There are two basic questions. First, what should be done to resolve the current crisis—that is, to stop the financial failures and the excess downward spiral of house prices? And, second, what should be done to reduce the risk and the severity of future financial crises? 

08 Book.indb 17

2/13/09 3:58:12 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next Charles W. Calomiris

A sound banker, alas! Is not one who foresees danger and avoids it, but one who, when he is ruined, is ruined in a conventional and orthodox way along with his fellows, so that no one can really blame him.

—John Maynard Keynes, “The Consequences to the Banks of the Collapse in Money Values,” 1931

Introduction and Executive Summary We are currently experiencing a major shock to the financial system, initiated by problems in the subprime mortgage market, which spread to securitization products and credit markets more generally. Banks are being asked to increase the amount of risk that they absorb (by moving off-balance sheet assets onto their balance sheets), but losses that the banks have suffered limit their capacity to absorb those risky assets. The result is a reduction in aggregate risk capacity in the financial system­—a bank credit crunch caused by a scarcity of equity capital in banks—as losses force those who are used to absorbing risk to have to limit those exposures. This essay considers the origins of the subprime turmoil, and the way the financial system has responded to it. There are both old and 19

08 Book.indb 19

2/13/09 3:58:12 PM

20

Charles W. Calomiris

new components in both the origins and the propagation of the subprime shock. With respect to origins, the primary novelty is the central role of agency problems in asset management. In the current debacle, as in previous real estate-related financial shocks, government financial subsidies for bearing risk seem to have been key triggering factors, along with accommodative monetary policy. While government encouragement of risky borrowing and loose money played a major role in the current U.S. housing cycle, investors in subprime-related financial claims must share the blame for making ex ante unwise investments, which seem to be best understood as the result of a conflict of interest between asset managers and their clients. In that sense, sponsors of subprime securitizations and the rating agencies— whose unrealistic assumptions about subprime risk were known to investors prior to the runup in subprime investments—were providing the market with investments that asset managers demanded in spite of the obvious understatements of risk in those investments. With respect to the propagation of the shock, much is familiar—the central role of asymmetric information is apparent in adverse selection premia that have affected credit spreads, and in the quantity rationing of money market instruments—but there is an important novelty, namely the ability of financial institutions to have raised more than $434 billion (as of the end of the third quarter of 2008) in new capital to mitigate the consequences of subprime losses for bank credit supply. The ability and willingness to raise capital is especially interesting in light of the fact that the subprime shock (in comparison to previous financial shocks) is both large in magnitude and uncertain in both magnitude and incidence. In the past, shocks of this kind have not been mitigated by the raising of capital by financial institutions in the wake of losses. This unique response of the financial system reflects the improvements in U.S. financial system diversification that resulted from deregulation, consolidation, and globalization. Another unique element of the response to the shock has been the activist role of the Fed and the Treasury, via discount window operations and other assistance programs that have targeted assistance to

08 Book.indb 20

2/13/09 3:58:12 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

21

particular financial institutions. Although there is room for improving the methods through which some of that assistance was delivered, the use of directly targeted assistance is appropriate, and allows monetary policy to be “surgical” and more flexible (that is, to retain its focus on maintaining price stability, even while responding to a large financial shock). In light of these new and old elements of the origins and propagation of the subprime turmoil, the essay concludes by considering the near-term future of financial and macroeconomic performance, and the implications for monetary policy, regulatory policy, and the future of the structure of the financial services industry. Downside risks associated with the credit crunch increased in the wake of the financial upheaval of September 2008. At this writing, a comprehensive plan to recapitalize the financial system is being considered by Congress. An intervention based on preferred stock injections into banks would be preferable to the Fed-Treasury TARP proposal of government purchases of bank assets. Although credit conditions are a major concern, dire forecasts of the outlook for house prices reflect an exaggerated view of effects of foreclosures on home prices. Inflation and inflation expectations have risen and pose an immediate threat. Monetary policy should maintain a credible commitment to contain inflation, which would also facilitate U.S. financial and nonfinancial firms’ access to capital markets. Regulatory policy changes that should result from the subprime turmoil are numerous, and include reforms of prudential regulation for banks, an end to the longstanding abuse of taxpayer resources by Fannie Mae and Freddie Mac, the reform of the regulatory use of rating agencies’ opinions, and the reform of the regulation of asset managers’ fee structures to improve managers’ incentives. It would also be desirable to restructure government programs to encourage homeownership in a more systemically stable way, in the form of downpayment matching assistance for new homeowners, rather than the myriad policies that subsidize housing by encouraging high mortgage leverage.

08 Book.indb 21

2/13/09 3:58:12 PM

22

Charles W. Calomiris

What long-term structural changes in financial intermediation will result from the subprime turmoil? The conversion of Morgan Stanley and Goldman Sachs from stand-alone investment banks to commercial (depository) banks under Gramm-Leach-Bliley is one important outcome. The perceived advantages of remaining as a stand-alone investment bank—the avoidance of safety net regulation and access to a ready substitute for deposit funding in the form of repos—diminished as the result of the turmoil. Long-term consequences for securitization will likely be mixed. For products with long histories of favorable experiences—like credit cards—securitization is likely to persist and may even thrive from the demise of subprime securitization, which is a competing consumer finance mechanism. In less time-tested areas, particularly those related to real estate, simpler structures, including on-balance sheet funding through covered bonds, will substitute for discredited securitization in the near term, and perhaps for years to come. I.

What’s Old and What’s New About the Origins of the Turmoil?

The financial turmoil that began in the summer of 2007 continues, and likely will continue, through the end of 2008, and perhaps beyond. The turmoil has many dimensions in addition to the obvious statistics of falling asset prices, increased foreclosures, and widening default spreads—the “financial revulsion” (a wonderfully descriptive term that unfortunately has fallen out of use in recent decades) marks the end of a boom in housing prices, the collapse of the young subprime mortgage market, and the demise of a recent wave of complex securitization structures engineered by Wall Street to share risk and conserve on financial intermediaries’ capital (the so-called originateand-distribute model of financial intermediation). It also marks the end of one the longest periods of high profitability, ample equity capital, and abundant credit supply in U.S. banking (1993-2006). For these reasons, the turmoil is much more than a cyclical readjustment in prices, risk appraisal, risk tolerance, or credit supply; it represents an end to important secular trends in asset prices, financial innovation, and financial intermediation, which persisted for more than a decade.

08 Book.indb 22

2/13/09 3:58:12 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

23

From the perspective of a longer-term view of financial shocks, such reversals are not new. The Great Depression saw similar longterm trend reversals. Asset prices that had boomed in the 1920s collapsed in the 1930s. The stock-issues boom and the tendency of retail investors to become stockholders on a large scale (both of which can be regarded as financial structural changes of the 1920s), were brought to an end in the 1930s (for roughly thirty years). And much like the securitized mortgage finance sector today, the high-fliers of the 1920s, the utilities companies, went from a booming sector that thrived on the new funding sources of the 1920s to struggling enterprises and wards of the state. The Great Depression is not the only example of an historical financial crisis that witnessed a long-term reversal in financial structure trends. Indeed, the Depression was quite different from the current turmoil in its origins; there are many better historical parallels to choose from.1 When searching history for precedents and lessons it is important to recognize distinctions among financial crises (exemplified in Table 1). Some entail severe losses (losses from the dot-com collapse were greater than the large losses from the current subprime turmoil); others do not (e.g., the Penn Central crisis, or the panics of the national banking era). In some cases, the incidence of losses across the economy is easy to discern (e.g., in the dot-com collapse); in others (like the current subprime debacle, the Penn Central crisis, or the national banking era panics) losses are not easy to measure or locate within the financial system. Some revolve around bank lending behavior (like today’s problems); others are located mainly in stock and bond markets (e.g., the dot-com collapse). Some are closely related to real estate (the agricultural problems of the 1920s and the 1980s); others are not. What are the typical historical ingredients of crises that are most similar to the current turmoil? What has caused severe credit collapses linked to real estate booms and busts in the past? Accommodative monetary policy has been a key factor in historical credit and asset pricing cycles of all types historically (Bordo and Wheelock 2007a, 2007b, Bordo 2007). This has long been recognized by commentators on financial crises. In reviewing White’s (1996) edited compendium of prominent articles on financial crises, Calomiris (1998) noted an

08 Book.indb 23

2/13/09 3:58:12 PM

24

Charles W. Calomiris

Table 1 Illustrating the Diversity of U.S. Financial Shocks Financial Shock

Banking Problem?

Real Estate Related?

Importance of Asymmetric Information in Relevant Market

Severity of Financial Shock (relative to size of overall economy)

Panic of 1893

Yes

Partly

High

Low

Panic of 1907

Yes

No

High

Low

Agriculture Distress 1920-1930

Yes

Yes

Low

High

Crash of 1929

No

No

Low

High

Banking Distress 1931-1933

Yes

Partly

Occasional, mainly regional

High

Penn Central 1970

No

No

High

Low

Agricultural Distress Early 1980s

Yes

Yes

Low

Moderate

Bank and S&L Distress 1980-1991

Yes

Yes

Varied

High

Crash of 1987

No

No

Low

High

Dot-com Crash of 2001

No

No

Low

High

Subprime Shock

Yes

Yes

High

High

overarching theme of the collection: the most severe financial crises typically arise when rapid growth in untested financial innovations coincided with very loose financial market conditions (that is, an abundance of the supply of credit). In historical and contemporary real estate-related financial crises, a third factor has also been key to causing the most severe losses: the presence of government subsidies encouraging widespread underpricing of risk, which makes the costs of financial collapses particularly large (see Calomiris 1989, 1990, 1992, 2008, Caprio and Klingebiel 1996a, 1996b, Dermirguc-Kunt, Kane, and Laeven 2008). In exploring the roots of the subprime debacle, it is reasonable to begin the search for causes in this familiar territory. Can one conclude that the current turmoil offers simply another illustration of familiar broad themes that are well known to financial historians? Is the current mess just another example of what happens when one mixes loose monetary policy (magnified by the so-called global savings glut of the past several years),2 distortionary policies that subsidize risk-taking (like various government subsidies for leveraging real estate, discussed

08 Book.indb 24

2/13/09 3:58:12 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

25

below), and financial innovations that complicate risk assessment (an innovative, fast-growing market for securitized assets)? Real estate debacles are common historically. A little more than 100 years ago, five of the financial collapses of that era (Argentina 1890, Australia 1893, Italy 1893, the Western United States 1893, and Norway 1900) all displayed similar trend reversals in real estate markets, albeit to different degrees.3 Four of these crises (Australia, Argentina, Italy, and Norway) constitute the most severe banking crises of the 1875-1913 period worldwide, where severity is measured in terms of the negative net worth of failed banks as a proportion of annual GDP.4 All four of these cases have been linked in the economic history literature to government subsidies in real estate finance that gave rise to booms in real estate investment. The most severe ones (Australia and Argentina, both of which resulted in nearly unprecedented resolution costs of roughly 10% of GDP) clearly were cases in which particularly large government subsidies financing land development drove extraordinary booms in land markets that ended badly. The Argentine financial collapse of 1890 was at its core the end of an experiment in the subsidization of real estate risk in the pampas. Argentina’s banks were permitted to originate mortgages (cedulas) that were guaranteed to be paid by the state if the borrower was unable to do so. These mortgages traded at par with Argentine government securities in the London money market. This arrangement was designed to expand credit supply for land (the political brainchild, of course, of the recipients of the subsidy). In the process, it also encouraged extreme risk-taking by lenders (the incentive consequences of guaranteeing mortgage repayment are essentially the same as guaranteeing deposit repayment or GSE liabilities in our modern financial system). The Australian case was a bit different; financial market policies toward the private sector were not the primary means through which the government promoted the land boom that preceded the bust of 1893. The pre-1890 Australian economic expansion was largely an investment boom in which the government played a direct role in investing in land and financing farmers’ investments. Government investments in railroads, telegraphs, irrigation, and farms were financed by government

08 Book.indb 25

2/13/09 3:58:12 PM

26

Charles W. Calomiris

debt floated in the British capital market and by government-owned savings banks and postal savings banks (M. Butlin 1987, N. Butlin 1964, S. Butlin 1961, Davis and Gallman 2001). The smaller losses during the Norwegian and Italian land busts reflected less aggressive, more regionally-focused government policies that promoted land development. In Norway, that was achieved through a government-sponsored lender and an accommodative central bank; in Italy, through liability protection for the Banca di Roma, which famously financed a Roman land boom at the behest of the Pope, who had lobbied for national government insurance of the bank’s liabilities (Canovai 1911). The Norwegian banks’ losses amounted to roughly 3% of GDP, and the Italian banks’ losses (which largely reflected exposures to the Roman land market) were roughly 1% of GDP (Calomiris 2008). The agricultural finance collapse of the 1890s in the Western U.S. (concentrated in Kansas and Nebraska) was a different matter; it had little to do with government policy. Here, mortgage brokers and local bankers mistook the quality and riskiness of the newly settled lands of the so-called “middle border,” and in retrospect, invested far too much in lands that failed to meet those expectations; those overly optimistic initial assessments were brought to light during the droughtstricken years after 1887 (Bogue 1955, Calomiris and Gorton 1991). It is noteworthy that bank failures during the U.S. crisis of 1893 were highly concentrated in the states whose lands had produced surprising losses; the losses of failed banks for the U.S. as a whole were small as a fraction of GDP (less than one-tenth of 1%)—in sharp contrast to the other four cases—reflecting the region-specific nature of that crisis and the absence of an active role of government subsidization of real estate risk, which was present in the other four cases. In the 20th century, boom-and-bust cycles in agricultural land prices, sometimes with dramatic consequences for farm and bank failures, were also apparent, and the most severe of these episodes (the farmland price collapses of the 1920s and the early 1980s)—like the land booms and busts of Australia, Argentina, Italy, and Norway in the 1890s—were traceable to government policies that subsidized real estate financing.

08 Book.indb 26

2/13/09 3:58:12 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

27

Following a typical wartime pattern, agricultural prices were bid up substantially during World War I. Some optimistic, risk-loving farmers in some states in the United States substantially expanded their land under cultivation in response to that short-term change (wrongly inferring a permanent change had occurred), while others did not. Interestingly, not all states empowered optimistic farmers to the same degree. In North Dakota, South Dakota and Nebraska, the losses from overly optimistic agricultural lending that came home to roost in the 1920s were much larger than in adjoining states. Those three states empowered land-value optimists by establishing large land financing subsidies in the form of mandatory deposit insurance systems for state-chartered banks. Optimistic farmland speculators could organize small new banks and attract funds easily in the presence of deposit insurance. All state banks shared (via mutual liability for each other’s deposits) any losses that occurred. The result was that these three states’ state-chartered banks expanded their agricultural lending at a much faster pace than other states and did so through the establishment of new, small (very undiversified) rural banks with very low equity capital (Calomiris 1990, 1992).5 A similar pattern repeated itself at the national level during the agricultural boom of the 1970s. Carey (1994) constructed a theoretical and empirical model of how credit subsidies administered through the Farm Credit System “fed the optimists” during the 1970s. As land prices escalated, non-Farm Credit System lenders withdrew from financing loans collateralized by obviously overbought land, while government lenders did not (and eventually constituted 100% of the marginal loan supply for agricultural loans). Carey’s empirical evidence of the existence of a land bubble in the 1970s is unusually convincing; unlike in residential real estate (where projections of fundamentals relating to permanent income and demographic trends make it difficult to establish the existence of a bubble) by focusing on agricultural land, whose value can be clearly linked to soil productivity and crop price trends (which are observable characteristics), one can measure the extent to which land values deviate from reasonable projections of the net present value of income earned from the land.

08 Book.indb 27

2/13/09 3:58:12 PM

28

Charles W. Calomiris

In summary, real estate-related financial crises with the most disastrous loss consequences have typically been the result of government financial policies that subsidized the taking of real estate risk.6 How relevant are these historical cases—Australia and Argentina in the 1890s, the Dakotas and Nebraska in the 1920s, or the U.S. farm boom and bust of the 1970-1985 period—for understanding the current turmoil? Did government investment and credit subsidies drive the current boom and bust in the same manner as it drove these most severe trend-reversing real estate busts of the past, which resulted in huge macroeconomic declines and enormous taxpayerborne resolution costs? Clearly, U.S. financial policy subsidizes the bearing of risk in financing residential real estate. The U.S. government subsidizes homeownership in several ways, but each of those subsidies is delivered in a way that promotes financial fragility in the real estate market. The primary subsidies are: (1) the deductibility of mortgage interest on one’s home,7 (2) FHA programs to provide credit to buyers (which permit 97% leverage at origination, and permit cash-out refinancing that leaves leverage as high as 95%), (3) government funding subsidies via Federal Home Loan Bank lending (which played a large role in financing IndyMac and Countrywide) and liability protection for Fannie Mae and Freddie Mac (formerly implicit, now explicit) along with political pressures on those institutions to increase their “affordable housing” programs, which increased demand for subprime mortgages by Fannie and Freddie, (4) government initiatives (including the Community Reinvestment Act, or CRA) that have pressured banks to increase the access of low-income and minority individuals to bank credit, and (5) default mitigation protocols, developed during the 1990s and early 2000s, which have required banks that originate loans held by Fannie, Freddie, and FHA to adopt standardized practices for renegotiating delinquent loans to avoid foreclosure. These five categories of initiatives either encourage creditworthy borrowers to increase their mortgage leverage (by establishing benefits of maintaining high leverage) or expand access to borrowing for people who would not otherwise be able to secure or retain mortgage loans. Over the last several decades, the government and private lenders

08 Book.indb 28

2/13/09 3:58:12 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

29

have both expanded the maximum allowable leverage on a home and reduced the minimum creditworthiness of individuals with access to mortgage finance, which has magnified the subsidies from these various credit programs. The most important of these influences in recent years seems to have been the role of congressional politics in encouraging Fannie and Freddie to grow their subprime portfolios. Accounting scandals at Fannie and Freddie in 2003 and 2004 galvanized the GSE reform movement. Critics, including Alan Greenspan, worried increasingly about the systemic risks posed by the growing size and portfolio risks of these institutions and undertook a concerted effort to rein in the housing GSEs, which culminated in proposed legislation by Senate Republicans in 2005 (Calomiris and Wallison 2008). Apparently, this drove the GSEs to redouble their efforts to appeal to congressional Democrats by substantially expanding their exposures to subprime mortgages from 2005 through 2007. As of 2008, Fannie and Freddie had a combined exposure to subprime and Alt-A mortgages of more than $1 trillion. Alternative means for subsidizing homeownership do exist in other countries. In particular, one alternative is a program of government matching of downpayments by new homebuyers. This offers an alternative, risk-reducing means of promoting homeownership (Calomiris 2001). But governments typically prefer promoting homeownership by subsidizing lending. The primary explanation for Congress’ and other governments’ preference for credit subsidies, historically and currently, revolves around the differing electoral politics of onand off-budget subsidies. Downpayment matching by the government is a budgeted transfer payment, while the costs of subsidizing housing via the five categories of credit intervention listed above are hidden (until a financial collapse makes them apparent). The desire of legislators to avoid visible budgeted costs in favor of hidden guarantee costs seems to be a consistent theme of political history. That has an important consequence: The powerful political interests that favor real estate subsidies receive their government largesse in a form that promotes financial instability. Undoubtedly, subsidies for mortgage leverage and government policies that have expanded access to credit were key drivers of the

08 Book.indb 29

2/13/09 3:58:12 PM

30

Charles W. Calomiris

current U.S. turmoil. This is not just a U.S. problem; in Germany, for example, the government-supported Landesbanken are the locus of some of the most severe losses. Clearly, it is desirable to reduce government subsidization of mortgage risk. But loose monetary policy and government encouragement of subprime investments by Fannie Mae, Freddie Mac, and other government interventions to promote affordable housing do not offer a complete explanation of the current mortgage mess in the United States. Subprime loan securitizations were bought by private sector players, not just by Fannie Mae or Freddie Mac. And the purchasers and originators of claims on these mortgages were not just regulated commercial banks (who had to meet CRA or other similar regulatory pressures), but included all classes of institutional investors. I will argue that another influence, namely an investment agency problem, was also important for understanding the timing and severity of the subprime shock. Before making that case, it is useful to review more comprehensively the frameworks used by economists to explain financial crises, and how well or poorly those competing frameworks perform in explaining the facts of the current subprime turmoil. Different Frameworks for Explaining Booms and Busts There are several non-mutually exclusive frameworks in economics that are capable of delivering what are variously termed “credit cycles,” “cycles of mania and panic,” “booms and busts,” and the like. I would characterize the literature as divided into four broad frameworks: (1) variation in fundamentals over time, (2) irrational myopia, (3) government subsidies that distort risk pricing, and (4) agency in asset management.8 I have already described the way government subsidies that distort risk pricing can produce booms and busts. I briefly review the other three frameworks to explain why asset management agency problems, in combination with loose monetary policy and a preexisting set of government policies that encouraged high leverage, played the dominant role in the origins of the current turmoil. The first framework, the “fundamentalist” model, posits that credit cycles reflect exogenous events, which alter rational perceptions of future cash flows and lead to endogenous changes in tolerances for

08 Book.indb 30

2/13/09 3:58:12 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

31

risk, reflected in leverage limits, risk pricing, and asset prices. Recent examples of such models include Von Peter (2008) and Geanakoplos (2008). According to these models, which build on prior theoretical work on credit cycles and business cycles by Bernanke and Gertler (1989, 1990) and others, agents behave rationally and respond to evolving news. Responses to news become especially pronounced in environments of asymmetric information, and can deliver large changes in leverage and asset pricing. One strength of this class of models is that it is capable of explaining why some credit cycles are much more severe than others—the severity of the cycle should depend on the size of the exogenous shock and on the financial condition (state variables such as leverage, liquidity, etc.) of financial intermediaries, firms, and consumers at the time news shocks arrive. This framework implies many testable implications (identifying shocks and measuring differences in responses that vary according to the state variables of the agents). There is a large literature measuring responses over the credit cycle and linking them to identifiable shocks and propagators. Importantly, this literature shows that severe credit events do not happen in every cycle. For example, Calomiris and Gorton (1991) show that the timing of the nationwide banking panics during the period 1870-1913 can be fully explained with a dual threshold criterion: If and only if the quarterly liabilities of failing businesses rose sufficiently (by 50%) while stock prices were falling sufficiently (by more than 9%), a banking panic ensued. Calomiris, Orphanides and Sharp (1997) find that firms’ investment contractions during recessions do depend on their preexisting leverage, but that dependence is complex and reflects the fundamental circumstances of individual firms; the combination of firms’ sales growth fundamentals and leverage is what matters, not just leverage, per se, when considering how severely firms are punished by contracting credit markets. Similarly, Calomiris and Mason (2003a) show that bank depositors varied their withdrawal responses to the shocks buffeting banks during the Great Depression according to the fundamental positions of their respective banks. Calomiris and Mason (2003b) show that regional variation in the extent of the credit crunch during the Great Depression was related to characteristics of the banking systems in different states.

08 Book.indb 31

2/13/09 3:58:13 PM

32

Charles W. Calomiris

Variation over time in the pricing of risk (as described by Bordo 2007, and Bordo and Wheelock 2007a, 2007b) arises in a fundamentals-based model of credit cycles. Asymmetric information problems in financial intermediation cause variation over time in the effective supply of credit available to borrowers, and the pricing of risk will vary with the supply of credit. For example, if reductions in the riskless interest rate are associated with increases in the value of bank equity capital, and if increases in equity capital in turn increase the supply of loanable funds, then credit spreads may fall with riskless interest rates. Indeed, this particular transmission mechanism of monetary policy was a key insight in Bernanke’s (1983) fundamentalist model of financial markets during the Great Depression, which found further microeconomic empirical support in Calomiris and Mason (2003b) and Calomiris and Wilson (2004). A limitation of the fundamentalist approach is that it explains variation in risk pricing, but not under- or over-pricing of risk. Several empirical studies have argued that risk pricing not only varies over time, but becomes excessively favorable during booms, implying a failure of markets to adequately protect against loss and to price underlying risk fully (Dell’Ariccia, Igan, and Laeven 2008, Jimenez, Ongena, Peydro-Alcalde, and Saurina 2007, and Mendoza and Terrones 2008). Indeed, I will argue below that there is strong evidence of the underpricing of risk in subprime lending from 2004 to 2007. How can mispricing of risk be explained? Hyman Minsky (1975) and Charles Kindleberger (1978) advocate a behavioral theory of manias (during booms) and panics (during crashes), which is rooted in the tendency of human nature to overreact. Myopia and herdlike behavior cause endogenous cycles of greed and fear to dominate investment behavior rather than rational long-term calculations of forecasted fundamentals. This theory posits the perpetual under- and over-pricing of risk as the result of human nature’s purported tendency to engage in cycles of euphoric greed, followed by fear and panic. Despite its appeal for explaining risk mispricing, the MinskyKindleberger approach suffers from an important empirical defect: As a theory about human nature, it should have nearly universal application. At least within the context of roughly similarly organized

08 Book.indb 32

2/13/09 3:58:13 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

33

financial markets and economies, boom and bust cycles should be pretty similar in their length and severity. That implication is a problem for the theory; some financial crises, as even the brief review of cases above illustrates, have much more severe consequences than others. This variation, of course, is precisely what fundamentalist models of financial cycles are capable of explaining. If one wants to know why this particular turmoil of 2007-2008 is so much worse than others in the past, the Minsky-Kindleberger view is going to be hard pressed to explain it. Neither the Minsky-Kindleberger view nor the “fundamentalist” model can explain the origins and peculiar severity of the current turmoil. The fundamentalist view cannot explain the private sector’s under-pricing of subprime risk. Furthermore, unlike the Russian/ Long-Term Capital crisis of 1998, or Sept. 11, there was no identifiable exogenous shock driving the current turmoil; the problem came from within the financial system. The Minsky-Kindleberger view, while capable of explaining under-pricing of risk, does not explain the relative severity of shocks like the current one. Furthermore, as discussed at length below, there is evidence that subprime risk underpricing was intentional, not the result of euphoria or ignorance. In my view, the three specific, key influences that worked together to produce the massive ex ante underpricing of risk in the two years prior to mid-2007 were: (1) the global savings glut (a surge in the supply of investable funds resulting from loose monetary policy and other global influences, including the exchange rate/reserve accumulation policy of China), (2) the massive increase in demand for subprime instruments by Fannie Mae and Freddie Mac, and (3) agency problems that led asset managers to purposefully deploy an increasing proportion of funds in bad investments. The three influences fed on each other. Fannie and Freddie bid up the prices of subprime instruments and seemed to offer a reliable source of growing, taxpayer-supported demand in support of subprime mortgage-backed securities’ prices. The global savings glut encouraged excessive risk-taking by providing a vast pool of resources available for investment; this factor, by itself, would tend to encourage excessive risk-taking by non-hedge fund money managers who

08 Book.indb 33

2/13/09 3:58:13 PM

34

Charles W. Calomiris

are compensated on the basis of the volume of risky assets that they manage. Indeed, the fact that LBO financing and other asset classes, not just subprime mortgages, seem to have been overpriced in 2006 and 2007 provides evidence of a general environment of excessive risk-taking. But the agency problem was especially pronounced for subprime investments because of the behavior of the GSEs, as well as the novelty of subprime lending and the particular loss experience on subprime foreclosures in 2001-2003, which created a unique moralhazard opportunity for asset managers to enjoy “plausible deniability” in the pricing of risk. Asset managers invested too much in risky assets because of an incentive conflict. If they had informed their clients of the truth—that the supply of good investments in risky assets has been outstripped by the flood of financial savings, and that consequently, the riskreward tradeoff does not warrant further investment in risky assets— then asset managers would have been required to return money to clients rather than invest in risky assets. Presumably, the money would then have ended up in bank deposit accounts or other investments. Returning the money to investors under these circumstances makes investors better off (given the poor return to bearing risk), but it can make asset managers worse off (if their compensation depends primarily on the size of the funds they manage), since the management fees earned grow in proportion of the amount of funds invested in risky assets.9 Agency in Asset Management: “Plausible Deniability” and the 6% Solution What is the evidence that asset managers who bought or retained securitization claims or other liabilities relating to subprime mortgages willingly over-invested their clients’ money in risky assets that did not adequately compensate investors for risk? Others (e.g., Mason and Rosner 2007a, 2007b, IMF 2008, Ellis 2008) describe in detail the faulty assumptions that underlay the securitization of subprime mortgages and related CDOs. Of course, it is always difficult to establish the ex ante unreasonableness of any assumptions. Nevertheless, some facts known to investors in advance of the subprime

08 Book.indb 34

2/13/09 3:58:13 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

35

collapse are hard to explain without appealing to an asset management agency problem. Ratings agencies and sponsors, who engineered the financing structure of subprime MBS through their chosen assumptions regarding the probability of default (PD) and loss given default (LGD) on portfolio pools (and other assumptions), assumed unrealistically low expected losses on subprime MBS pools prior to the crisis and failed to timely revise them upward, despite the high growth of subprime and changes in the population of originators and borrowers that should have been cause for concern. Indeed, ratings agencies and sponsors maintained highly optimistic assumptions about the market until the middle of 2007, long after clear signs of serious problems had emerged. The expected loss assumptions were unreasonably low, and independent observers drew attention to that fact far in advance of the summer of 2007. The low expected loss assumptions were fundamental to the growth of subprime MBS in the four years leading up to the crisis. A low assumed expected loss is crucial for explaining how subprime mortgages were able to finance themselves more than 80% in the form of AAA debts and more than 95% in the form of A, AA, or AAA debts issued by subprime MBS conduits. The low assumed expected loss had two parts: a low assumption of the probability of default (PD), and a low assumption of the loss given default (LGD), which is also called the “severity” of loss. It is hard to document the pre-2007 PD and LGD assumptions used by ratings agencies or sponsors.10 Data on expected losses for subprime pools, however, do exist (the product of LGD and PD). Assumed expected losses were roughly 4.5% circa 2004 and rose to roughly 6% in 2006. Realized losses on these cohorts are now projected to be several times these numbers.11 Where did the low loss assumptions come from, and how could institutional investors have accepted these as reasonable forward-looking estimates? Subprime was a relatively new product, which grew from humble beginnings in the early 1990s and remained small even as recently as several years ago (Table 2); not until the last three years did subprime originations take off. Given the recent origins of the subprime market, which postdates the last housing cycle downturn

08 Book.indb 35

2/13/09 3:58:13 PM

36

Charles W. Calomiris

Table 2 Mortgage Originations By Product and By Originator (Billions of Dollars) 2007 (6 mo)

2006

2005

2004

2003

2002

2001

FHA/VA

42

80

90

130

220

176

175

Conv/Conf

570

990

1090

1210

2460

1706

1265

Jumbo

242

480

570

510

650

571

445

Subprime

151

600

625

530

310

200

160

AltA

205

400

380

185

85

67

55

HELOC

200

430

365

355

220

165

115

1410

2980

3120

2920

3945

2885

2215

ARMs

460

1340

1490

1464

1034

679

355

Refis

765

1460

1572

1510

2839

1821

1298

TOTAL

Top 10 Originators Countrywide (CA)

245

Wells Fargo (IA)

148

Citi (MO)

116

Chase (NJ)

109

B of A (NC)

96

WaMu (WA)

83

Resid. Cap. (NY)

58

Wachovia (NC)

55

IndyMac (CA)

48

Am Home Mort (NY)

35

TOTAL for Top 10

993

TOTAL for Market

1410

Source: Originations data are from “Current Mortgage Market Conditions,” Housing Data Users Group, September 26, 2007.

08 Book.indb 36

2/13/09 3:58:13 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

37

in the U.S. (1989-1991), how were ratings agencies able to ascertain what the LGD would be on a subprime mortgage pool? A significant proportion of subprime mortgages defaulted in the wake of the 2001 recession, although the volume of outstanding subprime mortgages was small at that time (Chart 1). In fact, only in the last quarter has the default rate on subprime mortgages exceeded its 2002 level. The existence of defaults from 2001-2003 created a default loss record, which provided a basis for low expected loss projections. Subsequent experience was even better; the 2003 cohort of subprime mortgages realized cumulative losses of only 3% prior to July 2007 (Merrill Lynch 2007, p. 9, note 11). There were two major problems with using the 2001-2003 experience as a basis for a forward-looking forecast of future losses from subprime foreclosures. First, and most importantly, the loss experience of 2001-2003 occurred in the wake of a very unusual (almost unique) macroeconomic event, namely a recession (in 2001) during which the housing market continued to boom. Low realized losses reflected the fact that housing prices grew dramatically from 2000 to 2003 (see Chart 13). In a flat or declining housing market—the more reasonable forward-looking assumption for a high-foreclosure, recessionary state of the world—both the probability of default (PD) and the LGD would be much greater (as today’s experience demonstrates). The PD would be greater in a declining housing market because borrowers would be less willing to make payments when they have little equity at stake in their homes.12 The LGD would be greater in a declining housing market because of the effect of home price appreciation on lenders’ losses.13 This error was forecastable. For the most part, the housing cycle and the business cycle coincide very closely. Most of the time in the past (and presumably, in the future) when recession-induced defaults would be occurring on subprime mortgages, house prices would be not be appreciating. Thus, it is reasonable to assume that times of high foreclosure are also times of high LGD. This implies that the loss experience of 2001-2003 (when house prices rose) was not a good indicator either of the probability of foreclosure or of the LGD for subprime mortgage pools on a forward-looking basis. Anyone

08 Book.indb 37

2/13/09 3:58:13 PM

38

Charles W. Calomiris

Chart 1 Foreclosure and Delinquency Rates 12 All Mortgages: Foreclosure Inventory (pre-1998) All Mortgages: Payments Past Due 90 Days (pre-1998) All Mortgages: Foreclosures Started (pre-1998) Subprime Mortgages: Foreclosure Inventory (post-1998) Subprime Mortgages: Payments Past Due 90 Days (post-1998) Subprime Mortgages: Foreclosures Started (post-1998) Prime Mortgages: Foreclosure Inventory (post-1998) Prime Mortgages: Payments Past Due 90 Days (post-1998) Prime Mortgages: Foreclosures Started (post-1998)

10

Percent of Mortgages

8

6

4

2007-Q1

2005-Q1

2003-Q1

2001-Q1

1999-Q1

1997-Q1

1995-Q1

1993-Q1

1991-Q1

1989-Q1

1987-Q1

1985-Q1

1983-Q1

1981-Q1

0

1979-Q1

2

Source: Mortgage Bankers Association, National Delinquency Survey. FHA and VA mortgages, and jumbo mortgages, are included in the pre-1998 aggregate data, but VA and FHA mortgages are not included in the post-1998 samples of prime and subprime mortgages; jumbo mortgages are included in those samples.

estimating future losses sensibly should have arrived at a much higher expected loss number than the 4.5%-6% numbers used during the period 2003-2006. Another reason that the expected losses were unrealistically low relates to the changing composition of loans. Even if 6% had been reasonable as a forward-looking assumption for the performance of the pre-2005 cohorts of subprime borrowers, the growth in subprime originations from 2004 to 2007 was meteoric, and was accompanied by a significant deterioration in borrower quality (Ellis 2008).14 Was it reasonable to assume that these changes would have no effect on the expected loss of the mortgage pool? The average characteristics of borrowers changed dramatically (resulting in substantial increases in the PD, which were clearly visible by 2006 even for the 2005 cohort, as is apparent in Chart 2). Mason and Rosner (2007a, 2007b) raised these and many other criticisms of subprime underwriting standards before August 2007. As early as the summer of 2006, critics pointed to the implausible

08 Book.indb 38

2/13/09 3:58:14 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

39

Chart 2 Default Paths of Different Mortgage Cohorts (60+ day delinquencies, in percent of balance) 30

30

Subprime 2006

2005

25

25

2000

20

20 2004

15

15 2007 2003

10

10

5

5

0

0 0

14

10

20

30

40

50

60

14

Alt-A 2000

12

12 2006 10

10

8

8

6

6

2007 2005

4

4

2004

2

2

2003

0

0 0

10

20

30

40

50

60

Source: IMF Global Financial Stability Report, April 2008, p. 6.

08 Book.indb 39

2/13/09 3:58:14 PM

40

Charles W. Calomiris

loss assumptions of subprime mortgage pools, and the need to stresstest them with a housing downturn. This was not rocket science. Even more remarkably, subprime and Alt-A originations for 2006 and early 2007 continued despite mounting evidence of performance problems in existing portfolios, which were discussed openly by the ratings agencies. Gary Gorton, in his oral comments at the 2008 Kansas City Federal Reserve Bank’s Jackson Hole Conference, described the originations in 2006 and 2007 by Merrill, UBS, and Citibank as “shocking.” As Gorton’s (2008) paper emphasizes, the core assumption on which subprime lending had been based was the permanent appreciation of home prices. By the middle of 2006, that assumption came into question. Gorton (2008) shows that the ABX market had become concerned about subprime performance by the middle of 2006. According to Fitch’s (2006a, p. 21) extremely negative discussion of subprime prospects, the environment became increasingly negative after the first quarter of 2006, as reflected in the fact that “the number of sub-prime downgrades in the period between July and October 2006 was the greatest of any four-month period in Fitch’s history for that sector” (up to that point). Fitch (2006a, p. 21) correctly predicted that “the sensitivity of sub-prime performance to the rate of HPA [home price appreciation] and the large number of borrowers facing scheduled payment increases in 2007 should continue to put negative pressure on the sector. Fitch expects delinquencies to rise by at least an additional 50% from current levels throughout the next year and for the general ratings environment to be negative, as the number of downgrades is expected to outnumber the number of upgrades.” Nevertheless, in the midst of all this negative news, the originations continued at a feverish pace (Table 2), and not until the middle of 2007 did serious problems become reflected in significant changes in modeling assumptions by the ratings agencies.15 Institutional investors managing the portfolios of pensions, mutuals, insurance companies and banks continued to buy subprime-related securitization debt instruments, and banks that sponsored these instruments continued to retain large amounts of the risk associated with the subprime MBS and CDO securitizations they packaged,

08 Book.indb 40

2/13/09 3:58:14 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

41

through purchases of their own subprime-related debts and credit enhancements for subprime conduits. Were the bankers who created these securitizations and retained large exposures for their banks related to them, and other sophisticated institutional investors who bought subprime-related securities, aware of the flawed assumptions regarding PD and LGD that underlay the financial engineering of subprime MBS by ratings agencies? These assumptions were widely publicized as part of the process of selling the securities. Did they object? Apparently not. There is also evidence that bankers who securitized subprime mortgages put the worst of the subprime mortgages into their securitization portfolios (retaining the better subprime mortgages on their balance sheets). Keys, Mukherjee, Seru, and Vig (2008) examine a dataset on securitized subprime mortgage loans and find that lenders purposely placed inferior subprime mortgages into securitization portfolios. Specifically, although the mortgages in the pools appeared to be similar to non-securitized mortgages, based on prima facie credit indicators (such as FICO scores), those that were securitized ultimately had substantially higher default rates. These results suggest that securitization was associated with the purposeful adverse selection of risk. In other words, securitizations purposely created hidden risks for buyers, including the sponsoring institutions that retained much of the risk created by their own securitizations. Why did bankers create these risks for their own and other institutions, and why did other sophisticated institutional investors buy these overpriced securities? One answer is that asset managers were placing someone else’s money at risk and earning huge salaries, bonuses and management fees for being willing to pretend that these were reasonable investments. And furthermore, they may have reasoned that other competing banks and asset managers were behaving similarly and that they would be able to blame the collapse (when it inevitably came) on a surprising shock. The script would be clear, and would give “plausible deniability” to all involved. “Who knew? We all thought that 6% was the right loss assumption! That was what experience suggested and what the rating agencies used.” Plausible deniability may have been a coordinating device for allowing asset

08 Book.indb 41

2/13/09 3:58:14 PM

42

Charles W. Calomiris

managers to participate in the feeding frenzy at little risk of losing customers (precisely because so many participated). Because asset managers could point to market-based data and ratings at the time as confirming the prudence of their actions on a forward-looking basis, they were likely to bear little cost from investor losses. If the understatement of subprime risk was so clear, then why didn’t hedge funds sell these investments short? As Gorton (2008) discusses, individual subprime MBS and CDO debt instruments were not traded widely. The ABX market, which traded in aggregate subprime-related indexes, developed only in January 2006; before that time, it was not possible for informed investors to express opinions about the level of risk in this market by buying or selling the various subprime indexes. This account does not place the primary blame for the mispricing of risk on sponsors or rating agencies. After all, sponsors were only supplying what asset managers of their own institutions or outside buyers were demanding. And the rating agencies were also doing what the investors wanted—going through the mechanical process of engineering conduit debt structures and rating them based on transparently rosy assumptions. I doubt that rating agencies were deceiving sophisticated institutional investors about the risks of the products they were rating; rather they were transparently understating risk and inflating the grading scale of their debt ratings for securitized products so that institutional investors (who are constrained by various regulations to invest in debts rated highly by NRSROs) would be able to invest as they liked without being bound by the constraints of regulation or the best interests of their clients. Many observers wrongly attribute rating agencies’ behavior to the fact that sponsors, rather than investors, paid for the ratings. But that fact seems irrelevant; sponsors and investors alike knew what was going on, and if the investors had not wanted the ratings to be inflated, then the ratings agencies would not have inflated them. Ratings grade inflation was demand-driven. Another fact confirms that conclusion. Collateralized debt obligations (CDOs), which increasingly repackaged subprime mortgages, grew dramatically alongside the subprime mortgage boom. From

08 Book.indb 42

2/13/09 3:58:14 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

43

2000 to 2005, the percentage of non-conforming mortgages that became securitized as MBS increased from 35% to 60%, while the percentage of conforming mortgages securitized rose from 60% to 82%. In 2005, 81% of new CDO pools consisted of MBS, and as of October 2006, 39.5% of existing CDO pools covered by Moody’s consisted of MBS, of which 70% were subprime or second-lien mortgages (Mason and Rosner 2007a, p. 28). CDO issuance roughly doubled in 2006 (Chart 3). Were institutional investors aware that rating agencies were rating CDOs using a different scale from the normal corporate bond ratings? Yes. Moody’s published restrospective data on the probability of default (as of the end of 2005) for Baa CDO tranches and for Baa corporate debts. As of 2005, the Baa CDO offerings had a roughly 20% five-year default probability, compared to a roughly 2% five-year default probability for corporate Baa bonds.16 Despite the rhetoric rating agencies publish claiming to maintain uniformity in their ratings scale, it was common knowledge before and during the subprime boom that investment grade debt issues of subprime MBS and CDO conduits were much riskier than their corporate counterparts. Indeed, this fact had been known about securitization debt issues since the early 1990s and was the topic of a high-profile article published by two New York Fed economists (Cantor and Packer 1994). An anecdote conveyed to me by a rating agency executive supports the view that asset managers, not sponsors and rating agencies, were driving the market’s decision to overpay for risky debts. It is well known that sponsors of CDOs engage in an activity called ratings shopping. Sponsors ask rating agencies to tell them, hypothetically, how much AAA debt they would allow to be issued against a given pool of securities being put into the CDO portfolio. If a rating agency gives too conservative an answer relative to its competitors, the sponsor just uses another rating agency. On one occasion, when one agency was uninvited by a sponsor from providing a rating (because the rating agency did not offer to approve as high a percentage limit for AAA debt as the other agencies), the agency warned a prominent institutional investor not to participate as a buyer, but was rebuffed with the statement: “we have to put our money to work.” Clearly, the institutional investors understood and controlled the rating process.

08 Book.indb 43

2/13/09 3:58:14 PM

44

Charles W. Calomiris

Chart 3 Annual Cash CDO Issuance 500

500

450

450

400

400

350

350

300

300

250

250

200

200

150

150

100

100

50

50

0

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

0

Sources: Mason and Rosner (2007), derived from Lucas, Goodman and Fabozzi (2006).

They were sophisticated and informed buyers, and because they controlled the cash, they determined what constituted acceptable risk measurement by sponsors and rating agencies. To what extent is it plausible to argue that the novelty of securitization products (subprime MBS, CDOs, etc.) made investors and rating agencies unable to gauge risk properly? As I have already noted, data were available prior to the turmoil that showed (1) that assumptions regarding subprime losses were unrealistically low, and (2) that the ratings given to debts issued by securitization conduits exaggerated the quality of those debts. Furthermore, the novelty of a securitization product, in and of itself, was an indicator of a need to adjust estimates of risk upward. Experience suggests that rating agencies frequently underestimate the risks of new products and learn from major credit or fraud events that their risk measures and controls are inadequate. Experience prior to the subprime collapse (in credit card securitization, in delinquent consumer account receivable securitization, and in other areas) has shown that the learning curve related to underestimation of risk can be steep. Decades of experience with steep learning curves in new securitization products indicates yet another reason that properly incentivized institutional investors should

08 Book.indb 44

2/13/09 3:58:15 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

45

have been cautious about the new, fast-growing markets in subprime mortgages and CDOs. Indeed, it is particularly strange to look at the measurement of subprime risk in contrast to the measurement of risk in the much older credit card securitization business. In credit card securitization, market participants paid close attention to the identities of originators, to their performance in the past, to the composition of portfolios, and to how compositions changed over time, and originators were rewarded with greater leverage tolerances for “seasoned” receivables with good track records (Calomiris and Mason 2004a). In contrast, until the middle of 2007, the ratings of subprime portfolios (based largely on the 6% or below expected loss assumption) seem to have been extremely insensitive to changes in borrower quality, product type (which is correlated with unobservable aspects of borrower quality), or the state of the housing market. And there was dramatic new entry into subprime origination in 2004-2006, yet these new entrants offering new, riskier products to new customers seem to have been able to raise funds under more or less the same low loss assumptions as old originators who offered older, lower-risk products.17 The principles learned over twenty years in the credit card securitization business were thrown out the window. Various regulatory policies unwittingly encouraged the “plausible deniability” equilibrium. Regulation contributed in at least four ways. First, insurance companies, pension funds, mutual funds, and banks all face regulations that limit their ability to hold low-rated debts, and the Basel I and II capital requirements for banks also place a great deal of weight on rating agency ratings. By granting enormous regulatory power to rating agencies, the government encouraged rating agencies to compete in relaxing the cost of regulation (through lax standards). Rating agencies that (in absence of regulatory reliance on ratings) saw their job as providing conservative and consistent opinions for investors changed their behavior as the result of the regulatory use of ratings and realized huge profits from the fees that they could earn from underestimating risk (and in the process provided institutional investors with plausible deniability).

08 Book.indb 45

2/13/09 3:58:15 PM

46

Charles W. Calomiris

Second, unbelievably, Congress and the SEC were sending strong signals to the rating agencies in 2005 and 2006 to encourage greater ratings inflation in subprime-related CDOs! In a little known subplot to the ratings-inflation story, the SEC proposed “anti-notching” regulations to implement Congress’ mandate to avoid anti-competitive behavior in the ratings industry (Calomiris 2007a). The proposed prohibitions of notching were directed primarily at the rating of CDOs and reflected lobbying pressure from ratings agencies that catered most to ratings shoppers. Notching arose when CDO sponsors brought a pool of securities to a rating agency to be rated that included debts not previously rated by that rating agency. For example, suppose that ratings shopping in the first generation of subprime securitization had resulted in some MBS securities that were rated by Fitch but not Moody’s (i.e., perhaps Fitch had been willing to bless a higher proportion of AAA debt relative to subprime mortgages than Moody’s). When asked to rate the CDO that contained those debts issued by that subprime MBS conduit, Moody’s would offer either to rate the underlying MBS from scratch, or to notch (adjust by a ratings downgrade) the ratings of those securities that had been given by Fitch. Rating agencies that offered more favorable subprime MBS ratings reportedly lobbied Congress to prohibit notching, complaining that this constituted an anti-competitive practice, and arguing that the dominant players (Moody’s and S&P) should instead accept ratings of other agencies without adjustment when rating CDO pools. This effectively would have further emboldened the most lenient rating agencies to be even more lenient to ratings shoppers, since it effectively would have required the relatively conservative agencies (e.g., Moody’s) to accept the ratings of other agencies in repackaging securities rated by others. Unbelievably, the SEC agreed that notching was anti-competitive and proposed to prohibit notching. In light of the CDO debacle and a flood of criticism from academics (including myself ), the SEC quietly withdrew this proposed anti-notching regulation (at least for the time being). But it still contributed to the subprime rating problem. In the face of the threatened anti-notching rule, the likely response by the relatively conservative rating agencies

08 Book.indb 46

2/13/09 3:58:15 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

47

was to loosen their ratings standards on subprime MBS and CDOs. This policy constituted an attack on any remaining voices of conservatism within the ratings industry that argued for the importance of preserving long-run reputational capital: Trying to swim against the tide of grade inflation would put conservative rating agencies at risk of running afoul of their regulator. Third, changes in prudential bank capital regulation introduced several years ago relating to securitization discouraged banks from retaining junior tranches in securitizations that they originated and gave them an excuse for doing so. This exacerbated agency problems by reducing sponsors’ loss exposures. The regulatory changes relating to securitization raised minimum capital requirements for originators retaining junior stakes in securitizations. Sponsors switched from retaining junior stakes to supporting conduits through external credit enhancement (typically lines of credit of less than one year), which implied much lower capital requirements.18 Sponsors that used to retain large junior positions (which in theory should have helped to align origination incentives) no longer had to worry about losses from following the earlier practice of retaining junior stakes. Indeed, one can imagine sponsors explaining to potential buyers of those junior claims that the desire to sell them was driven not by any change in credit standards or higher prospective losses, but rather by a change in regulatory practice—a change that offered sponsors a plausible explanation for reducing their pool exposures.19 More fundamentally, the prudential regulatory regime lacked any device for ensuring that bank risk would be adequately measured or that capital would be commensurate with risk. As Adrian and Shin (2008) show, both risk and leverage increased during the subprime boom, which provides prima facie evidence of the regulatory failure to measure risk and budget capital accordingly. Interestingly, Calomiris and Wilson (2004) show that in the 1920s this was not the case. During that lending boom, as banks’ risks increased, market discipline forced banks to reduce their leverage in order to limit the riskiness of their deposits. In the presence of deposit insurance and anticipated too-big-to-fail protection, however, debt market discipline is now lacking. If prudential regulation fails to limit risks, banks

08 Book.indb 47

2/13/09 3:58:15 PM

48

Charles W. Calomiris

may fail to maintain adequate capital cushions. The recent failure of banks to maintain adequate capital in the face of rising risk suggests a need for fundamental reform of prudential regulation, which is explored in detail in Section III. Fourth, the regulation of compensation practices in asset management likely played an important role in the willingness of institutional investors to invest their clients’ money so imprudently in subprime mortgage-related securities. Casual empiricism suggests that hedge funds (where bonus compensation helps to align incentives and mitigate agency) have fared relatively well during the turmoil, compared to other institutional investors, and this likely reflects differences in incentives of hedge fund managers, whose incentives are much more closely aligned with their clients. The standard hedge fund fee arrangement balances two considerations: the importance of incentive alignment (which encourages long-term profit sharing by managers) and the risk aversion of asset managers (which encourages limiting the downside risk exposure for managers). The result is that hedge fund managers share the upside of long-term portfolio gains but have limited losses on the downside. Because hedge fund compensation structure is not regulated, and because both investors and managers are typically highly sophisticated people, it is reasonable to expect that the hedge fund financing structure has evolved as an “efficient” financial contract, which may explain the superior performance of hedge funds. The typical hedge fund compensation structure is not permissible for some other, regulated asset managers. Mutual fund managers must share symmetrically in portfolio gains and losses; if they were to keep 20% of the upside, they would have to also absorb 20% of the downside. Since risk-averse fund managers would not be willing to expose themselves to such loss, mutual fund managers typically charge fees as a proportion of assets managed and do not share in profits. This is a direct consequence of the regulation of compensation, and arguably has been a source of great harm to investors, since it encourages asset managers to maximize the size of the funds that they manage, rather than the value of those funds. Managers who gain from the size of their portfolios rather than the profitability of their investments will face

08 Book.indb 48

2/13/09 3:58:15 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

49

strong incentives not to inform investors of deteriorating opportunities in the marketplace and not to return funds to investors when the return relative to risk of their asset class deteriorates. To summarize, the subprime debacle is best understood as the result of a particular confluence of circumstances in which incentive problems combined with unusual historical circumstances. The longstanding problems of asset management agency problems and government distortions in real estate finance got much worse in 2003-2006. The specific historical circumstances that drove this included (1) loose monetary policy, which generated a global savings glut, (2) GSE politics in Congress that drove Fannie Mae and Freddie Mac to expand their purchases of subprime assets, (3) prudential regulatory policies that increasingly encouraged lax risk management, and (4) the historical accident of a very low loss rate during the early history of subprime mortgage foreclosures in 2001-2002. Monetary, regulatory, and GSE policies combined with the historically low loss rates to give incentive-conflicted asset managers, rating agencies, and securitization sponsors a basis of “plausible deniability” on which to base unreasonably low projections of default risk. Government actions must bear a significant share of the blame for this outcome, and not just because regulators failed to prevent bank sponsors from behaving more prudently. GSE purchases of subprime assets, increased regulatory reliance on ratings, regulatory actions that encouraged grade inflation, ineffective bank capital regulations including rules that discouraged sponsors from retaining junior risk exposures, proposed SEC anti-notching rules, and regulatory limits on profit sharing by asset managers all contributed to the “plausible deniability” equilibrium. II.

What’s Old and What’s New about the Propagation of the Turmoil?

What aspects of the reactions of financial markets to the subprime shock have been similar to, or different from, the propagation of financial shocks in the past? As in the case of the origins of the subprime shock, the propagation of the subprime shock in the financial system shares many features with previous responses to financial

08 Book.indb 49

2/13/09 3:58:15 PM

50

Charles W. Calomiris

shocks. The role of uncertainty about the size and incidence of the shock across different financial institutions (“asymmetric information” about losses) has produced a wide variety of familiar market responses, which I review (widening credit spreads, ebbs and flows of optimism and pessimism, quantity rationing in money markets, a contraction in the supply of credit, and lender of last resort interventions by the central bank). Nevertheless, there are three elements to the current turmoil that are quite new, and surprisingly so, when considered together. The first novelty is that the shock is unusually severe, as it combines the worst features of previous historical shocks (namely, on the one hand, a large realization of loss, and on the other hand, large uncertainty about the precise size and location in the financial system of that loss). The second novelty is that financial institutions have been unusually willing to raise capital and successful in doing so, and have thereby mitigated the consequences of the subprime shock. This second feature is even more remarkable when considered in combination with the first. A third novelty has been the aggressive use of coordinated Fed and Treasury assistance to particular financial institutions through the discount window and special programs. This section first reviews aspects of the current turmoil that are qualitatively familiar from the history of financial system responses to similar financial shocks, then discusses the three novel aspects of the adjustment to the shock. With respect to the second novelty, the special role of the evolution of the structure of the banking system in the past two decades is described (through a combination of deregulation, consolidation and globalization), which helps to explain the unprecedented ability and willingness of banks to issue new equity in the wake of losses. What’s Old About the Financial System’s Responses to the Shock Subprime mortgages either served as backing for MBS, or were held on balance sheet. Subprime MBS was sometimes repackaged into CDOs, increasingly so leading up to the 2007 collapse of the subprime market. Subprime MBS and CDO conduits issued debts of

08 Book.indb 50

2/13/09 3:58:15 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

51

various ratings which were sold to institutional investors (AAA debts constituted the vast majority—roughly 80% of subprime MBS pools and an even larger percentage of CDO pools). Sponsors of MBS and CDOs did not sell all the securities issued by their conduits. Banks, in particular, purchased substantial amounts of their own conduits’ AAA debts (which enjoyed favored risk weights as assets from the standpoint of bank capital regulation), and many of those debt purchases ended up being parked in ABCP conduits or SIVs run by the sponsoring bank.20 These conduits financed themselves primarily or largely by asset-backed commercial paper, which was sold to MMMFs and other money market investors (Fitch 2005). Additional exposures to these pools also took the form of so-called “external credit enhancements,” by sponsors and other intermediaries (especially monoline insurance companies), who provided various types of liquidity or credit guarantees to the MBS, CDO, and ABCP conduits. The sequence of events relating to the subprime shock and its spread is described in several papers (IMF 2008, Brunnermeier 2008, Buiter 2008, Greenlaw, Hatzius, Kashyap, and Shin 2008, Herring 2008), and in numerous press accounts, and will not be reviewed in detail here. The important elements of the story are that it became clear very quickly in the late summer and early fall of 2007 that losses were growing rapidly on the large amount of subprime mortgages that had been originated in the previous three years, and that the models that had quantified the risks on those mortgages had grossly underestimated prospective losses. The precise size of the future loss was (and remains) hard to gauge, since the structures of the securities are so complex (Gorton 2008) and these new products have such limited track records, particularly in a declining house price envirnoment. The problem was not just the novelty of the product itself, but the fact that its early years of growth had occurred in a booming housing market; there was no way to predict accurately how defaults would evolve in a soft housing market. Furthermore, underwriting standards had deteriorated, as “no-docs” and “low-docs” subprime mortgages proliferated. That meant that the experience of prior cohorts of subprime borrowers offered little reliable evidence on future defaults even if housing conditions did not soften materially.

08 Book.indb 51

2/13/09 3:58:15 PM

52

Charles W. Calomiris

Not only was the aggregate size of loss related to subprime exposures hard to gauge, the incidence of those losses was also hard to measure. Some subprime MBS had been repackaged into complex CDOs and CDO-squareds. And sponsors of CDO conduits, including some of the largest banks, had placed significant amounts of the debts issued by those CDO conduits into their own ABCP and SIV conduits, which in turn financed themselves with commercial paper and various notes. External credit enhancements for the various conduits issuing all these securities were complex, and exposures of guarantors were not easy to quantify. The precise size of portfolios held by different intermediaries, and the proliferation of external credit enhancements that entailed uncertain loss exposures made loss estimation difficult. Markets in the debt instruments were virtually nonexistent, so there was little hope of marking to market. Estimates of the total loss from subprime and other relatively risky (Alt-A) mortgages within the first several months of the turmoil were in the neighborhood of $100-400 billion, which reflected widely disparate views of the probability of default and the loss given default. These losses remain uncertain. At the moment, reasonable estimates fall at the high end of that range. Additional losses related to other consumer, corporate and commercial real estate lending will, in aggregate, likely reach a similar magnitude. Confusion about the size of loss and its incidence led to a flight to quality, as investors sought liquidity. Thus, in addition to the initial (uncertain) shocks to net worth of financial institutions, liquidity risk became a major factor. As emphasized by Mishkin (1991) and Calomiris and Gorton (1991), in historical financial crises the incidence of shocks was hard to gauge (e.g., 1893 or 1907). Asymmetric information about the true financial positions of borrowers and banks led to a contraction in the willingness of parties to lend to each other, which resulted in a flight to quality. In the 2007-2008 turmoil, rising default risk, market illiquidity and the flight to quality were visible in rising long-term debt default risk spreads, and falling Treasury bond yields, as shown in Charts 4 and 5, which plot the CDS spread, the 10-year Treasury yield, and the spread between the Baa corporate rate and Treasuries. Chart 6 shows that the spread between jumbo mortgage interest rates

08 Book.indb 52

2/13/09 3:58:15 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

53

Chart 4 CDS Swap Spread, 10-Yr

100

100

90

90

80

80

70

70

60

60

Nov

2007

Dec

Jan

Feb

Mar

Apr

2008

May

Jun

Jul

Aug

Sep

Source: Bloomberg.

Chart 5 S&P 500 vs. 10-Year Treasury Yields vs. Spread Between Moody’s Seasoned Baa Corporate Bonds and 10-Year Treasury Yields 1800

6

1600 5 1400 4

1000 3 800 600

Yield (%)

S&P 500 Index

1200

2

400

S&P 500

1

10-Year Treasury

200

9 /10 /2008

8 /13 /2008

6 /18 /2008

7 /16 /2008

5 /21 /2008

4 /23 /2008

3 /26 /2008

2 /27 /2008

1 /30 /2008

1 /2 /2008

11 /7 /2007

12 /5 /2007

9 /12 /2007

10 /10 /2007

8 /15 /2007

7 /18 /2007

6 /20 /2007

5 /23 /2007

4 /25 /2007

3 /28 /2007

2 /28 /2007

1 /3 /2007

1 /31 /2007

Spread of Baa Corporate over 10-Year Treasury 0

0

Sources: Yahoo! Finance (http://finance.yahoo.com); Federal Reserve Statistical Release H.15.

08 Book.indb 53

2/13/09 3:58:16 PM

54

Charles W. Calomiris

and conforming mortgage interest rates widened, and both mortgage rates rose, despite the aggressive Fed rate cuts that drove money market rates lower. The widening jumbo-conforming spread reflects, in part, the relative liquidity of conforming mortgages, and in part, the fact that relatively expensive homes are more dependent on the private (non-GSE) securitization market, which saw a rise in its relative cost of funding. Widening of spreads is also visible between different money market instruments. The flight to quality was apparent in a widening spread between LIBOR and Treasury bill yields (Chart 7), the rising relative cost of longer-term LIBOR (Chart 7), and the rising cost of financial commercial paper relative to nonfinancial (Chart 6). The spread between overnight LIBOR and overnight fed funds (Chart 8) also rose. Both of these are costs of unsecured interbank borrowing for one day. Loans of fed funds, however, typically entail credit from small banks, while LIBOR loans are from large banks. The widening spread between overnight LIBOR and fed funds (which had generally remained within 5 basis point prior to the turmoil)21 reached almost 180 basis points toward the end of 2007 and over 400 basis points in September 2008. Large banks were unwilling to lend during the turmoil, either because they were scrambling for liquidity or because they doubted each other’s credit quality. Interestingly, although there is one primary underlying source of loss affecting the year-long period of July 2007-September 2008 being graphed in the various charts (namely, subprime and other losses on existing loans), the charts display large movements up and down in spreads, reflecting variation in estimated losses, adverse selection costs and market illiquidity as uncertainty about the size and consequences of the losses rose and receded in various waves, clearly visible in CDS spreads in Chart 4. This is a familiar pattern in the history of asymmetric information crises, including the national banking era crises and some of the regional banking crises during the Great Depression, which saw similar ups and downs in the perception of risk, and concerns about concentrations of risk in particular financial institutions, which arose in response to particular news events over time (see Sprague 1910, Wicker 1996, Calomiris and Mason 1997,

08 Book.indb 54

2/13/09 3:58:16 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

55

Chart 6 Commercial Paper Rates, LIBOR, and Mortgage Rates 9

9

8

8

7

7

6

Percent

6

5

5

4

4

3

3

Jumbo Mortgage Rate

2

Conforming Mortgage Rate

2

30-Day AA Financial Commercial Paper Interest Rate

1

9/9/2008

8/12/2008

7/15/2008

6/17/2008

5/20/2008

4/22/2008

3/25/2008

2/26/2008

1/1/2008

1/29/2008

11/6/2007

12/4/2007

9/11/2007

10/9/2007

8/14/2007

7/17/2007

6/19/2007

5/22/2007

4/24/2007

3/27/2007

2/27/2007

1/2/2007

0 1/30/2007

0

1

30-Day AA Nonfinancial Commercial Paper Interest Rate

Sources: Federal Reserve (http://www.federalreserve.gov/DataDownload/Choose.aspx?rel=CP); HSH Associates, www.hsh.com.

Chart 7 LIBOR, Treasury Bill, and Fed Funds Rates 7.0 U S L IB O R , 3 - M o n th 6.0

US L IB O R , O ve r n i g h t F e d e r a l f u n d s e f f e c t i ve r a t e

P e r ce n t

5.0

1-Month Treasury

4.0 3.0 2.0 1.0

9/24/2008

9/10/2008

8/27/2008

8/13/2008

7/30/2008

7/16/2008

7/2/2008

6/18/2008

6/4/2008

5/21/2008

5/7/2008

4/9/2008

4/23/2008

3/26/2008

3/12/2008

2/27/2008

2/13/2008

1/30/2008

1/16/2008

1/2/2008

12/5/2007

12/19/2007

11/21/2007

11/7/2007

10/24/2007

10/10/2007

9/26/2007

9/12/2007

8/29/2007

8/15/2007

8/1/2007

0.0

Sources: Federal Reserve Statistical Release H.15; British Bankers Association, Historic BBA LIBOR Rates (http:// www.bba.org.uk/bba/jsp/polopoly.jsp?d=141&a=627).

08 Book.indb 55

2/13/09 3:58:16 PM

56

Charles W. Calomiris

Chart 8 Overnight Libor-Fed Funds Spread 480 430 380 330

Basis Points

280 230 180 130 80 30

9/10/2008

8/27/2008

8/13/2008

7/30/2008

7/2/2008

7/16/2008

6/4/2008

6/18/2008

5/7/2008

5/21/2008

4/9/2008

4/23/2008

3/26/2008

3/12/2008

2/27/2008

2/13/2008

1/30/2008

1/2/2008

1/16/2008

12/5/2007

12/19/2007

11/7/2007

11/21/2007

10/24/2007

9/26/2007

10/10/2007

9/12/2007

8/29/2007

8/1/2007

8/15/2007

-20

Sources: Federal Reserve Statistical Release H.15; British Bankers Association, Historic BBA LIBOR Rates (http:// www.bba.org.uk/bba/jsp/polopoly.jsp?d=141&a=627).

and Bruner and Carr 2007). During historical banking panics, when confusion about the incidence of shocks produced large adverse selection costs in banking, actions by banks, clearing houses, and regulators that resolved uncertainties about the incidence of shocks helped to restore confidence, reduce adverse selection costs, restore liquidity and eventually brought the panics to an end.22 Similarly, during the past year, news that helped reassure market participants that the turmoil was being contained (e.g., Fed intervention to prevent a meltdown of Bear Stearns) produced reductions in spreads. It is difficult to decompose the various contributing factors that affect spreads during an asymmetric-information crisis. Four separate factors are at work: (1) increased expected loss for risky debts, (2) changes in the pricing of any risk of loss reflecting the reduced net worth of asset buyers (i.e., diminishing marginal utility of consumption), (3) changes in the pricing of risk relating to adverse-selection costs (reflecting the difficulty of observing risk), and (4) changes in the pricing of liquidity reflecting an increased desire for liquidity on the part of buyers. Recent research by Schwarz (2008) suggests that

08 Book.indb 56

2/13/09 3:58:17 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

57

during the past year changes in the pricing of liquidity have been more important than credit risk in explaining widening spreads (see also Allen and Carletti’s 2008 view of the central role of systemic liquidity problems in the current turmoil). LIBOR spread widening, in particular, largely has reflected the heightened liquidity demand of borrowers.23 Despite the progress made in disentangling the various influences on spreads, some aspects of the recent experience remain puzzling. Why, for example, did the spreads on Fannie Mae and Freddie Mac debts (over comparable-maturity Treasuries) not fall more as the result of government commitments to protect Fannie’s and Freddie’s debtholders from the risk of default in July 2008, which should have caused Fannie and Freddie debts to be viewed as close substitutes for U.S. Treasuries? An important aspect of financial system adjustment to severe shocks is the tendency for quantity rationing in money market instruments, which is a source of liquidity risk during financial crises. Short-term near money market instruments with a risk of loss—uninsured deposits, commercial paper, and repos—respond to increases in risk primarily through quantity rather than price adjustment. Thus, in addition to rising spreads in bond, CDS and money markets, a major part of the adjustment process to the subprime turmoil was a contraction in money market instruments. LIBOR deposits of maturities greater than a few days virtually disappeared from the banking system in the first months of the turmoil. This is consistent with the theoretical framework of Calomiris and Kahn (1991). Very short-term (demandable) debt becomes more necessary during difficult times owing to its superior ability to discipline bank risk-taking (through the threat of funding withdrawal) in an environment of highly asymmetric information; any bank that would attempt to borrow at longer term under difficult circumstances would both be avoiding discipline of short-term debt (giving rise to a moral-hazard cost) and revealing a desire to avoid that discipline (giving rise to an adverse-selection cost), and would thus pay a higher interest rate. Only banks with risky intentions or unobservably weak banks would try to lock in long-term credit. This explains why

08 Book.indb 57

2/13/09 3:58:17 PM

58

Charles W. Calomiris

longer term, one-month or three-month LIBOR lending was virtually nonexistent in the immediate aftermath of the shock. Asset-backed commercial paper issues, which were strongly connected to CDOs, were withdrawn rapidly from the market, while other commercial paper remained relatively unaffected (only in September and October of 2008 did nonfinancial paper rollover become a potential problem, as the liquidity crisis deepened). As Chart 9 shows, ABCP grew rapidly in 2006 and the first half of 2007, reflecting the close link between ABCP and CDO originations. ABCP fell even faster; most of the decline in outstanding commercial paper occurred in the immediate aftermath of the August-September 2007 shock and reflected mainly the contraction of ABCP; while other financial commercial paper contributed somewhat to the decline, nonfinancial commercial paper has remained virtually unchanged (at least through mid-September 2008). This shows that the initial fallout from the shock has mainly to do with the loss in confidence in the architecture of securitization per se and secondarily with rising adverse-selection costs for financial institutions. It is interesting to note that even within ABCP, it appears that a significant share of ABCP was being rolled over even during the period of sharp ABCP contraction. That is, the decline of ABCP appears to be substantially less than the decline that would have occurred if all maturing ABCP had been withdrawn from the market. Apparently, there was not a categorical refusal to roll over ABCP.24 Some of the apparent “rollover” of ABCP also likely reflects banks purchasing their own paper.25 Bear Stearns’ heavy reliance on overnight repos and high leverage to fund itself led to its collapse in March 2008 as counterparties became concerned about its increasing risk, and as mortgage-backed securities ceased to be acceptable in the market as collateral for overnight repos (a shock that would have been extremely difficult to anticipate even a few months before). Liquidity risk was an important part of that story, since by any reasonable estimate (Bernstein Research 2008a) Bear Stearns was not insolvent. But Bear’s heavy reliance on the risk-intolerant overnight repo market for its funding (Bernstein Research 2008b) meant that it could not continue to rollover its liabilities. Historical evidence from the Panics of 1893 and 1907 confirm

08 Book.indb 58

2/13/09 3:58:17 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

59

Chart 9 Commercial Paper Outstanding (Weekly, Seasonally Adjusted) 2.5

2.5

2.0

2.0

1.5

1.5

Trillion $

Trillion $

Asset-backed Commercial Paper

1.0

1.0

Financial Commercial Paper

0.5

0.5

9/10/2008

7/30/2008

5/7/2008

6/18/2008

3/26/2008

1/2/2008

2/13/2008

11/21/2007

8/29/2007

0.0 10/10/2007

6/6/2007

7/18/2007

4/25/2007

3/14/2007

1/31/2007

11/8/2006

9/27/2006

7/5/2006

8/16/2006

5/24/2006

3/1/2006

4/12/2006

1/18/2006

12/7/2005

9/14/2005

10/26/2005

12/20/2006

Nonfinancial Commercial Paper

0.0

Source: Federal Reserve (http://www.federalreserve.gov/DataDownload/Choose.aspx?rel=CP).

that quantity rationing in money markets can take the form of sudden runs (on deposits and repos) in response to an increase in risk even when the underlying risk of insolvency remains quite low.26 The risk intolerance of money market instruments has been visible historically and in recent times, both in response to idiosyncratic events at particular banks and firms and in response to aggregate shocks. Calomiris, Himmelberg, and Wachtel (1995) analyze the exit of contemporary commercial paper issuers, which occurs reliably and quickly in response to deterioration in earnings and sales growth. Calomiris (2007b) shows that, in response to sudden adverse news affecting a commercial paper issuer, orderly exit from the commercial paper market often occurs even before commercial paper matures; issuers remove their paper from the market, sometimes at a price equal to accrued par (to prevent investors from suffering any loss as the result of the adverse news event) as a means of preserving their reputations with the investor community, in hope of reentering the market subsequently. Uninsured bank deposits, historically and currently, also display patterns of rationing in response to adverse shocks. This

08 Book.indb 59

2/13/09 3:58:17 PM

60

Charles W. Calomiris

can occur as a sudden run on one bank or on many banks (Calomiris and Schweikart 1991, Calomiris and Gorton 1991, Calomiris and Kahn 1991), or as a more gradual response by depositors to reduce certain classes of deposits that are particularly risk-intolerant (Calomiris and Mason 1997, 2003a, Calomiris and Wilson 2004, Calomiris and Powell 2001). A final familiar theme from previous financial disturbances is that financial failures typically reflect fundamental weakness, not random market behavior. Bear Stearns was not insolvent in March 2008, and the same may be said of Lehman Brothers and AIG in September 2008; nevertheless, the unwillingness of creditors to permit Bear to continue in its weak state reflected its unusually large exposure to subprime risk and its unusually high leverage. The market properly singled out the investment bank with the weakest fundamentals. Similarly, Northern Rock was an observably weak institution with large asset-side risk and very high leverage. This non-random pattern of failure is important because it reminds us that financial market discipline is often well-informed, selective, and helpful in containing systemic loss by preventing weak institutions from continuing to operate. Similar patterns of informed, selective, and helpful market discipline have been apparent in historical banking crises, as well. That is not to say that market discipline is perfect; asymmetric information implies that not all financial institutions that lose the confidence of their creditors are as weak as their creditors fear. Furthermore, as the events of September 2008 illustrate, once a liquidity crisis becomes systemic, even institutions with little fundamental risk exposure (like Goldman Sachs and Morgan Stanley) find themselves at risk of being taken down. Still, market discipline has a fair record in identifying doubtful risks even in the midst of severe financial crises (Calomiris and Mason 1997, 2003a, Bruner and Carr 2007). What’s New About the Response to the Shock: Unprecedented Recapitalizations The greatest concern about the subprime turmoil and the collapse of the securitization markets that came with it, from the perspective of potential macroeconomic implications, is the possibility that the

08 Book.indb 60

2/13/09 3:58:17 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

61

failures of financial institutions and the large subprime-related losses within surviving financial institutions would substantially reduce equity capital available to support lending. Although many financial institutions have suffered substantial losses, the primary systemic concern for the macroeconomy is the health and lending capacity of commercial banks, given their central role in providing consumer and business credit. The losses in bank equity were occurring at a time when banks needed capital more than ever to absorb erstwhile securitized assets back onto their balance sheets and support new lending. From the beginning, policy makers worried that the combination of lost capital and reintermediation of securitized assets in the wake of the subprime shock could lead to a huge bank credit-supply contraction, similar perhaps in effect to the credit crunch of the Great Depression (Bernanke 1983, Calomiris and Mason 2003b, Calomiris and Wilson 2004) or the credit crunch of 1989-1991 (Bernanke and Lown 1991, Baer and McElravey 1993, Boyd and Gertler 1994). In the bank capital crunches of the 1930s and 1989-1991, despite the scarcity of bank equity capital, and consequent scarcity of credit, financial institutions suffering from large losses raised virtually no new equity capital (Calomiris and Wilson 2004). Financial economists attribute the lack of new equity offerings by banks in response to large losses to adverse selection problems that result from asymmetric information. Any bank trying to issue equity at a time where potentially large hidden losses remain unidentified will experience a large decline in its stock price, as the market infers that the offering institution may have unusually high losses that it wants to share with new shareholders. That price reaction would make a stock offering highly dilutive, and thus value-destroying, for existing shareholders. During the subprime turmoil, asymmetric information was high, and adverse selection costs were visible in money market spreads and bond spreads, and in money market quantity rationing. Those same information problems should be all the more costly to a bank trying to raise equity capital, since adverse selection problems are much greater for (junior) equity offerings than for (senior) short-term debts (Myers and Majluf 1984).

08 Book.indb 61

2/13/09 3:58:17 PM

62

Charles W. Calomiris

From the standpoint of the ability of banks to raise equity in response to losses, both the size of the shock and the ability to ascertain who will bear its costs are highly relevant. Adverse selection costs of raising equity are higher when shocks are large and uncertain in their incidence. From that perspective, one might have expected little equity to be raised in the wake of the subprime shock. Compared to other financial shocks, this one was both large and highly uncertain in its incidence. In financial history, for the most part, the largest financial shocks affecting banks (measured in units of loss as a percentage of GDP) have generally not been “asymmetric-information” shocks. The losses from the U.S. agricultural bust of 1920-1930, for example, were large, but for the most part, these shocks—which were visible in agricultural commodity price declines and consequent land value declines with clear consequences for local banks—were not shocks in which asymmetric information was very important. The classic asymmetric-information shocks of the national banking era panics of 1873, 1884, 1890, 1893, 1896, and 1907, in contrast, were not associated with large financial system losses, but rather with confusion about the incidence of those losses, which created problems for banks because of the risk intolerance of depositors. In that sense, the current shock is unusually severe in that it is both large (losses on subprime and Alt-A mortgages and related instruments could be as high as 4% of GDP) and markets have been quite uncertain about the incidence of those losses (Greenlaw, Hatzius, Kashyap, and Shin 2008). The large size and uncertain incidence of the subprime shock explains the protracted process of financial system adjustment to the shock. What it does not explain, however, is the remarkable fact that financial institutions have recapitalized themselves with over $434 billion of new capital over the year ending September 2008 (Chart 10). Banks showed an unprecedented capacity to mitigate the consequences of the subprime shock by raising new equity. In September 2008 alone, as Goldman Sachs and Morgan Stanley sought to insulate themselves from the liquidity crisis, and as Merrill, Wachovia, and Washington Mutual were acquired, the financial system raised capital in excess of $40 billion.

08 Book.indb 62

2/13/09 3:58:17 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

63

Chart 10 The Distribution of Total Writedowns ($590.8 billion) and Capital Raising ($434.2 billion) by Institution ($ Billions) Writedown & Loss

Capital Raised

Bank of Montreal U.S. Bancorp Alliance & Leicester Plc Marshall & Ilsley Corp. Bank Hapoalim B.M. Mitsubishi UFJ Financial Group Dexia SA Fifth Third Bancorp Royal Bank of Canada ABN AMRO Holding NV Commerzbank AG UniCredit SpA Landesbank Sachsen AG DZ Bank AG Bank of China Ltd Bear Stearns Companies Inc. Nomura Holdings Inc. Rabobank HSH Nordbank AG E*TRADE Financial Corp. BNP Paribas Dresdner Bank AG WestLB AG Landesbank Baden-Wurttemberg Lloyds TSB Group Plc Goldman Sachs Group Inc. Indymac Bancorp Inc Natixis National City Corp. Mizuho Financial Group Inc. Societe Generale ING Groep N.V. HBOS Plc Bayerische Landesbank Fortis Canadian Imperial Bank of Commerce Barclays Plc Credit Agricole S.A. Wells Fargo & Company Deutsche Bank AG Credit Suisse Group AG Lehman Brothers Holdings Inc. Royal Bank of Scotland Group Plc IKB Deutsche Industriebank AG Morgan Stanley JPMorgan Chase & Co. Bank of America Corp. HSBC Holdings Plc UBS AG Washington Mutual Inc. Merrill Lynch & Co. Wachovia Corporation Citigroup Inc. 0

10

20

30

40

50

60

70

80

Source: Yalman Onaran & Dave Pierson, Banks’ Subprime-Related Losses Surge to $591 Billion: Table, Bloomberg, Sep. 29, 2008.

08 Book.indb 63

2/13/09 3:58:18 PM

64

Charles W. Calomiris

That is not to say that new capital has prevented a credit crunch. The last year has seen a dramatic reduction in some securitization flows. For example, according to Bear Stearns (2007), commercial mortgage-backed securities issues that had averaged $18 billion per month for January through August 2007, fell to only $4 billion in September 2007. As Chart 11 shows, however, commercial and industrial lending expanded rapidly during the August and September upheaval, and continued to grow at a reasonably fast pace throughout the past year, an achievement that stands in sharp contrast the huge contractions in lending that occurred in the 1930s and in 1989-1991. This unprecedented achievement was not a random event, but was rather a predictable consequence of two sets of factors: (1) the favorable condition of banks’ balance sheets at the time the subprime shock hit, and (2) major structural changes in the financial system that made this unprecedented recapitalization occur. Those structural changes were a consequence of the consolidation, deregulation, and globalization of banking and finance that occurred in the past two decades. With these exceptional historical circumstances in mind, some observers foresaw that the unprecedented bank recapitalization would likely occur in response to the capital losses and argued that it could prevent the subprime turmoil from triggering a major recession, a forecast that at least thus far has proved to be accurate. First, with respect to the preexisting condition of U.S. banks at the time of the subprime shock, as Fed Chairman Ben Bernanke noted from the outset, commercial banks were otherwise doing reasonably well and had substantial equity capital. Although the capital position of U.S. banks as of 2007 was inadequate in light of the risks that they had taken, banks were in better shape than they had been in the 1980s. In the late 1980s, bank balance sheets were extremely weak, owing to the series of shocks banks had faced. Banks had suffered losses due to interest rate rises in the early 1980s, LDC loan problems, agriculture land value collapses in the mid-1980s, commercial real estate collapse in the late 1980s, and southwestern oil and real estate distress in the mid-to-late 1980s. Moreover, the overall economic environment was one of anemic macroeconomic performance. Banks were

08 Book.indb 64

2/13/09 3:58:18 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

65

1,700

1,700

1,500

1,500

1,300

1,300

1,100

1,100

All Commercial Banks 900

900

Weekly Reporting Large Domestic Commercial Banks

8/27/2008

6/4/2008

7/16/2008

4/23/2008

3/12/2008

1/30/2008

11/7/2007

12/19/2007

9/26/2007

7/4/2007

8/15/2007

4/11/2007

5/23/2007

2/28/2007

1/17/2007

12/6/2006

10/25/2006

8/2/2006

9/13/2006

6/21/2006

500

5/10/2006

500

3/29/2006

700

1/4/2006

700

2/15/2006

$ Billion

$ Billion

Chart 11 Commercial and Industrial Loans

Note: Data are seasonally adjusted. Source: Federal Reserve Statistical Release H.8 (http://www.federalreserve.gov/releases/h8/data.htm).

not well diversified regionally and had limited sources of income. By the end of the 1980s some money center banks were barely solvent. In contrast, U.S. banks enjoyed profitable and diverse operations and ample equity capital at the time the subprime shock hit. Their wide range of profitable ventures included nontraditional and traditional banking products, within and outside the United States. According to the Federal Reserve Board Statistical Release H8, large, domestically chartered U.S. commercial banks (the primary point of vulnerability in the financial system to the current securitization shock) maintained a seasonally adjusted capital account of $702.5 billion as of September 12, 2007, which was 12.1% of seasonally adjusted assets. Their assets included $1,346.9 billion in securities, most of which were U.S. Treasury and agency securities. These banks had significant capacity for absorbing additional loans and mortgage-backed securities while remaining in compliance with minimum regulatory capital requirements.27 As of December 2006, total equity for the largest 50 U.S. bank holding companies (which is distinct from the data on the chartered banks of those holding companies, cited above) was $819 billion, and tier 1 capital for these holding companies was $570 billion of that amount, while total holding company assets were $9.6 trillion. Thus, the tier 1 leverage ratio, on

08 Book.indb 65

2/13/09 3:58:18 PM

66

Charles W. Calomiris

average, was 6.17% for this group, implying that banks could accommodate substantial new mortgage originations and other lending on balance sheet in an orderly fashion. The diversification of banks’ portfolios, operations, and sources of income—especially those of large, global banks—were also significantly better circa 2007 than in 1989 or 1930. Banks hold much more diversified portfolios today than they used to, they are less exposed to real estate risk than they were in the 1980s, and much less exposed to local real estate risk, although U.S. banks’ exposure to residential real estate has risen since 2000 (Wheelock 2006). In prior episodes of real estate decline (the 1920s, 1930s, and 1980s), much banking distress resulted from exposures to regional shocks because of the absence of nationwide branch banking. In the 1980s, shocks associated with commercial real estate investments in the northeast and oil-related real estate problems in the southwest were particularly significant sources of banking distress.28 During the last two decades, however, banks have become much more diversified regionally, owing to state-level and federal reforms of branching laws, and internationally, as the result of the globalization of banking and finance. Although banks are likely to absorb roughly half of the losses from the subprime fallout according to most estimates, as Chart 10 shows, those bank losses have been distributed globally, not just within the United States. Banks also have a more diverse income stream due to the expansion of bank powers, which culminated in the 1999 Gramm-Leach-Bliley Act. Diversified banks should be able to weather the subprime shock much better than in the 1930s or late 1980s, when variation in regional circumstances led to significant shocks to regionally isolated banks and to the supply of bank credit. That the industrial organization of banking is crucial for facilitating banking systems’ abilities to adjust to shocks without experiencing major disruptions has been a consistent theme of banking history. Bordo (1985) emphasized the peculiar fragility of American banking in the late 19th and early 20th centuries, which reflected the geographical fragmentation of U.S. banks historically, and this theme has been echoed in many other studies.29

08 Book.indb 66

2/13/09 3:58:18 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

67

The superior condition and prospects of banks (relative to the 1980s), owing to their diversification and the highly profitable environment of the last 15 years, reflected the favorable influences of deregulation, consolidation, and globalization, which reshaped the U.S. banking system. Those influences not only helped mitigate the effects of the subprime shock by making the initial condition of banks stronger; they also helped banks raise new capital. The keys to raising capital are convincing the market that the downside of loss can be bounded reasonably, and that favorable future prospects exist (in pursuit of which new capital will be deployed). Banks that are stronger, larger, and more diverse are much better able to bound losses and credibly argue for favorable prospects. Deregulation also helped facilitate the orderly restructuring of large distressed investment banks in 2008. The acquisitions of Bear Stearns and Merrill Lynch by JP Morgan Chase and Bank of America would not have been possible without the repeal of Glass-Steagall. Clearly, the claim that “deregulation” produced the subprime crisis is a false diagnosis. Regulatory failure (especially with respect to the GSEs and prudential banking regulation) was a major contributor to the crisis. But deregulation of branching and bank powers over the past two decades has helped to mitigate the fallout from the crisis in many ways. Several other factors also favored bank recapitalization. First, despite what may seem a slow process of recognizing loss, in comparison with the loss recognition practices of banks and S&Ls in the 1980s, loss recognition has been fast. This reflects a substantially improved regulatory environment in which it is much harder for banks to disguise losses or delay their recognition.30 Second, many hedge funds and sovereign wealth funds were relatively unaffected by the subprime shock and had ample funds to invest. Thus, there were sophisticated investors with adequate resources available to recapitalize banks, if adverse-selection concerns could be overcome. Here again, globalization of finance has helped to cushion the subprime shock considerably. In addition to assisting in recapitalizing banks, nonbank investors (hedge funds and private equity firms) with ample

08 Book.indb 67

2/13/09 3:58:18 PM

68

Charles W. Calomiris

resources to invest are also taking pressure off of bank balance sheets by purchasing assets. What’s New About the Policy Response to the Shock: Unprecedented Activism Another new feature of the response to the current turmoil is the level of activism of the Fed and the Treasury. The number and boldness of their actions have been striking, even prior to the September 2008 campaign to implement the comprehensive TARP plan for massive purchases of financial assets. The terms of lending, and collateral requirements, were quite flexible. Primary dealers and Fannie and Freddie were granted access to the discount window, not just depository banks. A major Wall Street investment bank and the world’s largest insurance company were bailed out by the combined efforts of the Fed and Treasury. And Fannie Mae and Freddie Mac were rescued, as well, and they were subsequently placed in conservatorship, as the initial effort to keep them afloat proved inadequate. Not surprisingly, many people find all this a bit worrying. Government loans, guarantees and investments in troubled financial institutions (which even include potential capital infusions into the GSEs), not to mention government purchases of assets (as contemplated under the TARP plan) not only put taxpayers’ resources at risk today, they also change the risk-taking behavior of financial institutions going forward. If financial institutions know that the government is there to share losses, that makes risk taking a one-sided bet, and so more risk is preferred to less. There is substantial evidence from financial history—some of it very recent—that this “moral-hazard” problem can give rise to hugely loss-making, high-risk investments that are both socially wasteful and an unfair burden on taxpayers.31 Of course, the presence of moral-hazard cost does not mean that all government assistance is ill-advised. If assistance is provided only when the systemic consequences of not providing assistance are truly large, that will limit moral-hazard costs, and if assistance is structured to limit abuse, then assistance can be particularly worthwhile. Were these two conditions met? Were the interventions by the Fed, the Treasury

08 Book.indb 68

2/13/09 3:58:18 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

69

and the Congress justified by the systemic risks of failing to intervene, and did they structure assistance in a cost-minimizing manner? To address these questions, and to place the recent assistance decisions in context, it is useful to review the debate on the role of the lender of last resort as it has evolved in recent years. The debate about the potential gross benefits of assistance has revolved around the question of how important asymmetric information and adverse selection are during episodes of financial shocks. In the 1980s and early 1990s, several prominent economists argued that it might be desirable to abolish the discount window, on the theory that central banks should only manage the aggregate amount of liquidity in the system (via open market operations) and leave it to the financial system to (efficiently) determine the proper allocation of credit (Goodfriend and King 1988, Bordo 1990, Kaufman 1991, 1992, and Schwartz 1992). Proponents of abolishing the discount window recognized that in days of yore it served a purpose, but argued that in the modern era of an efficiently operating fed funds market and other efficient private markets for lending among financial institutions there was no point in Fed lending to banks. Calomiris (1994) challenged that view, and referred to the Fed’s use of the discount window during the Penn Central crisis as an example of how asymmetric information costs can cause erstwhile efficient markets to shut down, giving a role to the Fed in preserving market liquidity through specifically targeted assistance. During the Penn Central episode, which was in some ways similar to the recent turmoil, albeit on a much smaller scale, the market lost confidence in the screening apparatus of the rating agencies for determining access to the commercial paper market. The commercial paper market essentially shut down, and many borrowers faced significantly increased liquidity risk as they were unable to rollover their outstanding commercial paper. By targeting assistance to commercial paper issuers, via pass-through discount window lending channeled through banks, the Fed targeted a temporarily dysfunctional part of the financial system for assistance and prevented commercial paper borrowers from having to cut their investments and engage in a counterproductive scramble for liquidity. As the recent turmoil illustrates, despite

08 Book.indb 69

2/13/09 3:58:18 PM

70

Charles W. Calomiris

the ongoing technological improvements and sophistication of our financial system, asymmetric information problems that disrupt the operation of normally efficient markets remain an important ingredient of market reality. The discount window, therefore, remains an important component of the Fed’s toolkit. How should assistance be structured? Specifically, on what terms (how long a maturity, and at what interest rate?), and against what kind of collateral should loans be made? Should nonbanks be permitted access to the window? Are loans good enough, or are other investments sometimes warranted? An exploration of the full range of possible policy interventions to deal with financial shocks is beyond our scope here; the following is a selective review.32 Bagehot (1873) famously argued that the lender of last resort should lend freely at a penalty rate on good (but not perfect) collateral.33 This prescription still holds validity today, but the devil is in the details. The lender of last resort should lend at a penalty rate to avoid abuse of access to the window. The term of the loan should be long enough to relieve pressure in the market; too short a term forces borrowers to bear imminent rollover risk, which does little to assuage the flight to liquidity. It makes little sense for the lender of last resort to exclude systemically important financial institutions from receiving assistance, although once it is clear that nonbanks are eligible for assistance, they should be subjected to prudential regulations (analogous to those that apply to banks) to limit potential abuse of safety net access. An effective lender of last resort should not be too picky about collateral. Lending against collateral assets that are of higher average quality (lower risk) than the borrower’s overall asset portfolio may do harm rather than good. If a lender of last resort lends against very high-quality collateral, that effectively subordinates depositors of the bank and thereby increases the risk of depositor loss, which could counterproductively prompt deposit withdrawals. Indeed, Mason (2001) shows that this was precisely the problem with the first attempts of the Reconstruction Finance Corporation to provide assistance to banks during the Depression. The 1933 switch to preferred

08 Book.indb 70

2/13/09 3:58:19 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

71

stock investments (which were junior claims relative to deposits) made RFC assistance much more effective. As Meltzer (2003) shows, the Fed has never clearly enunciated a policy rule for its lender of last resort interventions. It prefers instead to make ad hoc interventions and has behaved inconsistently over time. Nevertheless, in theory, it is possible to justify a consistent rule that would contain most if not all of the assistance innovations of the Fed and Treasury—longer term discount window lending to banks and nonbanks on collateral of average quality (including mortgagebacked securities today) and even the proposed use of preferred stock injections into Fannie and Freddie as a substitute for lending. But in granting access to its resources, the lender of last resort still must adhere to two principles: (1) potential adverse systemic consequences with large social costs must be a real possibility (not just a chimerical convenience), and (2) the structure of assistance should minimize moral-hazard costs. Our financial leaders owe us a detailed explanation and justification of the various financial assistance packages that they have orchestrated and a coherent vision and set of rules to guide policy going forward that is consistent with these two principles lest wasteful and risk-increasing rescues become a habit. Neither the Fed nor the Treasury provided such a coherent vision in justifying their decisions regarding whether and how to assist Bear Stearns, Fannie Mae, Freddie Mac, Lehman or AIG. Neither did the Fed or the Treasury explain why the new comprehensive TARP approach was appropriate after September 18, 2008, but not before, or why this asset purchasing approach was superior to other means to stabilizing markets (notably, preferred stock purchases in banks, which have been favored as a superior alternative by most economists). Was intervention necessary and pursued in a least-cost manner in the three most controversial (pre-September 18, 2008) actions by the Fed and the Treasury, namely the assistance given to Bear Stearns, the GSEs, and AIG? The assistance provided to Bear Stearns seems defensible as an action to limit the risk of adverse systemic consequences of Bear Stearns’ failure. Bear was a counterparty to many derivatives transactions and a major repo issuer. A failure of Bear Stearns would have created

08 Book.indb 71

2/13/09 3:58:19 PM

72

Charles W. Calomiris

substantial confusion regarding the net positions of derivatives market participants and would have produced a major shock to the repo market and to money markets more generally. Assistance provided a means of orderly exit (the acquisition of Bear Stearns by JP Morgan Chase) and avoided what could have been substantial disruption in the repo market, derivative markets, and financial markets generally. Was the structure of assistance appropriate? In particular, was the $30 billion loss exposure accepted by the Fed and Treasury really necessary?34 It is not clear (and hard to second-guess in retrospect) whether the Fed and the Treasury could have gotten a better deal in their negotiations with JP Morgan Chase. By all accounts, JP Morgan Chase enjoyed a windfall from the transaction, even after the renegotiation of the Bear Stearns stock price by Bear shareholders, which raised the acquisition price from $2 a share to $10, after the bailout. On the other hand, there were few if any alternative qualified bidders, so the Fed’s (or Treasury’s) ability to bargain was limited. Most importantly, Bear Stearns’ stockholders suffered a huge loss (compared to their pre-acquisition stock price), and thus moral hazard should not be much encouraged by this episode. The promise of assistance to Fannie Mae and Freddie Mac that was given in July 2008 also seems to have been warranted in the sense that their role in the mortgage market was too important to ignore, and their ability to continue accessing the bond market had become questionable. The market wanted to know whether the long-anticipated implicit government backstop would, in fact, be forthcoming. Upon the announcement of the Fed and Treasury plan, the GSEs access to debt markets was initially restored, even before key aspects of the plan for assistance had been approved by Congress. After the July intervention, however, concerns about the GSEs mounted, and ultimately creditors demanded concrete injection of resources by the government, which was undertaken by placing the GSEs into conservatorships in September 2008. The government now has pledged to support the GSEs through preferred stock injections, as needed, to maintain the flow of mortgage credit and to support GSE obligations. These preferred stock injections may be desirable as a short-term measure, but there are several aspects of the proposal that are problematic.

08 Book.indb 72

2/13/09 3:58:19 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

73

First, GSE fragility reflected longstanding incentive problems and excessive risk-taking in anticipation of safety net protection. The GSEs made moral hazard a cornerstone of their business plan for decades. Critics of the GSEs argued that the government’s implicit protection warranted greater regulation, or privatization, or winding down, of GSE operations (Calomiris and Wallison 2008). The GSEs and their defenders responded that there was no implicit protection, and therefore, no need to prevent abuse. In the meantime, they built up subprime mortgage exposures of more than $1 trillion on a paperthin capital base. The short-term assistance program for the GSEs, even if legitimately motivated by systemic concerns, should have been accompanied by a clearly enunciated, long-term proposal to wind down the GSEs, or fully and credibly privatize them (and make them subject to a clearly specified receivership or conservatorship regime). Nationalization of the enterprises would have been another reasonable option. The July assistance legislation and the September creation of the conservatorships does neither, and simply leaves the long-term future of the GSEs open—a surefire method to maximize campaign contributions for influential members of Congress perhaps, but not a very helpful means either of stabilizing markets or providing a transition to proper market discipline. What about the government’s September 2008 decision not to intervene to rescue Lehman Brothers and its opposite decision to rescue AIG? The decision not to rescue Lehman has been criticized as causing much of the late-September 2008 liquidity strains in the market. That decision reflected the view by policy makers that the markets had been given ample time (six months) to adjust to the possibility of a Lehman failure, and that therefore Lehman’s failure would not have grave systemic consequences. In the case of AIG, the larger size, global ramifications, and suddenness of the increased risk of failure on the heels of AIG’s ratings downgrade may explain the government’s different course. Here the government provided assistance, albeit at the price of requiring 80% of the firm’s equity. The government changed course dramatically on September 18, 2008. Up to that point, ad hoc decisions whether and on what terms

08 Book.indb 73

2/13/09 3:58:19 PM

74

Charles W. Calomiris

to intervene had been the means of dealing with problems. On September 18, the Treasury and Fed proposed a comprehensive asset purchase (TARP) plan (alongside new prohibitions on short sales of financial stocks and insurance of money market mutual funds, which were experiencing large withdrawals after one prominent fund “broke the buck” of contributors’ principal in the fund). The best explanation for the change in course revolves around the “bear run” on the stock of Morgan Stanley and Goldman Sachs that occurred on the 17th and 18th of September. The previous policies of the government indicated that the government’s intervention to rescue an ailing firm was uncertain, but that when it did intervene, stockholders suffered large losses. That “punitive intervention” policy made sense from the perspective of limiting moral-hazard consequences of providing assistance, but it had one bad consequence: short sellers could be confident that they would very likely profit from shorting the stock of any financial firm experiencing liquidity trouble; if the insititution did not receive assistance, then short sellers would profit as the firm scrambled to raise cash, and if it did receive assistance, shares would plummet as the result of the policy of punitive intervention. The vulnerability of Morgan Stanley and Goldman Sachs despite the fact that neither of them had significant exposures to subprime problems may have convinced policy makers that the liquidity crisis had reached a new level of severity. III.

What’s Next?

In the first year since the subprime turmoil erupted, economic growth has been sluggish and the employment situation has worsened, but the ability of banks to reintermediate off-balance sheet positions without sharply curtailing credit supply (which was the consequence of banks’ preexisting regulatory capital cushion, their continuing earnings from other sources, as well as substantial capital flotations and dividend cuts) prevented the credit crunch from causing the sort of severe recession that otherwise would have accompanied a financial sector shock of this magnitude. The near-term outlook for the economy and the financial sector has deteriorated recently, as the financial sector was buffeted in

08 Book.indb 74

2/13/09 3:58:19 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

75

September by one of the most dramatic months in its history. Fannie Mae and Freddie Mac went into conservatorship. AIG was rescued by the government, Lehman Brothers failed, Merrill Lynch became part of Bank of America, Washington Mutual and Wachovia were acquired in FDIC-assisted transactions, and Morgan Stanley and Goldman Sachs became bank holding companies. By the end of September, the risk of further significant financial failures within the United States had been substantially reduced, if only by the fact that the fates of virtually all significant financial institutions had already been resolved. But European banks were beginning to experience severe strains and credit spreads were extremely elevated in the U.S. and abroad as equity and debt markets seized up, and the risk of a much more severe credit crunch loomed. At the same time, the inflation picture worsened. Many observers commented that the Fed’s aggressive fed funds rate cuts may have gone too far. There has been a substantial acceleration in inflation and a rise in at least one (controversial) measure of long-term inflation expectations (the Cleveland Fed measure shown in based on the spread between indexed and nominal 10-year Treasuries, shown in Chart 12). Many market participants commented that the failure by the Fed to convince the market that it would ensure price stability has been a significant drag on the stock market. Low U.S. stock prices, especially for banks, are a major cause for concern. Low stock prices discourage banks from raising new equity. Despite the enormous amount of equity raised thus far, unless stock prices rise to encourage banks to continue to raise equity capital, credit supply decline likely will accelerate. The Treasury and Fed have offered the TARP asset purchase plan as a means of staving off the risk of a severe decline in credit and economic activity. The remainder of this section (1) evaluates the TARP proposal, (2) evaluates the risks in the housing market related to the growing wave of foreclosures, (3) offers a few monetary and long-term regulatory policy recommendations, and (3) provides an assessment of how the subprime turmoil will reshape the structure of the financial system.

08 Book.indb 75

2/13/09 3:58:19 PM

76

Charles W. Calomiris

9/9/2008

1.7

8/12/2008

1.7 7/15/2008

1.9

6/17/2008

1.9

5/20/2008

2.1

4/22/2008

2.1

3/25/2008

2.3

2/26/2008

2.3

1/1/2008

2.5

1/29/2008

2.5

11/6/2007

2.7

12/4/2007

2.7

9/11/2007

2.9

10/9/2007

2.9

8/14/2007

3.1

7/17/2007

3.1

6/19/2007

3.3

5/22/2007

3.3

4/24/2007

3.5

3/27/2007

3.5

2/27/2007

3.7

1/2/2007

3.7

1/30/2007

Percent

Chart 12 Cleveland Fed 10-Year TIPS-Derived Expected Inflation

Source: Federal Reserve Bank of Cleveland, TIPS Expected Inflation Estimates (http://www.clevelandfed.org/research/ data/tips/index.cfm).

TARP and a Preferred Alternative The TARP proposal, which was pending before Congress at this writing, would have the U.S. government spend up to $700 billion acquiring distressed assets from financial institutions. The proposal has significant shortcomings. First, it places taxpayers in a first-loss position with respect to the assets they buy. To mitigate that problem, Congress added several proposed items, including the awarding of stock warrants to the government by asset-selling institutions, ex post assessments to be paid by all surviving financial institutions (to be designed subsequently) to recoup any ultimate taxpayer losses, limits on executive pay, and a variety of other features. These features reflect the desire to insulate taxpayers from the large potential risks associated with the acquisition of subprime-related assets and other assets, and entail significant uncertainties for taxpayers and participating institutions from their implementation. The asset purchases and the various risk-mitigating measures also provide extraordinary discretion to the Secretary of the Treasury.

08 Book.indb 76

2/13/09 3:58:19 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

77

Second, the plan is to purchase assets at above “fire-sale” prices but below “hold-to-maturity” value (to use Chairman Bernanke’s terms). This aspect of the plan reflects the recognition that purchasing assets at the lowest possible price in the midst of a liquidity crisis would do little to help banks, since it would not add to the capital of sellers and could force all banks to mark their portfolios to extremely low values. Given that most of these instruments do not trade in a secondary market, are highly heterogeneous and complex, and are not going to be purchased at the lowest (i.e., current market) price, it is hard to see how their prices will be determined. Discretionary authority combined with an ill-defined objective is a recipe for mischief, unaccountability, and even corruption. Third, the plan entails moving a huge amount of the financial system’s assets out of the private sector and into the public sector. This may be good news for the price of Northern Virginia’s real estate, but it will produce inefficient disposition of assets and reduce employment in New York’s financial center at a time when job losses there are already quite high. There is a better way. The Reconstruction Finance Corporation’s (RFC) preferred stock program, which began in 1933, was quite successful at giving banks needed capital and liquidity in the 1930s, and it did so at minimal risk to taxpayers. Infusing banks with preferred stock protects taxpayers against loss by making recipient bank stockholders bear the first tier of losses on their assets (thus avoiding the need for complex contracting schemes involving warrants, assessments and compensation limits), avoids the near-impossible task of pricing subprime-related securities, and keeps the workout of distressed assets in private hands (and in New York). The U.S. experience in the 1930s and Finland’s in the 1990s show that preferred stock injections can boost systemic stability with little risk to taxpayers (Mason 2001, Englund and Vihriala 2003, Calomiris and Mason 2004b). The RFC was successful in limiting the abuse of its preferred stock investments because it codified and followed clear practices specifically designed to limit abuse. Those included limiting common stock dividend payments, requiring recipients to devise a plan to increase capital, and retaining significant corporate governance

08 Book.indb 77

2/13/09 3:58:19 PM

78

Charles W. Calomiris

authority to limit abuse of protection. A properly designed RFC approach is head-and-shoulders better than the TARP approach being advocated by Messrs. Paulson and Bernanke. Will U.S. House Prices Collapse? If the above account of the origins of the turmoil is correct in placing significant blame on agency problems in asset management, then that implies an important corollary: Agency problems are also likely causing an overreaction to the subprime shock. Over-selling on the downside is a standard theoretical and empirical result in the literature on agency in asset management. It results from the desire of portfolio managers to avoid stocks that are seen by the public as obvious poor performers. The most dire predictions of financial sector loss begin with forecasts of a large decline in house prices. Using flawed measures of prices, many commentators believe that U.S. house prices have already fallen by more than 15% and may decline by substantially more in the near term. Such a decline implies that prime mortgages, not just subprime and Alt-A loans, could suffer substantial losses. The main worry is that a massive wave of subprime foreclosures and resulting distress sales of subprime borrowers’ houses will produce a steep house price decline for all houses, fueling further foreclosures (by “walkaway” prime borrowers) and leading to further price declines. Calomiris, Longhofer and Miles (2008) show that the empirical basis for this view is highly suspect.35 Roughly three-quarters of the U.S. mortgage market (measured in numbers of homes) is prime and conventional (non-subprime and non-jumbo). The value of these homes is accurately measured by the OFHEO indexes (there are two quarterly index numbers; one based on purchases of homes, the other based on both purchases and appraisals during refinancings— see Leventis 2007 for details). Regardless of which of the OFHEO indexes one employs, these price measures for the typical American home have not fallen much since the 2007 peak (Chart 13). Furthermore, even if dire foreclosure forecasts come true, Calomiris, Longhofer and Miles (2008) estimate that home prices as measured by the OFHEO index likely will not fall by very much (a peak-to-trough

08 Book.indb 78

2/13/09 3:58:19 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

79

Chart 13 U.S. Home Price Appreciation 240

240

220

200

180

180

160

160

140

140

120

120

100

100

2000-Q1 2000-Q2 2000-Q3 2000-Q4 2001-Q1 2001-Q2 2001-Q3 2001-Q4 2002-Q1 2002-Q2 2002-Q3 2002-Q4 2003-Q1 2003-Q2 2003-Q3 2003-Q4 2004-Q1 2004-Q2 2004-Q3 2004-Q4 2005-Q1 2005-Q2 2005-Q3 2005-Q4 2006-Q1 2006-Q2 2006-Q3 2006-Q4 2007-Q1 2007-Q2 2007-Q3 2007-Q4 2008-Q1 2008-Q2

Index: 2000Q1 = 100

200

220 Case-Shiller 10-city Composite Case-Shiller 20-city Composite Case-Shiller National OFHEO

Sources: S&P/Case-Shiller Home Price Indices (http://www2.standardandpoors.com/portal/site/sp/en/us/page.topic/indices_ csmahp/0,0,0,0,0,0,0,0,0,1,1,0,0,0,0,0.html); OFHEO, House Price Index (http://www.ofheo.gov/hpi_download.aspx).

decline of more than 5% would be a reasonable upper bound of average decline, even if foreclosures substantially exceed estimates for 2008 and 2009), although in roughly a dozen states declines will be (and already have been) severe (Chart 14). The OFHEO index is an accurate measure of the prices of houses financed in the prime mortgage market, and thus provides a clear indication of whether foreclosure activity is likely to produce significant price decline in that market. Other price indexes (the median sales price index, and the Case-Shiller national index—plotted in Chart 13) are biased measures of the overall housing market. CaseShiller, in particular, gives great weight currently to distressed subprime sales, and to jumbo sales, particularly in a few states (due to its uneven coverage of the national market). The OFHEO indexes, in contrast, mainly measure the value of houses in the prime market. Thus, there is little reason to believe that prime mortgages will suffer large losses from subprime foreclosures. If this upbeat assessment is correct, it is very good news for the recovery, since it indicates that the housing market is nearing its

08 Book.indb 79

2/13/09 3:58:20 PM

80

Charles W. Calomiris

Chart 14 Distribution of Forecasted Total House Price Changes between 2007Q2 and 2009Q4 Assuming Extreme Foreclosure Shock Extreme Foreclosure Shock

20

20

15

Number of States

15 12 11

10

10 8 6 5

5

5 3 1

1

1

1

1 0

0 -.24

-.21

-.18

-.15

-.12

-.09

-.06

-.03

0

.03

.06

.09

.12

Note: Alaska and New Hampshire are not included because of data limitations; the District of Columbia is included. Source: Calomiris, Longhofer and Miles (2008).

bottom. Recovery will not begin in earnest until markets become convinced that housing prices, which underlie so much of the uncertainty in the financial sector, have reached bottom. Calomiris, Longhofer, and Miles (2008) also argue that the OFHEO index is superior to Case-Shiller for measuring the consumption wealth effect of house price changes, since it is more representative of households whose consumption behavior is most likely to respond to house value decline. That argument reflects theoretical perspectives on the housing wealth effect (see Sinai and Souleles 2005, and Buiter 2008). Central banks’ macroeconomic models typically gauge the wealth effect using the OFHEO index as their measure of housing wealth, perhaps for similar reasons. The fact that the typical American home is unlikely to decline much in value over the period 2007-2010 due to the foreclosure wave buffeting the housing market, therefore, provides an optimistic perspective on consumption. The combination of a 5% OFHEO peak-to-trough price decline and a reasonable estimate of the housing wealth effect (a 3% elasticity) produces a very small decline in consumption.

08 Book.indb 80

2/13/09 3:58:20 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

81

Perspectives on Monetary Policy I have argued that the Fed’s aggressive actions with respect to the expansion in access to the discount window, the Fed-Treasury actions to prevent the collapse of Bear Stearns, and intervention to prevent the collapse of the GSEs, were appropriate responses to financial turmoil, although as many other commentators have correctly noted, in the case of the assistance to GSEs, government protection should have been delivered in a way that also committed to the right longterm resolution of the GSE problem. During an asymmetric-information shock, the central bank needs to be able to deliver targeted assistance. The discount window is a “surgical” tool used to combat localized problems (like the current securitization shock) without changing fed funds rates, and through them, interest rates throughout the financial system. Discount window lending inevitably entails some acceptance of risk by the central bank; to be useful, the collateral taken on loans should be good, but not riskless. At the same time, the discount window should not be used as a hidden means of transferring resources to insolvent borrowers (as the Fed did, and was roundly criticized for doing, during the 1980s). Being able to grant access to the discount window not only allows policymakers to target microeconomic assistance to put out fires with systemic consequences in the financial system, it also frees the monetary authority to keep the money supply and fed funds rate on an even keel, even during times of high stress. A bold use of the discount window, in other words, empowers the Fed to maintain a strong commitment to price stability even as it delivers assistance quickly where it is needed. Unfortunately, the Fed has not pursued a combination of bold lender-of-last-resort support alongside conservative policies to promote price stability. Aggressive fed funds cuts have permitted inflation to accelerate. During the turmoil, some voices within the Fed argued that core inflation provided a better indicator of long-term inflation, despite the fact that food and energy price inflation was obviously accelerating in a secular trend, rising alongside long-term inflation expectations and partly as a direct result of a weakening

08 Book.indb 81

2/13/09 3:58:20 PM

82

Charles W. Calomiris

dollar. This was unwise at best and disingenuous at worst. And Fed officials’ promises that rate cuts would be taken back in 2008 if inflation accelerated have proven hollow. To avoid a worsening economic contraction, banks and nonfinancial firms must be able to continue to access the stock and bond markets. U.S. corporations (whose debt capacity has improved over the past four years markedly, in response to the corporate leverage reduction wrought by the Bush dividend tax cuts—Chart 15) should be able to raise substantial funds in the bond market. But worries about inflation can limit buyers’ interest in new debt offerings. Ensuring price stability should be a priority for Fed policy, even from the standpoint of supporting the expansion of credit supply. Until the Fed raises the fed funds rate to demonstrate its concern about the acceleration of inflation, Fed pronouncements on price stability will be seen as cheap talk. Starting sooner rather than later, the Fed needs to raise the fed funds rate, slowly and predictably, to restore confidence in its continued commitment to price stability. Regulatory Policies With respect to regulatory policy, an important historical lesson is that bad regulations are often wrought in the wake of large financial shocks. Post-Depression regulatory changes (the separation of commercial and investment banks, the establishment of deposit insurance, the entrenchment of entry barriers across regions) are almost universally viewed by financial historians as mistaken reactions to the Depression which remained a source of major economic costs in the decades that followed (Calomiris 2000). It is important to emphasize that knee-jerk criticisms that blame the banking deregulation of recent decades for the subprime turmoil are dead wrong. As discussed above, bank deregulation and globalization over the past decades substantially reduced the costs of the subprime turmoil. But there have been regulatory mistakes, and they need fixing. This regulatory discussion focuses on six regulatory policy issues raised by the subprime turmoil:36 (1) prudential regulation of banks and other intermediaries, (2) policy toward the GSEs, (3) government

08 Book.indb 82

2/13/09 3:58:20 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

83

Chart 15 Corporate Leverage 0.60

0.60

0.50

0.50

0.40

0.40

0.30

0.30

Gross Corporate Leverage

0.20

0.20

Net Corporate Leverage

2008-Q1

2004-Q1

2000-Q1

1996-Q1

1992-Q1

1988-Q1

1984-Q1

1980-Q1

1976-Q1

1972- Q1

1968-Q1

1964-Q1

0.00 1960-Q1

0.00

1956-Q1

0.10

1952- Q1

0.10

Note: Gross corporate leverage is defined as liabilities divided by assets. Net corporate leverage is defined as liabilities, less cash, divided by assets. Cash is defined as total financial assets, less trade receivables, consumer credit, and miscellaneous assets. Sources: Federal Reserve Statistical Release Z.1, Table B.102 (http://www.federalreserve.gov/releases/z1/Current/data.htm).

policies designed to increase the rate of homeownership, (4) changes in the regulation of asset management, (5) the regulatory use of ratings for various purposes, and (6) foreclosure relief. Prudential regulation of banks has been shown to be inadequate, not just in retrospect, but in prospect. Critics of the status quo prior to the turmoil noted that the magic 8% number for total riskbased capital, and the lower limits on overall leverage enforced in the U.S., have long questioned whether these levels are adequate. Other longstanding criticisms have been that the chief pillars of Basel II—reliance on rating agencies opinions and reliance on internal models—have both been roundly discredited by the collapse of subprime. Many economists (see Repullo and Suarez 2008, for a review) have also noted the desirability of allowing minimum capital requirements to decline during downturns—to mitigate the credit supply contractions that accompany bank losses during downturns—but allowing such variation while also preserving sufficient equity buffers requires a substantial increase in the average minimum

08 Book.indb 83

2/13/09 3:58:20 PM

84

Charles W. Calomiris

capital ratio. This could be done at low cost to the economy if it were phased in over a long period of time (say over a decade or so). Once the economic recovery is underway, policy makers should begin the process of raising minimum capital requirements. The subprime debacle brings a deeper lesson, too. Banks used securitization to avoid prudential regulatory policies that tried to limit bank asset risk per unit of capital. If prudential regulation is going to be effective, it has to do more than make a new set of rules that clever bankers will innovate around. Regulation must take incentives into account and build rules that will be immune to creative accounting for risk. To accomplish that objective, capital requirements should also be made more dependent on debt market discipline, rather than just rating agency opinions or internal models. Many academics, within and outside the United States, have long favored the imposition of a minimum subordinated debt requirement as part of bank capital requirements (Shadow Financial Regulatory Committee 2000). While it is true that agency problems in asset management, like those revealed during the subprime turmoil, can weaken the accuracy of market opinions as expressed in the pricing of subordinated debts, the answer to that problem is to find ways to encourage better incentives by asset managers, not to give up on market discipline. Bankers who know that they will be subject to the risk judgments of sophisticated creditors, who place their own money at risk, will have strong incentives to limit the true underlying risk borne by those creditors. A minimum subordinated debt standard (which was supported by academic and Federal Reserve Board (1999) research, but killed by the political lobbying of the big banks in 1999), is the sine qua non of a credible approach to defeating regulatory arbitrage in banks’ risk-management practices. Largely in reaction to the disorderly LIBOR market over the past year, regulators are moving to require banks to meet minimum liquidity standards. It is likely that banks will be required to maintain adequate liquidity, not just adequate capital, as part of a reformed set of Basel requirements. Such a requirement would also reduce the dependency of banks on the Fed discount window during future financial shocks.

08 Book.indb 84

2/13/09 3:58:20 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

85

Another potential change in prudential regulation resulting from the subprime turmoil could be the imposition of prudential regulations on investment banks. Now that investment banks that are primary dealers have accessed the discount window and been the targets of other special Fed and Treasury intervention, is it possible to return to the status quo ex ante (where investment banks operate with neither the benefits of government protection nor the costs of adhering to strict guidelines for prudential regulation)? Much of the urgency of resolving that question was removed by the decisions of Morgan Stanley and Goldman Sachs to become bank holding companies under the regulation of the Federal Reserve Board. Still, the status of other investment banks, and of prospective entrants, remains unclear. The key unresolved issue is the extent of protection going forward. Unless the government can find a way to credibly avoid providing blanket protection to primary dealers that become troubled, prudential regulation of primary dealers would be necessary. On an optimistic note, reforms in over-the-counter markets are underway that would establish a central clearing house for some derivatives trading. This could substantially reduce and render more transparent the counterparty risks in derivatives trading. Doing so would reduce the potential costs of allowing a primary dealer to fail, and could thereby help limit the expansion of the safety net and the need to extend prudential regulation to the primary dealers. The genie is clearly out of the bottle with respect to GSE protection, which implies a pressing need to reform the GSEs. For over a decade, critics of the GSEs have been pointing out that the implicit protection afforded to them by the government invited abuse of taxpayers’ funds (Wallison 2001, Calomiris and Wallison 2008), and that there was no justification for preserving their unique mix of private ownership with government protection. Now that the government has bailed out the GSEs, taxpayers’ exposure is no longer implicit, it is explicit. The status quo ex ante is no longer acceptable. In the long term, the GSEs either should be divided into smaller institutions and credibly privatized, or should be wound down after

08 Book.indb 85

2/13/09 3:58:20 PM

86

Charles W. Calomiris

being nationalized. There are many acceptable ways to achieve one or the other of these options. The government has made a point of using credit subsidies as the primary means of encouraging homeownership—via tax deductibility of mortgage interest, FHA guarantees, support for GSEs and Federal Home Loan Banks, and pressures on lenders to expand access to credit for would-be homeowners. This has significantly contributed to unwise risk-taking and excessive leveraging in the real estate market, which promoted instability in the housing and financial markets. The argument typically made for subsidizing homeownership is that it increases people’s stake in their communities and makes them better citizens. A better way to achieve that objective is downpayment assistance for new homeowners (employed in Australia), which could deliver the same homeownership outcome in a way that stabilizes real estate markets and ensures that homeowners maintain a real stake in their homes. After all, how can homeownership significantly increase an individual’s stake in the community if the individual retains only a trivial stake in his or her home? Although it has received scant attention in the press, given the central importance of agency problems in asset management in triggering the recent turmoil, policy makers should be considering ways to reform the regulation of asset management to encourage better performance, greater competition, and more accountability. A good start would be the elimination of the symmetry requirement for profit sharing, which would permit asset managers to adopt compensation arrangements that would reward performance (along the lines of the arrangements employed by hedge funds). One can imagine other potential regulatory changes that might encourage greater competition and accountability on the part of institutional investors. This topic warrants more attention. The regulatory use of ratings, as discussed in Section 1, has contributed to ratings grade inflation, and given “plausible deniability” to value-destroying asset managers who made poor investments in subprime mortgage-related instruments.37 Unlike typical market actors, rating agencies are more likely to be insulated from the standard market penalty for being wrong, namely the loss of business. Issuers

08 Book.indb 86

2/13/09 3:58:20 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

87

must have ratings, even if investors don’t find them accurate. That fact reflects the unique power that the government confers on rating agencies to act as regulators, not just opinion providers. Portfolio regulations for banks, insurers, and pension funds set minimum ratings on debts these intermediaries are permitted to purchase. Thus, government has transferred substantial regulatory power to ratings agencies, since they now effectively decide which securities are safe enough for regulated intermediaries to hold. Ironically, giving rating agencies regulatory power reduces the value of ratings by creating an incentive for grade inflation and makes the meaning of ratings harder to discern. Regulated investors encourage grade inflation to make the menu of high-yielding securities available to them to purchase larger. The regulatory use of ratings changed the constituency demanding a rating from free-market investors interested in a conservative opinion to regulated investors looking for an inflated one. Grade inflation has been concentrated particularly in securitized products, where the demand is especially driven by regulated intermediaries. Even in the early 1990s, it was apparent how regulation was skewing the ratings industry. Cantor and Packer (1994) pointed out that grade inflation was occurring and that it was driven initially by ratings agencies other than Moody’s and S&P: “Rating-dependent financial regulators assume that the same letter ratings from different agencies imply the same levels of default risk. Most ‘third’ agencies, however, assign significantly higher ratings on average than Moody’s and Standard & Poor’s.” In fact, those “third” agencies were already pushing more heavily into structured finance than Moody’s and Standard & Poor’s, rating deals that the two main agencies did not. Moody’s and Standard & Poor’s eventually chose to join the others in what turned out to be an incredibly lucrative fast-growing product area, which accounted for roughly half of rating agencies’ fees. It is no use blaming the rating agencies, who are simply responding to the incentives inherent in the regulatory use of ratings. The right solution is for regulators to reclaim the regulatory power that has been transferred to rating agencies to both award ratings and determine the meanings attached to ratings. Such reform becomes even

08 Book.indb 87

2/13/09 3:58:20 PM

88

Charles W. Calomiris

more important in light of soon-to-be-adopted Basel II capital rules, which allow bond ratings to be used to measure default risk in regulating the portfolios of banks that do not develop their own models under Basel II’s Internal Risk-Based (IRB) Capital Rules. How can regulatory power be reclaimed? Regulating how rating agencies set standards is one possibility, but that would compromise rating agencies’ ability to use independent discretionary judgment. A better solution is to reform regulations to avoid the use of letter grades in setting standards for permissible investments by regulated institutions. In the absence of regulatory use of letter grades, banks and their regulators would look at the underlying risks of investments (their default probabilities and the expected losses given default), not letter grades. Indeed, rating agencies sell tools to investors that permit exactly this sort of analysis, and the IRB framework under Basel II presumes such data, which would render letter grades superfluous. Full disclosure of these new measures of portfolio risks and a greater reliance on market discipline to discourage excessive risk-taking would further improve the regulatory process. An even better reform would be to eliminate the regulatory use of ratings altogether. Regulation could substitute true market discipline through mandatory subordinated debt requirements, as discussed above.38 Not only would requiring banks to issue sub debt provide discipline from debtholders placing their funds at risk, the opinions of these market participants are publicly observable in bond prices and thus provide useful information to other investors and regulators. Congress and many states are considering various ideas for helping homeowners to avoid foreclosure. Many homeowners, particularly highly levered subprime borrowers who are facing rising interest rates as the result of teaser rate contracts, are facing a high risk of foreclosure. Compassion, and the desire to remove downward pressure on home prices from distress sales, motivate various aid proposals. The costs of such aid could be large, and the benefits in the form of higher home prices have been exaggerated (again, see Calomiris, Longhofer and Miles 2008). Costs include the moral-hazard consequences of encouraging high-risk borrowing in the future. To the extent that aid is provided, it should be targeted (e.g., to limit foreclosures on

08 Book.indb 88

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

89

primary residences of low-income homeowners) and should depend on renegotiation by creditors and lenders, not government intervention into the foreclosure process. Any aid should require lenders to make significant concessions to reduce borrowers’ leverage and reduce the risk of default going forward, and post-assistance cash-out refinancing should be strictly prohibited for borrowers participating in assistance programs. Long-Term Structural Consequences of the Subprime Turmoil Will securitization remain an important feature of financial intermediation, or has it been discredited too much by the subprime debacle? Over the last two decades securitization transformed financial intermediation. Advocates of efficiency gains from securitization point to the flexibility of securitization structures in carving up and distributing risk to meet different investors’ preferences for duration, default risk, interest rate risk, and prepayment risk. Securitization also can efficiently reduce the equity capital needed to absorb the risk of the assets being intermediated. Securitization mechanisms can perform that function by promoting learning about securitized assets over time (which reduces adverse selection costs) or by employing subtle contractual devices that improve the incentives of sponsors to manage risk (Calomiris and Mason 2004a). Critics see securitization as a means of promoting too much systemic risk by allowing banks to maintain inadequate minimum capital requirements, while retaining most or all of the risk of the assets being securitized. The absorption of much of the loss by sponsors of conduits has left many observers questioning whether securitization really does reallocate risk and whether it does so in a transparent fashion. The lack of reliability of the risk modeling for subprime MBS and CDOs has undermined confidence in the apparatus for engineering conduits and measuring the risks of their debt issues. Securitization of subprime and CDO conduits has given securitization a bad name, and the long-term future of securitization remains uncertain. But already we are seeing that the negative impact on securitization depends on the product line. For example, on the one hand, credit card securitizations seem to holding their own. They

08 Book.indb 89

2/13/09 3:58:21 PM

90

Charles W. Calomiris

have been around for decades, have operated through several business cycles, and have a well-understood track record. The master trusts under which debts are issued have evolved over time, and their complex structures (including early amortization structures that protect issuers and debtholders) have stood the test of time well. Deal flow in credit card securitizations remains high, and one could even argue that credit card securitization will benefit from the demise of subprime and other housing related products. On the other hand, more recent and exotic products, especially related to the residential or commercial mortgage sector, have been severely affected over the past year. Commercial MBS debt tranches with low loan-to-value ratios (e.g., 70% LTV tranches that are rated A) have seen yields in the high teens or even higher, and deal flow has been substantially reduced. Financial institutions are seeking to find a substitute mechanism in product areas where the market is less receptive to securitization. Covered bonds provide one possible solution. Indeed, one could argue that covered bonds are a more transparent version of the financial arrangements that previously characterized securitized assets. They similarly allow sponsors to carve up and redistribute risk, and permit separate categories of assets to serve as the bases for funding financial intermediation (rather than lumping everything together on the bank’s balance sheet and raising funds for the bank as a whole). Covered bonds are obligations of the issuing bank that issues them, but they are also linked directly to a set of assets that provide the first line of defense for repaying the cash flows promised to bondholders. This permits covered bond issuers to be rewarded for the performance of the asset pools on which the bonds are issued, as in a securitization, and it allows complex carving up of risks and targeting of risks to different (relatively junior and senior) bondholders. But debt service on covered bonds is a claim on the cash flows of the financial institution that issues them, not just the cash flows from the assets earmarked to support them, and covered bonds also are backed by the net worth of the issuing financial institution. While securitized assets enjoy the implicit backing of the sponsor’s holding company, this was conditional in the sense that there was no legal requirement by the sponsor to provide backing. Covered bonds entail a greater,

08 Book.indb 90

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

91

more explicit, and unconditional commitment for protection, and thus are quite different from securitization (Calomiris and Mason 2004a, Higgins and Mason 2004). That difference raises a concern for prudential regulation, namely cash flow and asset “stripping”—the possibility that the a bank’s commitment to its covered bond holders could cause a depletion of cash flow and assets that would otherwise support the institution as a whole (Eisenbeis 2008). So long as prudential regulation is effective, bank capital will be sufficient to provide protection against losses to other bank liabilities notwithstanding the use of covered bonds, but given the concerns noted above about the effectiveness of prudential regulation, it is worth recognizing that the use of covered bonds further reinforces the need for deep reforms of prudential regulation. Will Stand-Alone Investment Banks Disappear? Deregulation, culminating in the Gramm-Leach-Bliley Act of 1999, allowed commercial banks (i.e., those issuing deposits) to engage in a wide range of financial services. Why would a wholesale bank choose to remain as an investment bank after the deregulation of commercial banks’ powers? The primary advantage was avoiding the prudential regulations that applied to commercial banks. Although investment banks could not issue deposits, they could fund themselves with repurchase agreements (largely overnight), which substituted for short-term, low-interest rate deposits. The subprime crisis dramatically changed the perceived costs and benefits of remaining a stand-alone investment bank, as indicated by the disappearance of Lehman, the decisions by Morgan Stanley and Goldman Sachs to become bank holding companies, and the acquisitions of Bear Stearns and Merrill Lynch by JP Morgan Chase and Bank of America, respectively. It now seems likely that stand-alone investment banking will become the domain of small, niche players in the financial system. Obviously, the giant stand-alone investment banks didn’t want it to come to this. Why did they resist it for so long, and what does this

08 Book.indb 91

2/13/09 3:58:21 PM

92

Charles W. Calomiris

tell us about the downside of their capitulation for the structure and efficiency of the American financial system going forward? The investment banks’ resistance until now largely reflected the regulatory costs and risk “culture” changes that come with regulated depository banking. Virtually all of the franchise value of Goldman and Morgan is human capital. These folks are the most innovative product developers and the most skilled risk managers that the world has ever seen. Depository bank regulation, supervision, and examination prizes stability and predictability over innovativeness, and banks bear a great compliance burden associated not only with their financial condition, but also their “processes” related to both prudential regulatory compliance and consumer protection. None of that is conducive to innovation and nimble risk taking. Goldman’s and Morgan’s moves, therefore, could have a big cost in trimming their upside potential and reducing the value of their human capital for developing new products and proprietary trading strategies. What about the benefits? First and foremost, they will be able to use reliable, low-cost deposit financing as a substitute for the shrinking collateralized repo market and other high-priced marketbased debt instruments. Second, they will be able to preserve their client advisory business and perhaps even compete better in underwriting activities. Stand-alone investment banks have lost market share in underwriting to universal banks over the past two decades because underwriting and lending businesses are linked, and nondepository institutions suffer a comparative disadvantage in funding their lending (see Calomiris and Pornrojnangkool 2008). In this sense, the capitulation of the stand-alones marks the final stage in the victory of the relationship banking/universal banking model. Those of us who argued in the 1980s that nationwide branching would allow commercial banks to serve as platforms for universal banks with large relationship economies of scope can now say that we told you so. Bank of America, JP Morgan Chase, and Citibank have all weathered the financial storm and are not under immediate threat of failure precisely because their geographical and product diversification has kept them resilient and even permitted them to engage

08 Book.indb 92

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

93

in acquisitions and new stock offerings during the worst shock in postwar financial history. But it is not progress, in my mind, to move toward a one-size-fitsall financial system based entirely on behemoth universal depository banks. Just as community banks still play an important role in small business finance (owing to their local knowledge and flat organizational structures), we need nimble, innovative risk takers like Goldman and Morgan in the system. Still, I am not too worried about the lost long-run innovative capacity of American and global finance for a simple reason: Ultimately, people are the innovators, not institutions; smart, innovative people can (and many will) find homes elsewhere. The financial landscape will shift, giving rise to new franchises and new structures (perhaps even spinoffs from the current investment banks) that combine the features of the old franchises that don’t fit comfortably under the Fed’s umbrella. Global competition, as always, will be a reliable driver of financial efficiency. The structure of U.S. financial intermediation will probably undergo significant changes over the next few years, many of which are hard to predict. History does not give a precise guide to those changes, but one pattern is likely to repeat: Financial sector problems breed new opportunities alongside losses. The American financial system, if it remains true to its history, will adapt and innovate its way back to profitability and high stock prices sooner than is suggested by the dire predictions that fill today’s newspapers.

Author’s Note: For comments on previous drafts of this paper, which was titled, “Not (Yet) a Minsky Moment,” the author is grateful to: Lee Alston, the late George Benston, Michael Bordo, Richard Cantor, Mark Carey, Bob Eisenbeis, Richard C.S. Evans, John Geanakoplos, Thomas Glaessner, Gikas Hardouvelis, Kevin Hassett, Richard Herring, Charles Himmelberg, Glenn Hubbard, Charles Jones, Matthew Jozoff, Ed Kane, George Kaufman, Stanley Longhofer, Sebastian Mallaby, Joseph Mason, Chris Mayer, Douglas McManus, Allan Meltzer, Zanny Minton Beddoes, Jacques Rolfo, Eric Santor, Amrit Sekhon, Suresh Sundaresan, Craig Torres,

08 Book.indb 93

2/13/09 3:58:21 PM

94

Charles W. Calomiris

Goetz Von Peter, Peter Wallison, David Wheelock; participants at the Kansas City Fed’s Jackson Hole Conference, and participants at the FDIC 2007 Annual Bank Research Conference, the 2008 Annual Fain Lecture at Brown University, the IMF Conference on Financial Cycles, Liquidity and Securitization, the Conference on Money and Financial Markets at Ryerson University, the CRETE conference in Naxos Greece; and participants in other seminars at Columbia University, the American Enterprise Institute, the University of Frankfurt, the Federal Reserve Bank of New York, the Bank of Canada, the Central Bank of Cyprus, and the Hellenic Bank Association. Jeff West provided excellent research assistance. – October 1, 2008.

08 Book.indb 94

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

95

Endnotes Although agricultural problems continued from the 1920s into the 1930s, the Depression was not caused by shocks relating to a real estate bust. The Great Depression was caused primarily by shocks relating to worldwide monetary and exchange rate policy, which were propagated, in part, through their effects on the financial system. For a recent review of the contributing factors to the Depression, see Parker (2007). 1

By some measures, monetary policy was unusually accommodative during the subprime boom. The real fed funds rate, measured less the core PCE, or less the University of Michigan five-year expected inflation measure, was persistently negative from 2002-2005 to a degree only seen once before in the post-World War II era, in 1975-1978. The effects of loose monetary policy (which is generally confined to lowering only short-term interest rates) was magnified by global factors that promoted correspondingly low long-term rates (the so-called “conundrum”). Caballero et al. (2008) argue that special circumstances relating to the comparative advantage of financial intermediation in the United States can explain the conundrum. 2

These episodes are discussed in detail in Calomiris (2008).

3

Brazil is excluded from the list due to lack of available data.

4

Wheelock and Wilson (1994) show similar patterns, cross-sectionally, within Kansas. Banks in Kansas that voluntarily entered the Kansas deposit insurance system operated less prudently and suffered larger losses than other Kansas banks. The compulsory systems of Nebraska and the Dakotas, however, offered greater subsidization of risk and resulted in greater loss. 5

Obviously, the cases discussed above are not a complete list. Including other examples would confirm the central conclusion of this discussion. For example, the U.S. Panic of 1837 and Panic of 1857 also were significant financial crises with real estate aspects, particularly related to infrastructure expansion. The 1830s saw overbuilding of canals by state and local governments, through a combination of government expenditures, state government bond flotations, and loans from statechartered banks whose charters specifically envisioned financing these projects. The series of events that triggered the Panic of 1837 is controversial (Temin 1969, Schweikart 1987, Rousseau 2002), but whatever the trigger, the Panic brought huge losses related to prior infrastructure investment. The westward expansion of the 1850s resulted primarily from private investments in railroads, which was undermined by adverse political news relating to the brewing conflict over western expansion between the North and the South (Calomiris and Schweikart 1991). Compared to the Panic of 1857, the Panic of 1837 resulted in far more severe losses for banks and securities investors who financed the government-promoted real estate investments of the 1830s. 6

Because owners of rented residential properties are permitted to deduct their mortgage interest expenses, the benefit of which presumably is passed on to renters, 7

08 Book.indb 95

2/13/09 3:58:21 PM

96

Charles W. Calomiris

it is wrong to say that permitting homeowners to deduct their mortgage interest subsidizes homeownership; rather, it is perhaps better to say that allowing homeowners to deduct interest avoids taxing homeownership. 8 There is another category of theoretical models (which have fallen out of fashion in the past decade) that posit financial crises resulting from knife-edge phenomena relating to multiple equilibria and endogenous liquidity scarcity. I discuss this class of models elsewhere (Calomiris 2008), where I show that these models are of little use for understanding the likelihood, timing and varying severity of financial crises.

If this account is correct, it implies a testable hypothesis for future empirical work: Institutional investors who were investing their own money, or who are properly incentivized to focus on the long-run performance of their portfolios (i.e., many hedge fund managers), should have been more choosey about their investments in subprime mortgages and related CDOs. Casual empiricism is consistent with this prediction, although I am not aware of any formal analysis that supports it. 9

The modeling assumptions used in rating subprime pools have become much more transparent since the middle of 2007, and it is now possible to know LGD assumptions by type of product and by cohort, but this sort of information seems to be unavailable retrospectively. 10

The original collateral pool loss expectations for the 2006 subprime vintage were in a range between 5.5% and 6%, according to Moody’s (2007e). In 2004, some industry sources indicate that Moody’s expected loss assumption for subprime pools was 4.5%. 11

Foreclosure is a strategic decision on the part of borrowers and lenders, and thus reflects changes in house values. Calomiris, Longhofer and Miles (2008) show that negative shocks to house prices produce increases in foreclosures. 12

13 According to Fitch (2006b, p. 6), the 2004, 2005, and 2006 cohorts of subprime mortgages had average loan-to-value ratios of 81.5% and average loan sizes of roughly $163,000. On average, foreclosure costs on a home in the U.S. average roughly $59,000, which is a large fraction of the size of subprime mortgages (Getter 2007). Foreclosures, on average, are completed eighteen months after the first missed payment (Getter 2007). Costs consist of lost loan principal, real estate taxes and insurance payments, maintenance, real estate commissions, legal fees, and other physical collection costs. When house prices rise, some of the costs lenders bear in foreclosure are recoverable, although not all foreclosure costs can be recovered, even when home prices of foreclosed homes rise dramatically (Mason 2007). In essence, the LGD is kinked as a function of home price change: Home price declines have a one-for-one dollar effect on realized losses (since they reduce the ability to recover principal, accrued interest, and other recoverable costs one-forone), but home price appreciation only has a fractional effect on foregone losses, since some expenses cannot be recovered from the proceeds of the sale of the house.

08 Book.indb 96

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

97

For example, under an assumption of a 15% prospective decline in house prices, as of January 2008, JPMorgan projected that the LGD for a sample of Prime Alt-A Hybrid ARM portfolios that were originated in 2003 was 12.8%, but the LGD for the 2006 cohort of similar mortgages, under the same price change assumption, was 44.6%. That difference reflected the fact that on average the 2003 cohort had substantial equity (25.5% equity at origination plus 27.3% estimated appreciation from origination to January 2008), while the 2006 portfolio had 24.3% equity at origination and house prices were estimated to have declined 17.0% from origination to January 2008. Thus, the huge 31.8% estimated difference in LGD was attributable to a 44.3% difference in price change (less than a one-for-one effect). A reasonable forward-looking average LGD assumption for subprime mortgages prior to 2007 would probably have been upwards of 40%, consistent with realistic foreclosure cost estimates and a zero-price change housing outlook (Merrill Lynch 2007, p. 9), not the lower LGD numbers actually assumed prior to 2007. The low LGD assumptions employed reflected unrealistic assumptions of continuing home price appreciation, which persisted into the middle of 2007. 14 Low LGD assumptions also help to explain the rise of “no-docs” or “low-docs” subprime mortgages (less graciously called “liar” mortgages) that produced the uniquely loss-creating loan cohorts of 2005, 2006, and 2007 (Ellis 2008). The probability of default (PD)—which increases when screening is relaxed—matters less when the LGD is low. Cutting processing costs and time delays by adopting a no-docs process and charging a few extra percentage points of interest may be a more profitable way to run a subprime origination business, despite the adverse selection consequences for the pool of adopting this practice, if you believe that the LGD is low.

In July 2007, as problems in subprime started to appear, loss assumptions increased substantially to roughly 8-11% (Merrill Lynch 2007, Moody’s 2007a, 2007b, 2007c, 2007d). By the end of 2007, loss estimates had grown much more; in some subprime portfolios, estimated pool losses could exceed 50%. 15

According to Bloomberg Markets (July 2007, p. 56): “Corporate bonds rated Baa, the lowest Moody’s investment grade rating, had an average 2.2 percent default rate over five-year periods from 1983 to 2005, according to Moody’s. From 1993 to 2005, CDOs with the same Baa grade suffered five-year default rates of 24 percent, Moody’s found.” Long before the recent turmoil, Moody’s was aware that its Baa CDO securities were about 10 times as risky as its Baa corporate bonds. There was improvement in default experience on CDOs in 2006, and the default rate fell to 17%, reflecting that some previous impairments were cured in 2006. Nevertheless, the gap between corporate bonds and CDOs remained large. Based on additional data, through 2006, the comparable numbers are 2.1% and 17.0%. Moody’s refers to missed payments in CDOs as “impairments,” which are curable prior to maturity. Despite ratings’ agencies statements that letter grade ratings should represent consistent portrayals of risk across different debt instruments (e.g., corporate debt and debts from securitizations), in fact, this has not been the 16

08 Book.indb 97

2/13/09 3:58:21 PM

98

Charles W. Calomiris

case. For statements by ratings agencies affirming that ratings should have a consistent meaning “without regard to the bond market sector,” see Mason and Rosner (2007b, pp. 7-8, 19). 17 Interestingly, Moody’s (2007a) found that performance varied greatly across different subprime portfolios in ways that had not been foreseen; the identity of the originator was a very important determinant of differences in loss experience.

There were two important regulatory changes that took place in the last several years. In 2001, regulatory capital requirements were increased on junior stakes retained by sponsors; effectively, retaining a first-loss position in a securitization conduit required the sponsoring institution to maintain an equal amount of capital to the size of the retained position (http://www.occ.treas.gov/ftp/bulletin/2001-49a. pdf ). In contrast, holding AAA debts issued by the sponsor’s conduit required a 1.6% capital position against those AAA securities held (8% of a 20% risk weight). In 2004, regulators exempted conduit sponsors from the newly enacted GAAP consolidation rules for securitization (which in some cases would have otherwise required securitized assets to be treated as on-balance sheet assets for purposes of calculating capital requirements). Those 2004 regulations also established new rules for capital requirements on liquidity and credit enhancements from sponsors for their conduits (http://www.occ.treas.gov/fr/fedregister/69fr44908.pdf ). For example, an asset-backed commercial paper conduit with $100 million in securities as assets, issuing $90 million in commercial paper, with liquidity enhancement from the sponsor in the form of a line of credit of less than one year had to maintain $720,000 in capital against that credit line (8% x 10% “credit conversion factor” x $90 million). These regulations seem to have encouraged banks to use external enhancements and to hold AAA issues from their conduits, rather than hold firstloss positions in their conduits. 18

Of course, either through external enhancement or voluntary provision of support to their conduits, sponsors may still be taking a position that could result in large losses, and of course, many did so by absorbing losses that otherwise would have been born by other investors. 19

Arteta, Carey, Correa and Kotter (2008) analyze the risk choices of banks that established commercial paper issuing conduits. European banks were particularly heavy users of this means of finance. The authors argue that the relative reliance on this form of financing reflected several influences, including moral-hazard problems in risk management for heavy users. 20

Bartolini, Hilton, and Prati (2005) examine the LIBOR-fed funds spread prior to the turmoil and find that since 1990 (which marked an important regulatory change, eliminating reserve requirements on interbank borrowing in the Libor market) the two markets have been closely integrated. They find that during the 660 days of trading from February 11, 2002 to September 24, 2004, using actual transactions data from the two markets to compute hourly and daily spreads between the two markets, the two rates were always very similar. Using hourly data, 21

08 Book.indb 98

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

99

the two rates never diverge by more than 15 basis points, and reveal temporally scattered observations of gaps of 10-15 basis points only for 20 hours of trading during the 660-day period. Daily differences between the two rates are even smaller; spreads only exceed 5 basis points on 5 out of the 660 days, and never exceed 8 basis points. Chart 8, therefore, marks an unprecedented departure from the previously observed behavior of these two interest rates. The spread peaks August 10 at 128 basis points, and averages 49 basis points in the period August 9 to September 11. Bartolini, Hilton and Prati (2005) point out that “the Eurodollar market may draw a greater share of larger, more internationally-oriented institutions, which are more likely to operate foreign branches or International Banking Facilities through which they can borrow Eurodollars.” Bartolini, Hilton and Prati (2005) emphasize, therefore, that the counterparty risks in the two markets may not be identical. That observation suggests that the widening spread during the turmoil of August and September reflects adverse-selection problems that increased the counterparty risks for large-size transactions involving large, international banks (possibly the European banks with the large ABCP exposures discussed above), or rising liquidity demands by large banks that reflected their exposure to the subprime shock. The fed funds market, which often entails smaller transactions between small bank lenders and large bank borrowers should have been less affected by the liquidity demands of large banks or their adverse-selection problems, and apparently it was less affected. 22 See Sprague (1910), Gorton (1985), Calomiris and Gorton (1991), Calomiris and Schweikart (1991), Calomiris and Mason (1997), and Bruner and Carr (2007).

Schwarz (2008) is able to isolate default risk and liquidity effects on LIBOR spreads by comparing synthetic spreads (in which no financial instrument is held, and only default risk should affect pricing) with actual deposit transactions (in which both default risk and liquidity affect pricing). 23

Even at the height of the ABCP “run,” the aggregate liquidity risk for U.S. banks from the contraction of ABCP appears to not have been very large, although Citigroup stands out as the U.S. bank with more than its share of liquidity risk exposure (including its so-called structured investment vehicles, or SIVs, which issue a variety of debts, including ABCP). Much of U.S. ABCP consists of paper issued by so-called “multiseller issuers,” which tends to be maturity-matched so that liquidity risk is minimal. Most of the remaining ABCP can suffer from significant liquidity risk due to the mismatch between longer maturing assets (which include a wide variety of securities, loans, receivables, swaps, and repos) and short-term commercial paper liabilities. Most of that paper, however, was issued by foreign institutions. According to data from Moody’s, on average during the first quarter of 2007, of the $1.3 trillion in average ABCP outstanding administered (and, to a first approximation, issued by) the top 20 ABCP administrators, Citibank accounted for $98 billion, Bank of America accounted for $49 billion, and JPMorgan Chase accounted for $45 billion. Given the shrinkage in ABCP that has occurred over the past weeks, the total remaining liquidity risk exposure to U.S. banks from 24

08 Book.indb 99

2/13/09 3:58:21 PM

100

Charles W. Calomiris

ABCP issues, including any ABCP issued from SIVs, is roughly $100 billion, with Citigroup accounting for about half of that. This is a very small liquidity risk for the three American banks, given the sizes of their balance sheets and their liquid asset holdings. This discussion draws on data from Moody’s ABCP Program Index, March 31, 2007, and descriptions in JPMorgan Securities Inc. (2007). 25 From a regulatory capital standpoint, under Basel I rules, banks may have an incentive to purchase ABCP rather than fund its retirement via a line of credit, since a loan has a full risk weight, but commercial paper does not. Banks may also wish to purchase ABCP to resell it once market liquidity improves. It is unclear the extent to which ABCP that remains outstanding according to these data is being effectively retired by being purchased by banks that run the ABCP conduits.

In 1873, 1893, and 1907, suspension of convertibility stopped runs on New York City banks from continuing. Discount rates on cashier drafts on New York banks immediately after suspension show that market perceptions of risk of deposit loss were quite small even at times of extreme withdrawal pressure (just before suspensions), according to data reported in Sprague (1910). 26

Regulatory requirements include a 4% tier 1 risk-based capital requirement (as a fraction of risk-weighted assets), an 8% tier 1 plus tier 2 risk-based capital requirement (as a fraction of risk-weighted assets), and a leverage requirement (“adequately-capitalized” banks generally must maintain 4% of tier 1 capital relative to total assets; “well-capitalized” banks must maintain a ratio of 5% of tier 1 capital relative to total assets). It is highly desirable for banks to be considered “wellcapitalized,” and banks maintain a buffer above their minimum requirements. The leverage requirement is probably the most binding of these constraints going forward, especially since banks will be re-intermediating mortgage assets, which have less than a full risk weight, and likely will continue to maintain less than a full risk weight under Basel II. 27

Wheelock (2006) finds that, in the 1980s, substantial declines in real estate prices translated into significant deterioration in local banking condition. 28

For a review of branching deregulation and its positive effects on banking sector performance, see Calomiris (2000) and Jayaratne and Strahan (1996). Evidence on the role of regional shocks in banking distress and credit contraction during the 1980s is provided in Wheelock (2006); for the 1920s and 1930s, see Alston, Grove, and Wheelock (1994), Alston (1984), Calomiris (1992), Calomiris and Mason (1997, 2003a, and 2003b), and Calomiris and Wilson (2004). 29

30 It seems unlikely that fair value accounting has been of great use during the recent turmoil. Many market observers believe that fair value accounting has exaggerated losses (given the absence of useful transacting data, and the illiquidity of markets) and produced unreliable statements of earnings (Wallison 2008). More significant, to my mind, is the credibility of the regulatory environment, which allows investors to have some confidence that disclosures of bank exposures are

08 Book.indb 100

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

101

reasonably accurate. 31 There is a large literature measuring the moral-hazard costs of protection. These costs take various forms. For example, Alston (1984) shows that the foreclosure relief measures instituted to combat the agricultural distress of the 1920s and 1930s raised credit market costs for non-defaulting borrowers. Additionally, there is the cost of wasteful resource allocation from increased risk-taking. The academic literature looking at the adverse consequences for risk management of protecting banks is large. See, for example, Calomiris (1990), Barth, Caprio and Levine (2006), and Demirguc-Kunt, Kane, and Laeven (2008), among many others.

For a broader treatment of alternative mechanisms, see Calomiris, Klingebiel, and Laeven (2005). 32

For many interesting discussions of the application of this principle historically, see Meltzer (2003) and Capie and Wood (2007). 33

Although the exposure to loss was on the Fed’s balance sheet, it was indemnified by the Treasury, so it may be best to think of this arrangement as a Treasury action, facilitated by the Fed, rather than a Fed lending decision. 34

The study develops a quarterly Panel Vector Autoregressive model, using quarterly data at the state level since 1980 on employment, house sales, house permits, house prices, and foreclosures. We simulated house price declines for each of the states through 2009 by combining the model’s parameter estimates with state-level foreclosure estimates for 2008 and 2009 from economy.com. 35

36 Many other topics also warrant discussion, but not all can be treated here. The future of derivatives trading is of particular interest. Many observers are arguing that counterparty risk could be reduced by simplifying and homogenizing derivative contracts and encouraging their trading on exchanges, and by creating more efficient management of clearing and netting of positions. The allocation of regulatory and supervisory authority is another complex area of increasing debate. In particular, there are reasons to favor removing the Federal Reserve from the dayto-day business of supervision and regulation, as suggested by Secretary Paulson (see Calomiris 2006).

The discussion here relies heavily on Calomiris and Mason (2007).

37

For evidence of the desirability and feasibility of employing greater market discipline, see Board of Governors (1999), Mishkin (2001), and Barth, Caprio and Levine (2006). 38

08 Book.indb 101

2/13/09 3:58:21 PM

102

Charles W. Calomiris

References Adrian, Tobias, and Hyun Song Shin (2008). “Financial Intermediaries, Financial Stability and Monetary Policy,” Working Paper, Princeton University. Allen, Franklin, and Elena Carletti (2008). “The Role of Liquidity in Financial Crises,” Working Paper, University of Pennsylvania. Alston, Lee J. (1984). “Farm Foreclosure Moratorium Legislation: A Lesson from the Past,” American Economic Review, Vol. 74, June, 445-57. Alston, Lee J., Wayne A. Grove, and David C. Wheelock (1994). “Why Do Banks Fail? Evidence from the 1920s,” Explorations in Economic History 30, October, 409-31. Arteta, Carlos, Mark Carey, Ricardo Correa, and Jason Kotter (2008). “Revenge of the Steamroller: ABCP as a Window on Risk Choices,” Federal Reserve Board, September. Baer, Herbert L., and John N. McElravey (1993). “Capital Shocks and Bank Growth,” Economic Perspectives, Federal Reserve Bank of Chicago, July/August. Bagehot, Walter (1873). Lombard Street: A Description of the Money Market, Reprinted by Richard D. Irwin, Homewood, Illinois, 1962. Barth, James R., Gerard Caprio, Jr., and Ross Levine (2006). Rethinking Bank Regulation: Till Angels Govern, Cambridge University Press. Bartolini, Leonardo, Spence Hilton, and Alessandro Prati (2005). “Money Market Integration,” Working Paper, Federal Reserve Bank of New York. Bear Stearns (2007). “Global Commentary: Slowdown or Reset Button,” September 28. Bernanke, Ben S. (1983). “Non-Monetary Effects of the Financial Crisis in the Propagation of the Great Depression,” American Economic Review 73 (June), 257-76. Bernanke, Ben S., and Mark Gertler (1989). “Banking in General Equilibrium,” New Approaches to Monetary Economics, edited by William Barnett and Kenneth Singleton, Cambridge: Cambridge University Press. Bernanke, Ben S., and Mark Gertler (1990). “Financial Fragility and Economic Performance,” Quarterly Journal of Economics 104 (February), 87-114. Bernanke, Ben S., and Cara S. Lown (1991). “The Credit Crunch,” Brookings Papers on Economic Activity, No. 2, 205-48. Bernstein Research (2008a). “Bear Stearns: Sale Price $240 million, Break-Up Value of ‘Good’ Businesses Calculated at $7.7 billion,” March 17.

08 Book.indb 102

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

103

Bernstein Research (2008b). “U.S. Securities Industry: Funding Concerns Roil the Brokers,” March 17. Board of Governors of the Federal Reserve System (1999). “Using Subordinated Debt as an Instrument of Market Discipline,” Staff Study 172, December. Bogue, Allan (1955). Money at Interest: The Farm Mortgage on the Middle Border, Lincoln: University of Nebraska Press. Bordo, Michael D. (1985). “The Impact and International Transmission of Financial Crises: Some Historical Evidence, 1870-1933,” Revista di Storia Economica 2, 41-78. Bordo, Michael D. (1990). “The Lender of Last Resort: Alternative Views and Historical Experience,” Review, Federal Reserve Bank of Richmond (January/ February), 18-29. Bordo, Michael D. (2007). “The Crisis of 2007: The Same Old Story, Only the Players Have Changed,” Working Paper, Rutgers University. Bordo, Michael D., and David Wheelock (2007a). “Stock Market Booms and Monetary Policy in the Twentieth Century,” Review, Federal Reserve Bank of St. Louis, 89 (March/April). Bordo, Michael D., and David Wheelock (2007b). “When Do Stock Market Booms Occur: The Macroeconomic and Policy Environments of 20th Century Booms,” in The Origins and Development of Financial Markets and Institutions, edited by Jeremy Atack and Larry Neal, Cambridge: Cambridge University Press, forthcoming. Boyd, John H., and Mark Gertler (1994). “The Role of Large Banks in the Recent U.S. Banking Crisis,” Quarterly Review, Federal Reserve Bank of Minneapolis, Winter. Bruner, Robert E., and Sean D. Carr (2007). The Panic of 1907: Lessons Learned from the Market’s Perfect Storm, John Wiley & Sons. Brunnermeier, Markus K. (2008). “Deciphering the 2007-08 Liquidity and Credit Crunch,” Journal of Economic Perspectives, forthcoming. Buiter, Willem H. (2008). “Lessons from the North Atlantic Financial Crisis,” Working Paper, London School of Economics. Butlin, Matthew W. (1987). “Capital Markets,” in The Australian Economy in the Long Run, edited by Rodney Maddock and Ian W. Mclean, Cambridge University Press. Butlin, Noel G. (1964). Investment in Australian Economic Development, 1861-1914, Cambridge: Cambridge University Press. Butlin, Sydney J. (1961). Australia and New Zealand Bank: The Bank of Australasia and the Union Bank of Australia Limited, 1828-1951, London: Longmans.

08 Book.indb 103

2/13/09 3:58:21 PM

104

Charles W. Calomiris

Caballero, Ricardo J., Emmanuel Farhl, Pierre-Olivier Gourinchas (2008). “An Equilibrium Model of ‘Global Imbalances’ and Low Interest Rates,” American Economic Review, 358-93. Calomiris, Charles W. (1989). “Deposit Insurance: Lessons from the Record,” Economic Perspectives, Federal Reserve Bank of Chicago, May/June, 10-30. Calomiris, Charles W. (1990). “Is Deposit Insurance Necessary?” Journal of Economic History 50 (June), 283-95. Calomiris, Charles W. (1992). “Do Vulnerable Economies Need Deposit Insurance? Lessons from U.S. Agriculture in the 1920s,” in If Texas Were Chile: A Primer on Bank Regulation, edited by Philip L. Brock, The Sequoia Institute. Calomiris, Charles W. (1994). “Is the Discount Window Necessary? A Penn Central Perspective,” Review, Federal Reserve Bank of St. Louis, 76 (May/June), 31-56. Calomiris, Charles W. (1998). Review of Stock Market Crashes and Speculative Manias, edited by Eugene N. White, Journal of Economic History, September 1998, 614-617. Calomiris, Charles W. (2000). U.S. Bank Deregulation in Historical Perspective, Cambridge University Press. Calomiris, Charles W. (2001). “An Economist’s Case for GSE Reform,” in Serving Two Masters Yet Out of Control, edited by Peter J. Wallison, AEI Press. Calomiris, Charles W. (2006). “The Regulatory Record of the Greenspan Fed,” AEA Papers and Proceedings 96 (May), 170-73; a longer version of the paper is at http://www.aei.org/publications/filter.all,pubID.28191/pub_detail.asp. Calomiris, Charles W. (2007a). “Letter to Ms. Nancy Morris, Secretary, Securities and Exchange Commission, In Re: File No. S7-04-07 (SEC Proposed Rules Implementing Provisions of the Credit Rating Agency Reform Act of 2006),” May 4. Calomiris, Charles W. (2007b). “Expert Report In Re: Enron Creditors Recovery Group,” United States Bankruptcy Court, Southern District of New York, October 29. Calomiris, Charles W. (2008). “Victorian Perspectives on Modern Banking Crises,” Working Paper, Columbia Business School. Calomiris, Charles W., and Gary Gorton (1991). “The Origins of Banking Panics,” in Financial Markets and Financial Crises, edited by R. Glenn Hubbard, University of Chicago Press, 33-68. Calomiris, Charles W., Charles P. Himmelberg, and Paul Wachtel (1995). “Commercial Paper, Corporate Finance, and the Business Cycle: A Microeconomic Approach,” Carnegie-Rochester Series on Public Policy, 42, 203-250.

08 Book.indb 104

2/13/09 3:58:21 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

105

Calomiris, Charles W., and Charles M. Kahn (1991). “The Role of Demandable Debt in Structuring Optimal Banking Arrangements,” American Economic Review 8 (June), 497-513. Calomiris, Charles W., Daniela Klingebiel, and Luc Laeven (2005). “Financial Crisis Policies and Resolution Mechanisms: A Taxonomy from Cross-Country Experience,” in Systemic Financial Crises: Containment and Resolution, edited by Patarick Honohan and Luc Laeven, Cambridge University Press, 25-75. Calomiris, Charles W., Stanley D. Longhofer, and William Miles (2008). “The Foreclosure-House Price Nexus: Lessons from the 2007-2008 Housing Turmoil,” Working Paper, Columbia Business School. Calomiris, Charles W., and Joseph R. Mason (1997). “Contagion and Bank Failures During the Great Depression: The June 1932 Chicago Banking Panic,” American Economic Review, Vol. 87, December, 863-83. Calomiris, Charles W., and Joseph R. Mason (2003a). “Fundamentals, Panics, and Bank Distress During the Depression,” American Economic Review 93 (December), 1615-47. Calomiris, Charles W., and Joseph R. Mason (2003b). “Consequences of Bank Distress During the Great Depression,” American Economic Review, Vol. 93, June, 937-47. Calomiris, Charles W., and Joseph R. Mason (2004b). “How to Restructure Failed Banking Systems: Lessons from the U.S. in the 1930s and Japan in the 1990s” (with Joseph Mason), in Governance, Regulation, and Privatization in the AsiaPacific Region, edited by Takatoshi Ito and Anne Krueger, University of Chicago Press, 375-420. Calomiris, Charles W., and Joseph R. Mason (2004a). “Credit Card Securitization and Regulatory Arbitrage,” Journal of Financial Services Research, Vol. 26, 5-27. Calomiris, Charles W., and Joseph R. Mason (2007). “We Need a Better Way To Judge Risk,” Financial Times, August 23. Calomiris, Charles W., Athanasios Orphanides and Steven Sharpe (1997). “Leverage as a State Variable for Employment, Inventory Accumulation, and Fixed Investment,” in Asset Prices and the Real Economy, Forrest Capie and Geoffrey Wood, eds., Macmillan, 169-193. Calomiris, Charles W., and Thanavut Pornrojnangkool (2008). “Relationship Banking and the Pricing of Financial Services,” Working Paper, Columbia Business School.

08 Book.indb 105

2/13/09 3:58:21 PM

106

Charles W. Calomiris

Calomiris, Charles W., and Andrew Powell (2001). “Can Emerging Market Bank Regulators Establish Credible Discipline: The Case of Argentina, 1992-99,” in Prudential Supervision: What Works and What Doesn’t, edited by Frederic S. Mishkin, 147-96. Calomiris, Charles W., and Larry Schweikart (1991). “The Panic of 1857: Origins, Transmission, and Containment,” Journal of Economic History, 51, December, 807-34. Calomiris, Charles W., and Peter J. Wallison (2008). “The Last Trillion-Dollar Commitment: The Destruction of Fannie Mae and Freddie Mac,” AEI Financial Services Outlook, September. Calomiris, Charles W., and Berry Wilson (2004). “Bank Capital and Portfolio Management: The 1930s ‘Capital Crunch’ and Scramble to Shed Risk,” Journal of Business, Vol. 77, July, 421-56. Canovai, Tito (1911). The Banks of Issue in Italy, National Monetary Commission, U.S. Senate, 61st Congress, 2d Session, Document No. 575. Cantor, Richard, and Frank Packer (1994). “The Credit Rating Industry,” Federal Reserve Bank of New York Quarterly Review, Vol. 19, Summer-Fall, 1-26. Capie, Forrest H., and Geoffrey E. Wood, editors (2007). The Lender of Last Resort, London: Routledge. Caprio, Gerard, and Daniela Klingebiel (1996a). “Bank Insolvencies: Cross-Country Experience,” Working Paper No. 1620, World Bank. Caprio, Gerard, and Daniela Klingebiel (1996b). “Bank Insolvency: Bad Luck, Bad Policy or Bad Banking?” Annual Bank Conference on Development Economics, World Bank. Carey, Mark (1994). “Feeding the Fad: The Federal Land Banks, Land Market Efficiency, and the Farm Credit Crisis,” Working Paper, Federal Reserve Board of Governors. Citigroup Global Markets (2007). “Understanding the CP Crunch,” August 16. Cutts, Amy Crews, and Richard K. Green (2004). “Innovative Servicing Technology: Smart Enough to Keep People in Their Houses?” Working Paper BABC 04-19, Joint Center for Housing Studies, Harvard University, February. Davis, Lance E., and Robert E. Gallman (2001). Evolving Financial Markets and International Capital Flows: Britain, the Americas, and Australia, 1865-1914, Cambridge: Cambridge University Press. Dell’Ariccia, Giovanni, Deniz Igan, and Luc Laeven (2008). “Credit Booms and Lending Standards: Evidence from the Subprime Mortgage Market,” Working Paper, IMF, April.

08 Book.indb 106

2/13/09 3:58:22 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

107

Demirguc-Kunt, Kane, and Laeven, Editors (2008). Deposit Insurance Around the World, Cambridge, MA: MIT Press. Eisenbeis, Bob (2008). “Sliced Bread or Double Dipping? More on Covered Bonds,” Cumberland Advisors, July 30. Ellis, Luci (2008). “The Housing Meltdown: Why did it Happen First(?)/Only(?) in the United States?” Working Paper, Bank for International Settlements. Englund, Peter, and Vesa Vihriala (2003). “Financial Crises in Developed Economies: The Cases of Sweden and Finland,” Pellervo Economic Institute Working Paper No. 63, March. Fitch Ratings (2005). “Leveraged Super-Senior Credit Default Swaps,” July 26. Fitch Ratings (2006a). “2007 Global Structured Finance Outlook–Economic and Sector-by-Sector Analysis,” December. Fitch Ratings (2006b). “Mortgage Principles and Interest,” August. Geanakoplos, John (2008). “Leverage Cycles and the Anxious Economy,” Working Paper, Yale University. Getter, Darryl E. (2007). “Understanding Mortgage Foreclosure: Recent Events, the Process, and Costs,” CRS Report for Congress, Congressional Research Service, Order Code RL34232, November 5. Goodfriend, Marvin, and Robert G. King (1988). “Financial Deregulation, Monetary Policy, and Central Banking,” in Restructuring Banking and Financial Services in America, edited by William Haraf and Rose Marie Kushmeider, AEI Press, 216-53. Gorton, Gary (1985). “Clearing Houses and the Origin of Central Banking,” Journal of Economic History, 45, June, 277-83. Gorton, Gary (2008). “The Panic of 2007,” Working Paper, Yale University. Greenlaw, David, Jan Hatzius, Anil K. Kashyap, and Hyun Song Shin (2008). “Leveraged Losses: Lessons from the Mortgage Market Meltdown,” Working Paper, University of Chicago Graduate School of Business, February 29. Herring, Richard J. (2008). “The Darker Side of Securitization: How Subprime Lending Led to a Systemic Crisis,” Presentation to the Shadow Financial Regulatory Committee, May 4. Higgins, Eric, and Joseph R. Mason (2004). “What Is the Value of Recourse to Asset Backed Securities: A Study of Credit Card ABS Rescues,” Journal of Banking and Finance 28 (April), 857-74. International Monetary Fund (2008). Global Financial Stability Report, April 2008.

08 Book.indb 107

2/13/09 3:58:22 PM

108

Charles W. Calomiris

Jayaratne, Jith, and Philip E. Strahan (1996). “The Finance-Growth Nexus: Evidence from Bank Branch Deregulation,” Quarterly Journal of Economics, Vol. 111, August, 639-70. Jimenez, Gabriel, Steven Ongena, Jose Luis Peydro-Alcalde, and Jesus Saurina (2007). “Hazardous Times for Monetary Policy: What Do Twenty-Three Million Bank Loans Say About the Effects of Monetary Policy on Credit Risk?” Discussion Paper No. 6514, Center for Economic Policy Research, October. JPMorgan Securities Inc. (2007). “Short-Term Fixed Income Research Note,” August 16. Kaufman, George G. (1991). “Lender of Last Resort: A Contemporary Perspective,” Journal of Financial Services Research (October), 95-110. Kaufman, George G. (1992). “Lender of Last Resort, Too Large to Fail, and Deposit Insurance Reform,” in The Reform of Federal Deposit Insurance: Disciplining the Government and Protecting the Taxpayer, edited by James R. Barth and R. Dan Brumbaugh, Jr., New York: Harper Business, 246-58. Keys, Benjamin J., Tanmoy K. Mukherjee, Amit Seru, and Vikrant Vig (2008). “Did Securitization Lead to Lax Screening? Evidence from Subprime Loans,” Working Paper, London Business School. Kindleberger, Charles P. (1978). Manias, Panics, and Crashes: A History of Financial Crises, New York: Basic Books. Leventis, Andrew (2007). “A Note on the Differences between the OFHEO and S&P/Case-Shiller House Price Indexes,” OFHEO, July 25. Lucas, Douglas L., Laurie S. Goodman, and Frank J. Fabozzi (2006). Collateralized Debt Obligations: Structures and Analysis, 2nd Edition, Wiley. Mason, Joseph R. (2001). “Do Lender of Last Resort Policies Matter? The Effects of Reconstruction Finance Corporation Assistance to Banks During the Great Depression,” Journal of Financial Services Research 20, 77-95. Mason, Joseph R. (2007). “Mortgage Loan Modification: Promises and Pitfalls,” Working Paper, Drexel University, September. Mason, Joseph R., and Joshua Rosner (2007a). “How Resilient Are Mortgage Backed Securities to Collateralized Debt Obligation Market Disruptions,” Working Paper, Drexel University. Mason, Joseph R., and Joshua Rosner (2007b). “Where Did the Risk Go? How Misapplied Bond Ratings Cause Mortgage Backed Securities and Collateralized Debt Obligation Market Disruptions,” Working Paper, Drexel University. Meltzer, Allan H. (2003). The History of the Federal Reserve: Volume 1, 1913-1951, Chicago: University of Chicago Press.

08 Book.indb 108

2/13/09 3:58:22 PM

The Subprime Turmoil: What’s Old, What’s New and What’s Next

109

Mendoza, Enrique G., and Marco E. Terrones (2008). “An Anatomy of Credit Booms: Evidence from Macro Aggregates and Micro Data,” Working Paper, University of Maryland, April. Merrill Lynch (2007). “Mortgage Credit Losses: How Much, Where and When?” July 20. Minsky, Hyman P. (1975). John Maynard Keynes, Columbia University Press. Mishkin, Frederic S. (1991). “Asymmetric Information and Financial Crises: A Historical Perspective,” in Financial Markets and Financial Crises, edited by R. Glenn Hubbard, University of Chicago Press, 69-108. Mishkin, Frederic S. (Ed.) (2001). Prudential Supervision: What Works and What Doesn’t, University of Chicago Press. Moody’s Investors Service (2007a). “Moody’s Structured Finance Teleconference and Web Cast: RMBS and CDO Rating Actions,” July 12. Moody’s Investors Service (2007b). “U.S. Subprime—Overview of Recent Refinements to Moody’s Methodology: July 2007,” July 24. Moody’s Investors Service (2007c). “U.S. Subprime—Overview of Recent Refinements to Moody’s Methodology: July 2007,” August 2. Moody’s Investors Service (2007d). “US Subprime Mortgage Market Update: July 2007,” July 24. Moody’s Investors Service (2007e). “Testimony of Warren Kornfeld, Managing Director, Moody’s Investors Service, Before the Subcommittee on Financial Institutions and Consumer Credit U.S. House of Representatives,” May 8. Myers, Stewart C., and Nicholas Majluf (2004). “Corporate Finance and Investment Decisions When Firms Have Information That Investors Do Not Have,” Journal of Financial Economics, Vol. 5, 187-221. Parker, Randall E. (2007). The Economics of the Great Depression, Cheltenham, UK: Edward Elgar. Repullo, Rafael, and Javier Suarez (2008). “The Procyclical Effects of Basel II,” Working Paper Center for Economic Policy Research. Rousseau, Peter L. (2002). “Jacksonian Monetary Policy, Specie Flows, and the Panic of 1837,” Journal of Economic History 62 (June), 457-88. Schwartz, Anna J. (1992). “The Misuse of the Fed’s Discount Window,” Review, Federal Reserve Bank of St. Louis, (September/October), 58-69. Schwarz, Krista (2008). “Mind the Gap: Widening Risk Spreads Explained By New Measures of Credit and Liquidity,” Working Paper, Columbia Business School.

08 Book.indb 109

2/13/09 3:58:22 PM

110

Charles W. Calomiris

Schweikart, Larry (1987). Banking in the American South from the Age of Jackson to Reconstruction, Baton Rouge: Louisiana State University Press. Shadow Financial Regulatory Committee (2000). Reforming Bank Capital Regulation, Washington: AEI Press. Sinai, Todd, and Nicholas Souleles (2005). “Owner Occupied Housing as a Hedge Against Rent Risk,” Quarterly Journal of Economics 120 (May 2005), 763-89. Sprague, Oliver M.W. (1910). A History of Crises Under the National Banking System, National Monetary Commission, Reprinted by Augustus Kelley, Fairfield, 1977. Temin, Peter (1969). The Jacksonian Economy, New York: W.W. Norton. U.S. Shadow Financial Regulatory Committee (2000). Reforming Bank Capital Regulation, AEI Press. Von Peter (2008). “Asset Prices and Banking Distress: A Macroeconomic Approach,” Working Paper, Bank for International Settlements. Wallison, Peter J. (2001). Serving Two Masters Yet Out of Control, AEI Press. Wallison, Peter J. (2008). “Fair Value Accounting: A Critique,” Financial Services Outlook, American Enterprise Institute, July. Wheelock, David C. (2006). “What Happens to Banks When House Prices Fall? U.S. Regional Housing Busts of the 1980s and 1990s,” Federal Reserve Bank of St. Louis Review, September/October, 413-28. Wheelock, David C., and Paul W. Wilson (1994). “Can Deposit Insurance Increase the Risk of Bank Failure? Some Historical Evidence,” Review, Federal Reserve Bank of St. Louis, 76 (May/June), 57-71. White, Eugene N., Editor (1996). Stock Market Crashes and Speculative Manias, Cheltenham, UK: Edward Elgar. Wicker, Elmus (1996). The Banking Panics of the Great Depression, Cambridge: Cambridge University Press.

08 Book.indb 110

2/13/09 3:58:22 PM

Commentary: The Subprime Turmoil: What’s Old, What’s New and What’s Next? Michael Bordo

The Federal Reserve System was founded in 1913. One of its main goals was to deal with the problem of banking crises, which had plagued the National Banking era that followed the Civil War. The crisis of 1907 led to the call for the reform that created the National Monetary Commission in 1908, which in its report in 1912 called for the establishment of a U.S. central bank. The crisis of 1907-08 was the last of the major banking panics of the National Banking era. It led to the failures of numerous banks and ushered in a serious recession. It was noted for a rescue engineered by J.P. Morgan and ended, as did earlier panics, with the suspension of convertibility of deposits into currency. The fact that the U.S. Treasury was unable to resolve the crisis while J.P. Morgan did better was one of the causes of the popular movement to resolve the long-standing debate over the creation of a U.S. central bank. The crisis of 2007 also had some of the attributes of the current subprime turmoil. The crisis was initially centered on the New York trust companies, a financial innovation of that era. As it turned out, in the face of panic in October 1907, the trust companies were not covered by the safety net of the time—the New York Clearing House­—and the panic spread to the commercial banks.

111

08 Book.indb 111

2/13/09 3:58:22 PM

112

Michael Bordo

Today’s turmoil must be viewed in historical perspective. As Calomiris narrates, many of its attributes have been seen before. Chart 1 provides some background evidence for the U.S. over the past century. The upper panel from 1953 to the present shows the monthly spreads between the Baa corporate bond rate and the ten-year Treasury constant maturity bond rate. The spread, inter alia, represents a measure of the financial market’s assessment of credit risk and also a measure of financial instability reflecting asymmetric information (Mishkin 1991). Chart 2 takes a longer view and shows the data from 1921 to the present. Also displayed in both figures are National Bureau of Economic Research (NBER) recession dates and major financial market events, including stock market crashes, financial crises, and some major political events that affected financial markets. The lower panels of Charts 1 and 2 show policy interest rates—the federal funds rate since 1953 and the discount rate for the longer 20th century. As can be seen, the peaks in the credit cycle (proxied by the spreads) are often lined up with the upper turning points in the NBER reference cycles. Also, many of the events, especially the stock market crashes and the banking crises of the 1930s, occur close to the peaks. Moreover, the lower panel often shows the policy rate peaking very close to or before the peaks of the credit cycle. Its movements roughly reflect the tightening of policy before the bust and loosening in reaction to the oncoming recession afterwards. The rise in spreads in the recent episode are comparable to, but no higher than, what occurred in the last recession in 2001 and considerably lower than during the recessions of the 1970s, 1980s or the 1930s. I.

The Crisis

The crisis occurred following two years of rising policy interest rates. Its causes include lax regulatory oversight, a relaxation of normal standards of prudent lending and a period of abnormally low interest rates. The default on a significant fraction of subprime mortgages produced spillover effects around the world via the securitized mortgage derivatives into which these mortgages were bundled, to

08 Book.indb 112

2/13/09 3:58:22 PM

Commentary

113

Chart 1 Federal Funds Rate and Baa and 10-Year TCM Spread Percentage points

4

4

Bear Stearns Rescue

S and L crisis and Continental Illinois

Stock market crashes

Tech bust

3

3

Y2K

Penn Central

Subprime crisis

Kennedy assassination

2

2 Spread LTCM

1

0 1953

September 11

1

0 1963

1973

1983

1993

2003

20

20

16

16 Federal Funds Rate

12

12

8

8

4

4

0

0 1953

1963

1973

1983

1993

2003

Sources: Federal Reserve Board and NBER

08 Book.indb 113

2/13/09 3:58:23 PM

114

Michael Bordo

Chart 2 Discount Rate and Baa and Composite Treasury Over 10 Years Spread Percentage points

8

8

Bear Stearns Rescue

Banking Crises Tech bust 6 Stock market crashes

Penn Central

S and L crisis and Continental Illinois

Y2K

6 Subprime Crisis

Pearl Harbor September 11th

4 Kennedy assassination

4

Spread 2

2

LTCM 0

1921

1931

1941

1951

1961

1971

1981

1991

0

2001

16

16

14

14

12

12 Discount rate

10

10

8

8

6

6

4

4

2

2

0

1921

1931

1941

1951

1961

1971

1981

1991

2001

0

Sources: Federal Reserve Board and NBER

08 Book.indb 114

2/13/09 3:58:24 PM

Commentary

115

the balance sheets of investment banks, hedge funds and conduits (which are bank-owned but off their balance sheets), which intermediate between mortgage and other asset-backed commercial paper and long–term securities. The uncertainty about the value of the securities collateralized by these mortgages spreads uncertainty about the value of commercial paper collateral and uncertainty about the soundness of loans for leveraged buyouts. All of this led to the freezing up of the interbank lending market in August 2007 and substantial liquidity injections subsequently by the Federal Reserve and other central banks. Since then, the Fed has both extended and expanded its discount window facilities and also has cut the federal funds rate by over 200 basis points. The peak of the crisis was the rescue in March 2008 of the investment bank Bear Stearns by JP Morgan, backstopped by funds from the Federal Reserve and the creation of a number of discount window facilities whereby investment banks could access the window and which broadened the collateral acceptable for discounting. It was followed by a Federal Reserve/Treasury bailout of the GSEs, Fannie Mae and Freddie Mac, in July. The liquidity crisis subsequently turned into a credit crunch with recessionary potential as the drying up of securitization has forced banks to repatriate MBSs and CDOs to their balance sheets, putting pressure on their capital base and restricting their lending. II.

The Paper

Charlie Calomiris has written a masterful paper. He covers virtually all the bases on the subprime turmoil from an historical perspective. I learned a lot from reading the paper and find myself in agreement with most of his analysis and policy conclusions. According to Calomiris, the turmoil was caused by a number of factors: 1) loose monetary policy and the global savings glut; 2) financial innovation, especially the development and spread of securitization; 3) government regulation that encouraged excessive leverage in the mortgage market; 4) the “plausible deniability” hypothesis, aka an agency problem in asset management.

08 Book.indb 115

2/13/09 3:58:24 PM

116

Michael Bordo

The last point is the key contribution of this paper. According to the author, the main reason for the ballooning of securitization of subprime mortgages was that both investment managers and the ratings agencies accepted an unusually low estimate of the probability of default on subprime mortgages and of the losses to portfolios in the case of default. This estimate of 6% then allowed the ratings agencies to give AAA ratings to subprime mortgage-backed securities and investment managers to hold them in their portfolios. The low estimate was based on the default record and the losses incurred in the 2001-2003 housing recession when subprime mortgages were novel. This low estimate was then used to estimate future losses in subsequent years. The 6% estimate is downward biased because it was based on an episode when house prices were rising. This gradually reduced the losses incurred in the foreclosure process. In environments of flat or falling prices, which was the case after 2006, the losses would be much larger. However, everyone in the mortgage business accepted the low estimate of loss because it was in their interest to do so. It generated business for the ratings agencies who gave the high ratings on mortgage-backed securities and collateralized debt obligations and by the investment funds that were required by regulation to hold securities with AAA ratings. This pattern of excessive risk taking was fostered by the regulatory environment. According to Calomiris, the crisis was propagated, as in historical times, by asymmetric information manifest in rising spreads (flight to quality) and quantity rationing. The Bear Stearns collapse exhibited these characteristics. He argues that the crisis was managed appropriately by the Federal Reserve’s extension of its lender-of-last-resort facilities to encompass a wider range of collateral and the investment banks. He defends both the Bear Stearns rescue and that of the GSEs by saying that they were systemically serious and that they were dealt with in a way to minimize moral hazard. He argues, by extension of his earlier work on the Penn Central crisis of 1970, that discount window lending is a superior method compared to open market operations to overcome asymmetric

08 Book.indb 116

2/13/09 3:58:24 PM

Commentary

117

information problems leading to market failures (in the case of Bear Stearns, a collapse of the mortgage derivative markets; in the case of the GSEs, the collapse of mortgage financing). The author, however, is critical of the Fed’s cuts in the federal funds rate, which he views as both inflationary and conducive to a collapse of the dollar. Both factors, he posits, put downward pressure on the stock market, making it hard for banks to recapitalize themselves. Calomiris takes issue with the pessimists who see this crisis as sine qua non. His reasonable arguments are that the house price indexes used by the pessimists are severely downward biased and that the effect of the turmoil on bank capital has been greatly mitigated by the fact that the banks were in relatively good shape before the crisis reflecting deregulation and consolidation, and that globalization has enabled banks’ capital to be replenished by sovereign wealth funds. Finally, the author presents a list of policy reforms. He recommends regulatory reform including: the dethroning of the ratings agencies as official arbiters of quality; the use of subordinated debt as a disciplining device for the banks; either nationalizing or privatizing the GSEs; the use of covered bonds as a substitute for securitization; increasing capital requirements and imposing liquidity requirements on banks; and extending the Fed’s supervisory reach to include the investment banks that will ultimately become commercial banks. III.

Comments

Although I agree with much of the story, I have some reservations. First, I am not convinced by the case that Calomiris makes for providing liquidity through expanding access to the discount window rather than operating by open market purchases. He states that the discount window remains “an important component of the Fed’s toolkit” (pg.70). In fact, since the 1950s, use of the discount window has been minimal. A major change occurred in August 2007, leading with a cut in the discount rate. The change affected provision of credit directly to financial firms that the Fed deemed most in need of liquidity, in contrast to delivering liquidity to the market by open

08 Book.indb 117

2/13/09 3:58:24 PM

118

Michael Bordo

market purchases of Treasury securities and leaving the distribution of liquidity to individual firms to the market. Previously in 2003, the Fed set the discount rate to move just as did the fed funds rate. The choice of targeted lending instead of imperial liquidity provision by the market exposed the Fed to the temptation to politicize its selection of recipients of its credit. The Fed has created new programs for access to the discount window, including the Term Auction Facility (TAF), the Term Security Lending Facility (TSLF), and the Primary Dealer Credit Facility (PDCP). The oddest part of the creation of these new discount window loans is that they are sterilized. Nevertheless, net Fed assets have expanded 2-3% per year until recent weeks. Now they are up to 4-5% per year, which may account for the current 5% measured inflation rate. One question that arises is why this complicated method of providing liquidity has been introduced when the uncomplicated system of open market operations is available, and what has been achieved by the new facilities? A second question is why has the Fed reduced its holdings of government securities? How will the Fed be able to tighten monetary policy when it finally decides to combat the rise in the inflation rate? The only way to tighten is to sell government securities. The mortgage-backed securities now on the Fed’s balance sheet are not marketable. A second comment concerns the Bear Stearns rescue that Charlie approves of. Had Bear Stearns simply been closed and liquidated, it is unlikely that more demand for Fed credit would have come forward than that that actually occurred. The fact that general creditors and derivative counter parties of Bear Stearns were fully protected by the merger of the firm with JP Morgan Chase had greater spillover effects on the financial services industry than would have been the case had the Fed appointed a receiver and frozen old accounts and payments as of the date of the appointment. Fewer public funds would have been subjected to risk. When Drexel Burnham Lambert was shut down in 1990, there were no spillover effects.

08 Book.indb 118

2/13/09 3:58:24 PM

Commentary

119

A third comment is that Calomiris has not discussed the difficulty of pricing securities backed by a pool of assets, whether mortgage loans, student loans, commercial paper issues, or credit card receivables. Pricing securities based on a pool of assets is difficult because the quality of individual components of the pool varies, and unless each component is individually examined and evaluated, no accurate price of the security can be determined. As a result, the credit market—confronted by financial firms whose portfolios are filled with securities of uncertain value, derivatives that are so complex the art of pricing them has not been mastered—is plagued by the inability to determine which firms are solvent and which are not. Lenders are unwilling to extend loans when they cannot be sure that a borrower is creditworthy. This is a serious shortcoming of the securitization process that is responsible for paralysis of the credit market. Furthermore, the Fed has not recognized the solvency problem. It has emphasized providing liquidity to the market when that is not the answer to the problem of the market’s uncertainty about the solvency of individual or sectoral financial firms. No financial market can function normally when basic information about the solvency of market participants is lacking. The securities that are the product of securitization are the root of the turmoil in financial markets that began long before the housing market burst. Fourth, Calomiris is critical of the Fed’s cut in the federal funds rate by over 200 basis points since last August. However, given the Fed’s dual mandate to provide both price stability and high growth (full employment), the risk of recession consequent upon the credit crunch seems to be a reasonable rationale for a temporary easing of monetary policy. As the author does argue however, once inflationary expectations pick up, it behooves the Fed to return to its (implicit) inflation target. Finally, Charlie is critical of the Fed for allowing the dollar to weaken. In a freely floating exchange rate regime, the dollar should be at whatever level market forces dictate. The depreciation in the dollar until very recently, which reflected the cuts in U.S. policy rates with

08 Book.indb 119

2/13/09 3:58:24 PM

120

Michael Bordo

no similar changes abroad as well as greater recessionary pressure than abroad, has served to offset the recessionary pressures of the credit crunch and to reduce the feared current account deficit. Nevertheless, as he correctly points out, to the extent the weak dollar is reflecting inflationary expectations, that is a signal for policy tightening.

08 Book.indb 120

2/13/09 3:58:24 PM

General Discussion: The Subprime Turmoil: What’s Old, What’s New and What’s Next Chair: Martin Feldstein

Mr. Sinai: Charlie, I am puzzled by your logic of blaming the characteristics of crises on loose monetary policy. Chart 1 of Michael’s remarks underscores the opposite. It shows that the federal funds rate just prior to a bust is high really ahead of it. I would argue that most crises in the modern era have occurred in the presence of tight money, principally in response to high inflation. This time I think it is a balance sheet crunch—more than a credit crunch—that is a part of the upper turning point of the business cycle. That is new in its extent. Maybe you really mean that loose monetary policy helps create a later boom. Then the boom and higher inflation are part of the recession and bust in a crisis. This episode is different. You can’t put this at the door of the Federal Reserve. It is the financial system itself, the innards of it that have created the crisis. Loose monetary policy would to me make more sense as a cause if it were a prior precondition in time. Tight money and tight credit most often definitely have been characteristics of business cycle upper turning points. Mr. Blinder: Two questions for Charlie, although the first one will suggest an answer. The second one doesn’t suggest an answer. It really is a question.

121

08 Book.indb 121

2/13/09 3:58:24 PM

122

Chair: Martin Feldstein

At the end of your paper—you didn’t get there in your remarks­ —you talked about the prediction that stand-alone investment banks will wither away because of the superiority of raising money through deposits, which may be right. Ten or 12 pages earlier, you listed the establishment of deposit insurance in a list of changes “that are almost universally viewed by financial historians as mistaken reactions to the Depression.” First, I am not so sure how you square that. I am definitely sure it was a mistaken reaction to the Depression. Second is just a question, which is, You raised very interesting points about the compensation of asset managers. There is a related criticism that has been leveled, and I wonder what you think about that, which is that many asset managers get rewarded—and this goes right up the line (it is not just the trader)—on returns rather than risk-adjusted returns. Wouldn’t it be better if we could figure out a way to do it to reward them on risk-adjusted returns? Mr. Makin: Both Michael and Charlie discussed the issues of liquidity as connected with the Bear Stearns episode. Charlie suggested Bear Stearns was far from insolvency when it failed. I wanted to suggest a new concept—or maybe it’s not new. The problem that arose on that March 16 Sunday night was one of incipient insolvency. That is, Bear Stearns was experiencing a severe shortage of liquidity. Our concern was that, if they tried to address their liquidity problem by selling assets, the value of the assets of course would collapse and they would very quickly be insolvent. That same process could spread to other investment banks. That would bring on the systemic risk that Chairman Bernanke was talking about. I think it is a little dangerous to say, “Well, if we look at Bear ex post, they were far from insolvency.” They were going to be insolvent if there weren’t some intervention. The kind of intervention that Michael suggested perhaps was either receivership, where you shut the process down, or perhaps direct purchase of those assets. Either one of you might want to comment on that.

08 Book.indb 122

2/13/09 3:58:24 PM

General Discussion

123

Mr. Kashyap: Charlie, you implied the raising of equity can be viewed in isolation as something that is new in this crisis. Do you think if value-at-risk weren’t embedded in the risk-management procedures, this would have happened? I think not. What is new is the risk-management procedures that mandate when you see valueat-risk (VaR) skyrocketing, you either shrink the assets or raise the equity. The equity was a consequence of the VaR, not something that just happened by accident. From here forward we are going to keep seeing this dynamic of either you sell the assets or raise the equity. Mr. Hubbard: The principal point about the asset-management industry and compensation—of course, at Bear Stearns, there were already very high-powered incentives. The CEO lost $400 million in a month. My question for you is, Are you asking to change the 40 Act Rules for mutual funds? Is that what you had in mind? If so, are you trying to empower people to change structures, or do you want to regulate that? The issue is not just the rules for whatever mutual fund prices are, but the compensation of traders within an organization. Do you really want the Congress doing that? The second piece of that question is, There are two ways of going at this and they are not mutually exclusive. One is the compensation reforms you mentioned, and the other is letting capital requirements on institutions on the other side of the market move over the cycle. Do you see those as good complements? Would you prefer the comp reforms? Mr. Alexander: One of the questions I’d be interested in both the speaker and the commentator commenting on is, It seems to me one of the powerful sources of variation in this crisis is the fact that different institutions perform very differently. We think of this as something that was uniform but, even among large financial institutions like my own, perform differently than others. But also you have the different performance of hedge funds. I wondered what you make of that source of variation? A second question relates to cycles of innovation. One of the insights from behavioral finance is that people overreact to small amounts of information. One of the things that is relevant to this cycle is you have this very rapid development of subprime and related structure

08 Book.indb 123

2/13/09 3:58:24 PM

124

Chair: Martin Feldstein

credit. To a certain extent, people extrapolated from a period of good returns. That, it seems to me, is a very common phenomenon with respect to cycles of innovation. I would be interested in reaction as to how you see that in the historical record. Ms. Malmgren: I also have a question about the value-at-risk mathematical models, not in the aftermath but in the run-up to the crisis where essentially, because the models assume today’s price information is more important than yesterday’s, most of the catastrophic reference points had been knocked out of models, thus causing many institutions to say when volatility falls, we should double up the risk, and subsequent to the change in volatility, the lack of acquirement to update the volatility number. That is part of your agency risk. I wonder if you could comment on the relative contribution of this particular issue to the agency-risk picture? Mr. Meltzer: First, I compliment the paper as a comprehensive and excellent summary of what we know and what we need to know. One addition I would make would be something about the lenderof-last-resort policy. The Fed has had 95 years without ever enunciating a lender-of-last-resort policy. Sometimes it does, and sometimes it doesn’t. In the age of rational expectations, it is hard to justify allowing so much uncertainty about what its policy is going to be. Of course, one of the things that its lack of a stated and implicit policy encourages is the kind of intervention and pressures from Congress and Wall Street that have been so present in many, many crises and especially in the present crisis. A second comment concerns the role of regulation. Most regulation violates the first law of regulation. That is, the first law of regulation says that lawyers and bureaucrats make regulation and markets decide how to circumvent them. That has to really be borne in mind carefully in thinking about new regulation. Investment banks, for example, mark to market every day by borrowing short term. If they can’t borrow short term and they can’t mark to market, then they should become subject to FDICIA without having to go into all the portfolio analysis and argumentation.

08 Book.indb 124

2/13/09 3:58:24 PM

General Discussion

125

If you look at the history of regulation in the history of the Fed, regulating individual items just leads the Fed into wasting an enormous amount of time. For example, the implementation of Regulation Q got them involved in questions of whether the parking lots of banks should be counted as part of the interest payment that people receive, whether, if you gave the customer a safe deposit box, was that in lieu of interest and should you have to make a change for that, and there are just hundreds of these things which occupy the regulators all the time because they had a bad regulation and were having difficulty finding incentives that would make the regulation work. I think this was an excellent summary. Those are just two suggestions for additions. Mr. Goolsbee: The paper is pretty convincing in showing the ways that facilitated the speculation, but at two points I would caution you and ask you to justify a bit more of your argument. One is the presumption made at several points in the paper in talking about the changing in compensation rules is that because you are getting paid as a share of the assets you automatically are changing the desire for risk. The insight of the mutual fund literature is that just having a share of the assets doesn’t have any implication because the flows into and out of mutual funds are extremely performance sensitive. So it actually makes people more performance sensitive even though they are getting a share of the assets. The second place I would caution you is your conclusion that the government’s policies encouraging leverage have been somehow directly involved in this crisis. Almost all of the policies you cited are very long standing in origin. They are also, in many ways, not applicable to subprime. The mortgage interest deduction—a very large fraction of these people’s income is low enough and the mortgages are not high enough, so they are not even itemizing, so they weren’t even using that. Several of the regulations on banks were such that two-thirds of the subprime mortgages weren’t being made by banks, so the rules didn’t even apply to them. I would just caution you on those two issues.

08 Book.indb 125

2/13/09 3:58:24 PM

126

Chair: Martin Feldstein

Mr. Calomiris: I think Mike is doing something useful by posing explicitly the counterfactual of, What if we hadn’t used the discount window so much? My own view is, the surgical approach to targeting specific markets that have collapsed to prevent problems in those markets from spreading makes sense. That is the basic disagreement maybe that we have. That is, I start from the presumption that markets can collapse and stop functioning properly. The second thing I would say with regard to the Drexel Burnham example is, the world has changed a lot since Drexel Burnham. The Fed is very appropriately focusing on the need to improve the clearing process and the infrastructure now, so that it can allow investment banks to fail. I don’t want to paraphrase or try to say that is what I thought I heard the chairman saying, but that is certainly what I am arguing in this paper, and it is consistent with what the Fed is doing. Dollar weakening. I didn’t mean to say, if I did say, that the Fed should be targeting the dollar in some explicit sense. What I am saying is I am concerned that if there is a continuing loss in credibility for maintaining price stability that there could be a collapse of the dollar at some point in the future. I am less concerned about this than I was last month for a variety of reasons. But that would be a problem, because it would mean a collapse of consumption; it would mean a real recession; and it would also, of course, have dire implications for the stock and bond markets and for smoothing the effects of the credit crunch. Allen Sinai’s comment. I do explicitly mean what the papers I cite show, which is when interest rates are very low, you see credit standards deteriorating. And there is now a lot of microeconomic evidence for this direct link between times of very low interest rates and boom periods and the relaxation of credit standards. That is the point I am making. It was, I think, the way you interpreted it. Alan Blinder asked a lot of interesting questions. On stand-alone investment banks, I am simply suggesting the cost-benefit analysis has shifted. It is hard to predict. On deposit insurance, I would refer you to the paper Eugene White and I wrote and remind you all that

08 Book.indb 126

2/13/09 3:58:24 PM

General Discussion

127

the Treasury, the Fed, and President Roosevelt were very strongly opposed to deposit insurance, which was perceived at the time as special interest legislation being pushed by Henry Steagall, the Barney Frank of his time. It was an unnecessary innovation and very much a politically driven one. That doesn’t mean I am trying to get rid of deposit insurance now, which is not politically realistic. As far as long-term risk-reward tradeoffs, my view would be that this is exactly the topic we should be talking about in a whole conference. That is, how can we think about reforming risk-return tradeoffs? And Glenn’s question: Do I want to regulate this or deregulate it? I am mainly pushing, initially at least, in the direction of thinking that hedge fund incentives, which are more long-term directed toward value-maximization (which is the same as a risk-return tradeoff ), would be good to introduce more into the management of non-hedge fund investments. So it is a deregulation suggestion, not a new regulation. I recognize that I don’t have all the answers. There are some complicated issues there. Hedge fund incentives aren’t perfect either. I liked John Makin’s point, so I won’t comment more, except to say I’ll try to incorporate those. I also liked Anil’s point. I tried to make a version of that point, but maybe not as clearly. Moving to Lew Alexander, I would say that is the empirical paper I would next like to write, which is looking at how the different loss experiences in different financial institutions might reflect different incentives within those institutions. Also, Glenn raised an interesting point related to this, which is, How much do we need to worry about incentives—not just at the top like the Bear Stearns CEO­—but incentives all the way down to the asset managers? That is where I am focusing. I liked Pippa’s question. I don’t have a good answer to this question about measuring volatility. Of course, that was crucial for the quantitative equity trading that I didn’t talk much about. I didn’t think it was such a big deal actually, which is why I didn’t talk about it. Obviously, risk managers have to look beyond the current markedto-market volatility, particularly in an environment where the current pricing in the market might reflect asset management agency

08 Book.indb 127

2/13/09 3:58:24 PM

128

Chair: Martin Feldstein

problems. So you have to use your brain, not just mechanically plug in what the asset volatilities are that you are getting. Allan Meltzer made some great points. My hearing of Ben Bernanke’s comments was very sympathetic to this idea that we need to have a lender-of-last-resort policy, partly because it will reduce the problems of moral hazard. It sounded to me like there is a wonderful symphony of views on that idea. First law of regulation circumvention—absolutely! What is my suggestion? Subordinated debt requirements, of course, focus on market discipline in creating credible signals, forward-looking signals of risk, that might be more proactive. That is a big part of why I think they are a useful component. As far as Fed involvement, I basically agree with Secretary Paulson on this. The Fed needs to be a macro-prudential supervisor, but it needs to get out of the day-to-day supervision and regulation business like every other developed country in the world—other than New Zealand, which doesn’t have its own financial system after all. The Fed shouldn’t be deciding whether real estate brokerage is a financial activity, for example. That only weakens the Fed by politicizing it. Goolsbee’s point is the last one. Of course, it is relative performance that matters. The important point and the first point here is my Keynes’ quote that begins the paper by suggesting the following thinking may have been in operation, If you are a mutual fund manager and you know that all the others are taking the same bet you are taking, then you are not really at risk of a major relative performance problem, because you are all rising and falling together. In terms of leverage policies, I agree that in subprime the interest rate deduction is not an important incentive for risk taking— in fact, I even have a footnote saying it is not clear that we want to call the interest rate deduction a subsidy. But the other ones clearly are. Community Reinvestment Act pushes since the 1977 legislation, of course, have been a contributor to that problem.

08 Book.indb 128

2/13/09 3:58:24 PM

General Discussion

129

Mr. Bordo: Charlie basically discussed my main point, which is the counterfactual. I would really like to see what the counterfactual would have been had the Fed not did what it did with Bear Stearns and also with creating these special facilities. What if they had done what they had done before? What if they had used open market operations and used the discount window as it existed, given the legal mandate in the Federal Reserve Act, which does allow them and has allowed them to lend on the basis of many different kinds of collateral? So I wanted to know more about this assumption by everybody that the Fed did the right thing, that they were forced to do the right thing; and that if they hadn’t it would have been a disaster. I would like to see what the alternative counterfactual would have been. I am just not 100 percent convinced that counterfactual would show you that what the Fed did was the right thing. That is my reaction.

08 Book.indb 129

2/13/09 3:58:24 PM

The Panic of 2007 Gary B. Gorton

I.



Introduction With a full year elapsed since the panic of 1907 reached its crisis among this country’s financial markets, its banking institutions, and its productive industries, it ought to be possible to obtain an insight into the nature of that economic event such as could not easily have been obtained when the phenomena of the crisis itself surrounded us. —Alexander Noyes, “A Year after the Panic of 1907,” Quarterly Journal of Economics, February 1909.

We are now about one year since the onset of the Panic of 2007. The forces that hit financial markets in the U.S. in the summer of 2007 seemed like a force of nature, something akin to a hurricane, or an earthquake, something beyond human control. In August of that year, credit markets ceased to function completely, like the sudden arrival of a kind of “no trade theorem” in which no one would trade with you simply because you wanted to trade with them.1 True, thousands of people did not die, as in the recent natural disasters in Asia, so I do not mean to exaggerate. Still, thousands of borrowers are losing their homes, and thousands are losing their jobs, mostly bankers and others in the financial sector. Many blame the latter group for the plight of the former group; ironic, as not long ago the latter group was blamed 131

08 Book.indb 131

2/13/09 3:58:24 PM

132

Gary B.Gorton

for not lending to the former group (“redlining” it was called). The deadweight losses from bankruptcies, foreclosures, and job search are no doubt significant. Indeed, the feeling of the Panic of 2007 seems similar to that described by A. Piatt Andrew (1908A) a century ago, in commenting on the Panic of 1907: “The closing months of 1907 … were marked by an outburst of fright as wide-spread and unreasoning as that of fifty or seventy years before” (p. 290). Andrew (1908B) wrote that: “The autumn of 1907 witnessed what was probably the most extensive and prolonged breakdown of the country’s credit mechanism which has occurred since the establishment of the national banking system” (p. 497). The actions taken during that panic were extraordinary. They included legal holidays declared by governors and the extensive issuance of emergency currency through clearinghouses.2 It is true that today’s panic is not a banking panic in the sense that the traditional banking system was not initially at the forefront of the “bank” run as in 1907, but we have known for a long time that the banking system was metamorphosing into an off-balance sheet and derivatives world—the shadow banking system.3 Still, I would say that the current credit crisis is essentially a banking panic. Like the classic panics of the 19th and early 20th centuries in the U.S., holders of short-term liabilities (mostly commercial paper, but also repo) refused to fund “banks” due to rational fears of loss—in the current case, due to expected losses on subprime and subprime-related securities and subprime-linked derivatives. In the current case, the run started on off-balance sheet vehicles and led to a general sudden drying up of liquidity in the repo market, and a scramble for cash, as counterparties called collateral and refused to lend. As with the earlier panics, the problem at root is a lack of information.4 What is the information problem? The answer is in the details. Indeed, the details of the institutional setting and the security design are important for understanding banking panics generally. This should come as no surprise. Panics do not occur under all institutional settings or under all security designs. Contrary to most of the theoretical literature, historically it does not appear that panics are an

08 Book.indb 132

2/13/09 3:58:25 PM

The Panic of 2007

133

inherent feature of banking generally. This point has been made by Bordo (1985, 1986), Calomiris and Gorton (1991), and Calomiris (1993), among others. Bordo (1985), for example, concludes that: “the United States experienced panics in a period when they were a historical curiosity in other countries” (p. 73). Indeed, the same observation was made a century ago by Andrew (1908A): “In England no such general suspension of bank payments and no such premium upon money have occurred since the period of the Napoleonic wars; in France not since the war with Prussia…” (p. 290-91). Why is this point important? If one shares the viewpoint that panics are inherent to banking, then the details of panics perhaps do not matter. My viewpoint is that understanding panics requires a detailed knowledge of the setting.5 That is what I will try to provide here in the case of the Panic of 2007. How could a bursting of the house price bubble result in a systemic crisis?6 In this paper, I try to answer this last question. There are, of course, a myriad of other questions (many of them important, and some distractions from the real issues), but I focus on this one as the central issue for policy. I do not test any hypotheses in this paper, nor do I expound on any new economic theory. I include some anecdotal evidence, as well as observations from my own, and my colleagues’, experiences. I focus on describing the details of the financial instruments and structures involved and supply some very simple, stylized examples to illustrate their workings. Although I recognize that these details are probably rather boring for most people, I will argue that understanding the details of how the actual securities and structures involved are designed and intertwined is essential for addressing the most important questions.7 I develop the thesis that the interlinked or nested unique security designs that were necessary to make the subprime market function resulted in a loss of information to investors as the chain of structures—securities and special-purpose vehicles (SPVs)—stretched longer and longer. The chain of securities and the information problems that arose are unique to subprime mortgages— and that is an important message of this paper. Subprime mortgages are a financial innovation intended to allow poorer (and disproportionately minority) people and riskier borrowers access to mortgage finance in order to own homes. Indeed, these

08 Book.indb 133

2/13/09 3:58:25 PM

134

Gary B.Gorton

mortgages were popular. Subprime mortgage origination in 2005 and 2006 was about $1.2 trillion, of which 80 percent was securitized.8 The key security design feature of subprime mortgages was the ability of borrowers to finance and refinance their homes based on the capital gains due to house price appreciation over short horizons and then turning this into collateral for a new mortgage (or extracting the equity for consumption). The unique design of subprime mortgages resulted in unique structures for their securitization, reflecting the underlying mortgage design. Further, the subprime residential mortgage-backed securities (RMBS) bonds resulting from the securitization often populated the underlying portfolios of collateralized debt obligations (CDOs), which in turn were often designed for managed, amortizing portfolios of asset-backed securities (ABS), RMBS, and commercial mortgage-backed securities (CMBS). CDO tranches were then often sold to (market value) off-balance sheet vehicles or their risk was swapped in negative basis trades (defined and discussed below). Moreover, additional subprime securitization risk was created (though not on net) synthetically via credit default swaps (CDS) as inputs into (hybrid or synthetic) CDOs. This nesting or interlinking of securities, structures, and derivatives resulted in a loss of information and ultimately in a loss of confidence since, as a practical matter, looking through to the underlying mortgages and modeling the different levels of structure was not possible. And while this interlinking enabled the risk to be spread among many capital market participants, it resulted in a loss of transparency as to where these risks ultimately ended up. When house prices began to slow their growth and ultimately fall, the bubble bursting, the value of the chain of securities began to decrease. But, exactly which securities were affected? And, where were these securities? What was the expected loss? Even today we do not know the answers to these questions. In 2007, there was a run on off-balance sheet vehicles, such as structured investment vehicles (SIVs) and asset-backed commercial paper conduits (ABCP conduits), which were, to some extent, buyers of these bonds. Creditors holding the short-term debt, i.e., commercial paper, of these vehicles did not roll their positions, which was tantamount to a withdrawal

08 Book.indb 134

2/13/09 3:58:25 PM

The Panic of 2007

135

of funds. A number of hedge funds collapsed. As of this writing, the crisis is not over. An important part of the information story is the introduction, in 2006, of new synthetic indices of subprime risk, the ABX.HE (“ABX”) indices. These indices trade over-the-counter. For the first time information about subprime values and risks was aggregated and revealed. While the location of the risks was unknown, market participants could, for the first time, express views about the value of subprime bonds, by buying or selling protection. In 2007 the ABX prices plummeted. The common knowledge created, in a volatile way, ended up with the demand for protection pushing ABX prices down. The ABX information, together with the lack of information about location of the risks, led to a loss of confidence on the part of banks in the ability of their counterparties to honor contractual obligations. Securities wrapped by monoline insurers, such as auction rate notes, failed to re-auction and lost value, as monoline exposure to subprime was questioned. The entire financial system was engulfed when the ability to engage in repurchase agreements essentially disappeared. Collateral calls and the unwillingness to engage in repo transactions caused a scramble for cash. The bank-like system of off-balance sheet vehicles is beyond the reach of regulators, but migrates back to regulated institutions when things go bad.9 The assets of SIVs and conduits were absorbed back onto bank balance sheets. Liquidity for asset-backed securities and mortgage-backed securities, both cash and synthetic, dried up. Absent reliable market prices, accountants forced firms to “mark-to-market,” causing massive “write-downs” and resulting in reduced GAAP-based capital.10 Financial firms had to issue securities (at unfavorable terms) and sell assets, with the latter causing a further declines in prices—and subsequent further write-downs. Meanwhile, underneath all of this, millions of Americans face foreclosure on their homes due to being unable to refinance their mortgages or to make payments on their current mortgages.11 The information setting is complicated, but I try to develop the following story. The sell-side of the market (dealer banks, CDO, and SIV managers) understands the complexity of the subprime chain, while the buy-side (institutional investors) does not. Neither group

08 Book.indb 135

2/13/09 3:58:25 PM

136

Gary B.Gorton

knows where the risks are located, nor does either group know the value of every link in the chain. The chain made valuation opaque; information was lost as risk moved through the chain. The introduction of the ABX index revealed and aggregated values of the subprime bonds with centralized prices, until a breakdown of the index.12 At the root of the information story are the details of the chain. I detail the design of the various interlinked securities to develop the proposition that the uniqueness of these designs is at the root of the Panic. No other securitization asset class works like subprime mortgages, that is, no other asset class (e.g., credit card receivables, auto loans) is linked so sensitively to underlying prices. This distinction is important relative to the view of the Panic that seems to be coalescing into the common view. This view is known as the “originate-to-distribute” hypothesis, which very broadly claims that the last twenty–five years of change in banking has led to the current Panic because originators, it is alleged, have no incentive to maintain underwriting standards. I briefly discuss this hypothesis in a later section. In Section II, I briefly look at some background on mortgage markets and the development of the subprime mortgage market. Section III is devoted to explaining how subprime mortgages work. The focus is on implicit contract features, which link the functioning of these mortgages to home price appreciation. Subprime mortgage originators financed their businesses via securitization, but the securitization of subprime mortgages is very different from the securitization of other types of assets (e.g., prime mortgages, credit cards, auto loans). Subprime securitization has dynamic tranching as a function of excess spread and prepayment and is sensitive to house prices as a result. This is explained in Section IV. That is not the end of the story, because tranches of subprime RMBS were often sold to CDOs. Section V briefly explains the link to CDOs and the inner workings of these vehicles, the issuance of CDOs, links to subprime, and the synthetic creation of subprime RMBS risk. Section VI presents a very simplified example of the interlinked payoff structure of the securities to show the complexity and loss of information. The crisis also involves a widespread problem of liquidity, which is a topic

08 Book.indb 136

2/13/09 3:58:25 PM

The Panic of 2007

137

deserving of much more attention than I have space for here. Section VII is about the Panic itself, the falling house prices, the role of the ABX indices, the runs on the SIVs. I also try to summarize the information argument of the paper. In Section VIII I briefly discuss the liquidity crisis and some exacerbating factors: accounting and collateral calls. Section IX is devoted to the competing hypothesis, called “originate-to-distribute.” Concluding remarks are contained in Section X. II.

Some Background

In this section I begin with a very brief description of the evolution of subprime mortgages. Then I briefly look at the definition of “subprime” and the closely related category of “Alt-A” and review the issuance volumes and outstanding amounts of these mortgages. II.A. The Development of Subprime Mortgages Home ownership for low-income and minority households has been a long-standing national goal. Subprime mortgages were an innovation aimed at meeting this goal—and at making money for the innovators. The Harvard “1998 State of the Nation’s Housing Report” put it this way: In addition to a buoyant economy, the overall housing industry owes its enduring vigor to innovations in mortgage finance that have helped not only expand homeownership opportunities, but also reduce market volatility. Under market and regulatory pressure to make homebuying more accessible to low-income and minority households, financial institutions have revised their underwriting practices to make lending standards more flexible. In the process, they have developed several new products to enable more incomeconstrained and cash-strapped borrowers at the margin to qualify for mortgage loans. (Joint Center for Housing Studies, 1998, p. 8). In the same vein, Listokin, et al. (2000) noted:

08 Book.indb 137

2/13/09 3:58:25 PM

138

Gary B.Gorton

America’s housing and mortgage markets are in the midst of a dramatic transformation. After generations of discrimination and disinvestment, low-income and minority borrowers and neighborhoods now represent growth potential for homeownership and mortgage lending. In a movement that seems to reconcile socioeconomic equity with the imperatives of profitability in a competitive and turbulent industry, mortgage lending has emerged as the key to revitalizing the inner city, opening access to suburban housing markets, and promoting household wealth accumulation. Prodded by policy makers, the housing finance industry is now racing to tap new markets for homeownership by reaching traditionally underserved populations of racial and ethnic minorities, recent immigrants, Native Americans, and low- to moderateincome (LMI) households (p. 19). Subprime lending expanded during the 1990s, partly in response to changes in legislation affecting mortgage lending. See Temkin, et al. (2002) and Mansfield (2000) for the earlier history of subprime lending.13 Much of the change in mortgage products was due to technological change, which achieved efficiencies in standardizing loan products and allowed for the routinization of application procedures. For example, underwriting became automated, based on credit scoring models.14 The main issue to be confronted in providing mortgage finance for the unserved population was clearly that these borrowers are riskier. Subprime borrowers are, by definition, riskier than “prime” borrowers, so even if this risk is priced, there must be a decline in underwriting standards in order to provide mortgages to this segment of the population. But, more specifically, potential subprime borrowers have a number of issues which make them difficult bank customers. A Bank of America Mortgage study (cited by Listokin, et al., 2000, p. 98) noted the following problems: 1. Insufficient Funds for a Down Payment. Low-income or minority customers often are not able to save enough money for a down payment, particularly in rapidly appreciating markets. Intermittent employment and employment at lower-paying jobs

08 Book.indb 138

2/13/09 3:58:25 PM

The Panic of 2007

139

often make it hard for many such households to save (Smith, 1998).15 2. Credit Issues. BAMG (bank of America Mortgage) finds that roughly two-thirds of the LMI (low- and middleincome) population that it deals with has either no credit or lesser-rated credit, as measured by bureau or FICO scores (Smith, 1998). While it is the industry standard, the calibration of credit performance in bureau reports and FICO scores is deemed by BAMG to be far from a perfect measure when dealing with traditionally underserved populations. 3. Undocumented Income. The cash economy in many traditionally underserved communities means that “they [prospective home buyers] are earning income but cannot prove it in the way most lenders want them to, with a W-2” (Smith, 1998). 4. Lack of or Erroneous Information. As previously described regarding the Hispanic focus group study, many LMI, ethnic, and immigrant households are totally unfamiliar with the home-buying process or, worse, are misinformed on such matters as how much house they can afford and the minimum down payments required. BAMG underscores that there is not a monolithic underserved community, but rather that different segments of that community have varying problems. Some have strong credit but low savings, while others have some credit issues but have been better savers. To meet these different needs, BAMG introduced two new Neighborhood Advantage mortgages, Zero Down (launched April 1998) and Credit Flex (launched July 1998). Obviously, such households are risky propositions for lenders. If mortgages were to be extended to these borrowers, the underwriting standards would have to be different, and the structure of the mortgages would have to be different. For example, in 1998 Bank of America initiated two products to address this issue. One product, called the Neighborhood Advantage Zero Down, allowed low-tomoderate-income borrowers with good credit a 100 percent loan-tovalue (LTV) as well as gifts or grants to cover closing costs. The other

08 Book.indb 139

2/13/09 3:58:25 PM

140

Gary B.Gorton

product, called the Neighborhood Advantage Credit Flex, provided some flexibility to low-to-moderate-income borrowers subject to a documented alternative credit history. Other banks had similar products. See Listokin, et al. (2000). While the interest rate on a mortgage can be set to price the risk, such a rate is not likely affordable for these borrowers. So, the challenge was (and remains) to find a way to lend to such borrowers. The basic idea of a subprime loan recognizes that the dominant form of wealth of low-income households is potentially their home equity. If borrowers can lend to these households for a short time period, two or three years, at a high, but affordable interest rate, and equity is built up in their homes, then the mortgage can be refinanced with a lower LTV ratio, reflecting the embedded price appreciation.16 So, as detailed later, the mortgages were structured so that subprime lenders effectively have an (implicit) option on house prices. After the initial period of two or three years, there is a step-up interest rate, such that borrowers basically must refinance and the lender has the option to provide a new mortgage or not, depending on whether the house has increased in value. Lenders are long real estate, and are only safe if they believe that house prices will go up. This is detailed later. II.B. Subprime and Alt-A Mortgages The terms “subprime” and “Alt-A” are not official designations of any regulatory authority or rating agency. Basically, the terms refer to borrowers who are perceived to be riskier than the average borrower because of a poor credit history. However, the Interagency Expanded Guidance for Subprime Lending Programs defines a subprime borrower as one who displays one or more of the following features: • Two or more 30-day delinquencies in the last 12 months, or one or more 60-day delinquencies in the last 24 months; • Judgment, foreclosure, repossession, or charge-off in the last 24 months; • Bankruptcy in the last five years; • Relatively high probability of default as evidenced by, for example a FICO score of 660 or below;

08 Book.indb 140

2/13/09 3:58:25 PM

The Panic of 2007

141

• Debt service-to-income ratio of 50 percent or greater; or otherwise limited ability to cover family living expenses after deducting total debt-service requirements from monthly income. The market has adopted a somewhat larger, more ambiguous definition, one that is not standard across banks.17 As shown in Table 1, subprime borrowers typically have a FICO score below 640 and at some point were delinquent on some debt repayments in the previous 12 to 24 months, or they have filed for bankruptcy in the last few years.18 Whatever the definition, the innovation was successful, at least for a significant period of time. Tables 2 and 3, one for outstanding amounts and the other for issuance, show the size of the Alt-A and subprime mortgage markets relative to the total mortgage market and to the agency mortgage component of the market. The tables show: • The outstanding amounts of subprime and Alt-A combined amount to about one-quarter of the $6 trillion mortgage market. • Issuance in 2005 and 2006 of subprime and Alt-A mortgages was almost 30 percent of the mortgage market. • Over the period 2000-2007, the outstanding amount of agency mortgages doubled, but subprime grew 800 percent! • Since 2000, the subprime and Alt-A segments of the market grew at the expense of the agency share, which fell from almost 80 percent (by outstanding or issuance) to about half by issuance and 67 percent by outstanding amount. Many seem to hold the view that subprime mortgages are homogeneous. Aside from the attributes in the table of characteristics, this is not the case. Certainly, as is well-known, vintage of the mortgage is important. But also, even cross-sectionally, subprime mortgages are not homogeneous. That is, while they are all “subprime,” this does not mean that they are all the same across all dimensions, even holding vintage constant. Table 4 shows some of the heterogeneity of origination characteristics of the borrowers and the heterogeneity of experience of those borrowers across states from the 2006 vintage as of November 13, 2007. The table is from UBS (Mortgage Strategist, November 13, 2007,

08 Book.indb 141

2/13/09 3:58:25 PM

142

Gary B.Gorton

Table 1 Market Description of RMBS Categories Attribute

Prime

Jumbo

Alt-A

Subprime

Lien Position

1st Lien

1st Lien

1st Lien

Over 90% 1st Lien

Weighted Average LTV

Low 70s

Low 70s

Low 70s

Low 80s

Borrower FICO

700+ FICO

700+ FICO

640-730 FICO

500-660 FICO

Borrower Credit History

No credit derogatories

No credit derogatories

No credit derogatories

Credit derogatories

Conforming to Agency Criteria?

Conforming

Conforming by all standards but size

Non-conforming due to documentation or LTV

Non-conforming due to FICO, credit history, or documentation

Loan-to-Value (LTV)

65-80%

65-80%

70-100%

60-100%

Table 2 Non-Agency MBS Outstanding

Year

Total MBS

Agency

Total

Outstandings in $ Billions

Percent of Total MBS

Non-Agency Outstanding

Non-Agency Outstanding

Jumbo

Alt-A

Subprime

Agency

Total

Jumbo

Alt-A

Subprime

2000

3,003

2,625

377

252

44

81

87%

13%

8%

1%

3%

2001

3,409

2,975

434

275

50

109

87%

13%

8%

1%

3%

2002

3,802

3,313

489

256

67

167

87%

13%

7%

2%

4%

2003

4,005

3,394

611

254

102

254

85%

15%

6%

3%

6%

2004

4,481

3,467

1,014

353

230

431

77%

23%

8%

5%

10%

2005

5,201

3,608

1,593

441

510

641

69%

31%

8%

10%

12%

2006

5,829

3,905

1,924

462

730

732

67%

33%

8%

13%

13%

2007Q1

5,984

4,021

1,963

468

765

730

67%

33%

8%

13%

12%

Source: Federal Reserve Board, Inside MBS&ABS, LoanPerformance, UBS

08 Book.indb 142

2/13/09 3:58:25 PM

08 Book.indb 143

0.479

1.09

1.44

2.13

1.02

0.965

0.925

0.654

2000

2001

2002

2003

2004

2005

2006

7m 2007

Source: Inside MBS & ABS

Agency

Year

0.136

0.219

0.281

0.233

0.237

0.172

0.142

0.054

Jumbo

0.219

0.366

0.332

0.159

0.074

0.053

0.011

0.016

Alt-A

Non-Agency $ Bil.

0.176

0.449

0.465

0.363

0.195

0.123

0.087

0.052

Subprime

0.047

0.112

0.113

0.110

0.080

0.066

0.027

0.013

Other

1.23

2.07

2.16

1.88

2.72

1.86

1.35

0.615

Total MBS $ Bil.

53%

45%

45%

54%

78%

78%

80%

78%

Agency

11.0%

10.6%

13.0%

12.4%

8.7%

9.2%

10.5%

8.7%

Jumbo

Table 3 Gross Mortgage-Backed Security Issuance

17.8%

17.7%

15.4%

8.4%

2.7%

2.9%

0.8%

2.7%

Alt-A

14.3%

21.7%

21.6%

19.3%

7.2%

6.6%

6.4%

8.5%

Subprime

Percent of Total

3.8%

5.4%

5.3%

5.8%

2.9%

3.6%

2.0%

2.2%

Other

46.9%

55.3%

55.3%

45.9%

21.6%

22.3%

19.7%

22.1%

Non-Agency

The Panic of 2007 143

2/13/09 3:58:25 PM

144

Gary B.Gorton

p. 31). The last row is the total for the balances and is the weighted average for the characteristics.19 Table 4 shows: • The combined loan-to-value ratio (combo LTV) varies from about 80 percent to 91.5 percent. • All the state FICO scores are around 620. They vary from a low of 604 in West Virginia to a high of 644 in Hawaii. Note, however, that West Virginia’s percentage of loans that are 60 days or more delinquent is 6.67 percent, compared to a weighted national average of 16 percent. • The percentage of mortgages that are full doc varies from a minimum of 43.6 percent in New York to a maximum of 80.9 percent in Indiana. • Compared to “ALL,” note that the states Minnesota, California, Florida, Nevada, Rhode Island, Georgia, and Ohio are worse than the weighted average, in terms of percentage cumulative 60 days delinquent. In terms of cumulative loss, the experience varies from three basis points of loss in West Virginia to a maximum of 1.2 percent cumulative losses in Missouri. • House price appreciation (HPA) over the life of the loan, by state, shows a wide range of experience. • These are state averages, so the dispersion is undoubtedly greater. These observations are intended to convey the richness and complexity of the cross-sectional experience of different states. Even though subprime bond portfolios are fixed, and RMBS investors cannot easily choose state concentrations, there is some variation, which is relevant assuming house prices rise and defaults are idiosyncratic. But, portfolios tend to reflect the national concentrations of population, e.g., in California.

08 Book.indb 144

2/13/09 3:58:25 PM

08 Book.indb 145

Original Balance

$526,218,473

$1,849,884,555

$815,652,588

$14,428,873,327

$102,766,337,717

$5,292,370,638

$4,669,164,260

$1,194,568,797

$991,186,352

$43,832,887,130

$8,695,861,284

$3,018,554,281

$858,318,756

$1,415,015,589

$17,296,689,870

$2,885,253,658

$903,577,781

$1,317,753,384

$1,781,601,486

$9,065,659,267

$16,017,510,459

State

AK

AL

AR

AZ

CA

CO

CT

DC

DE

FL

GA

HI

IA

ID

IL

IN

KS

KY

LA

MA

MD

$10,727,182,750

$6,577,633,279

$1,539,635,309

$1,141,425,298

$699,004,465

$2,512,373,695

$11,903,745,425

$1,130,897,876

$683,838,875

$2,321,907,957

$6,981,317,691

$36,621,751,851

$794,565,683

$777,630,979

$3,861,877,916

$4,441,089,856

$82,358,162,338

$11,553,251,475

$697,886,978

$1,550,451,687

$399,461,897

Current Balance

68.5

73.8

89.0

89.8

79.7

88.7

69.7

82.4

81.0

78.7

81.5

86.4

83.8

65.8

85.2

86.9

82.3

83.2

87.5

85.1

76.4

Factor

84.9

84.9

89.1

90.4

90.8

90.7

88.7

86.7

90.6

83.0

91.3

85.5

85.8

79.3

84.7

91.4

86.3

85.8

89.9

89.5

88.4

Combo LTV

615

623

609

610

613

614

625

617

608

644

618

621

607

618

614

627

638

622

615

606

620

FICO

Table 4

62.7

55.8

68.8

78.7

78.8

76.0

56.9

70.2

80.9

46.2

67.1

50.8

71.1

52.5

60.8

70.4

46.6

58.7

73.1

76.1

70.4

% Full Doc

14.93

18.60

10.61

14.47

12.80

15.74

18.21

11.14

13.85

11.73

18.22

21.37

11.74

19.50

14.05

15.99

22.92

15.07

11.46

13.53

10.01

%60D+

1.00

1.70

0.65

1.24

1.05

1.45

1.16

1.58

0.94

1.14

2.51

1.26

0.77

1.81

1.20

2.50

2.33

1.53

1.50

1.24

0.96

%Cum Def

0.30

0.53

0.15

0.31

0.31

0.64

0.41

0.27

0.25

0.32

0.98

0.45

0.08

0.70

0.30

0.84

0.84

0.45

0.44

0.44

0.28

%Cum Loss

2.35

-1.82

5.93

3.36

3.56

2.42

3.71

7.32

3.13

4.21

4.11

1.33

4.39

0.63

0.95

1.44

-1.15

1.52

4.37

6.05

4.99

HPA Life

11.22

15.42

10.09

14.23

11.26

15.41

13.85

10.76

12.16

10.38

17.37

19.73

10.61

14.65

13.17

16.39

21.19

14.07

11.53

12.76

8.60

%Cum 60D+

The Panic of 2007 145

2/13/09 3:58:25 PM

08 Book.indb 146

$1,097,914,180

$6,820,690,521

$4,667,272,065

$3,654,696,377

$980,156,949

$410,267,389

$4,597,544,803

$93,805,229

$511,569,008

$1,361,125,986

$14,963,091,591

$1,377,416,203

$7,448,696,508

$22,383,244,240

$5,483,111,567

$1,221,051,933

$4,427,876,513

$6,978,493,823

$1,935,464,210

$2,359,469,767

$143,990,678

$3,863,653,816

$14,544,490,634

ME

MI

MN

MO

MS

MT

NC

ND

NE

NH

NJ

NM

NV

NY

OH

OK

OR

PA

RI

SC

SD

TN

TX

$12,691,323,091

$3,350,306,516

$125,448,463

$1,805,802,326

$1,506,722,871

$5,809,560,356

$3,595,736,620

$1,071,559,556

$4,690,730,151

$17,544,608,248

$6,276,562,378

$900,206,794

$10,011,731,473

$1,131,525,707

$448,252,110

$81,770,280

$3,520,500,657

$323,274,332

$855,069,697

$2,912,862,041

$3,835,369,086

$5,744,089,563

$793,716,799

90.4

88.5

88.5

78.0

79.6

86.0

83.7

90.5

87.1

79.7

87.6

66.9

68.0

86.3

89.6

88.2

78.1

81.5

89.1

81.4

83.6

85.4

74.4

89.6

91.5

91.0

88.1

84.9

85.5

87.2

90.3

90.6

84.3

88.2

87.1

83.8

85.0

91.4

91.3

89.7

85.5

89.8

89.5

89.6

89.8

84.2

616

615

616

612

621

608

629

610

613

633

631

615

620

614

614

616

613

617

605

607

626

613

615

66.7

75.7

75.7

70.7

55.5

70.0

70.3

76.8

76.3

43.6

54.5

68.7

48.8

63.8

77.7

77.4

73.5

65.7

74.7

74.0

64.5

66.5

62.7

Table 4 (continued)

11.51

12.38

11.53

13.18

19.87

10.88

9.59

12.39

17.89

18.58

19.61

9.01

18.12

13.34

10.98

8.59

11.16

8.64

15.19

15.40

23.92

22.31

15.16

2.05

2.21

0.28

1.14

2.53

0.53

1.17

1.12

1.08

1.60

1.89

0.69

1.10

1.27

1.28

1.09

1.31

1.45

1.69

2.91

1.72

1.79

0.61

0.74

0.74

0.07

0.29

0.95

0.12

0.23

0.33

0.36

0.38

0.60

0.12

0.26

0.35

0.51

0.15

0.31

0.10

0.56

1.20

0.70

0.86

0.15

6.43

6.09

4.37

6.41

-1.45

4.57

7.93

3.61

0.32

1.49

-1.60

7.46

1.77

0.38

2.19

5.96

6.89

8.06

5.56

3.43

0.61

-2.56

2.55

12.46

13.16

10.49

11.42

18.34

9.89

9.20

12.34

16.66

16.42

19.06

6.72

13.41

12.78

11.11

8.67

10.03

8.49

15.23

15.45

21.73

20.83

11.90

146 Gary B.Gorton

2/13/09 3:58:26 PM

08 Book.indb 147

$3,185,604,205

$10,125,147,122

$309,867,790

$9,550,742,478

$3,511,477,290

$454,297,185

$296,835,000

$378,382,004,715

UT

VA

VT

WA

WI

WV

WY

ALL

$299,265,424,087

$236,406,395

$347,335,187

$2,533,979,690

$7,505,680,840

$213,999,934

$7,702,473,341

$2,423,726,305

81.1

82.9

78.9

72.9

81.1

71.7

78.7

77.7

86.8

90.1

86.1

88.7

88.0

81.7

85.8

90.6

625

615

604

613

625

615

616

631

56.5

79.9

76.7

72.4

69.5

64.4

59.7

68.3

Table 4 (continued)

18.35

7.43

13.38

15.44

9.40

13.75

17.95

8.24

1.71

0.51

1.13

0.82

1.58

0.52

2.21

1.60

0.57

0.03

0.31

0.23

0.26

0.10

0.82

0.26

1.72

9.97

2.47

2.79

9.13

3.06

3.11

14.70

16.60

6.67

11.69

12.08

9.20

10.38

16.33

8.00

The Panic of 2007 147

2/13/09 3:58:26 PM

148

III.

Gary B.Gorton

Subprime Mortgage Design

The security design problem faced by mortgage lenders was this: How can a mortgage loan be designed to make lending to riskier borrowers possible? The defining feature of the subprime mortgage is the idea that the borrower and lender can benefit from house price appreciation over short horizons. The horizon is kept short to protect the lender’s exposure. Conditional on sufficient house price appreciation, the mortgage is rolled into another mortgage, possibly with a short horizon as well. The appreciation of the house can become the basis for refinancing every two or three years. In this section, I begin with an overview of subprime mortgages. The next subsection explains the details of how these mortgages work with a simple, stylized example. III.A. Overview The defining characteristic of a subprime mortgage is that it is designed to essentially force a refinancing after two or three years. Specifically, most subprime mortgages are adjustable-rate mortgages (ARMs) with a variation of a hybrid structure known as a “2/28” or “3/27.” Both 2/28 ARM and 3/27 ARM mortgages typically have 30-year amortizations. The main difference between these two types of ARMs is the length of time for which their interest rates are fixed and variable. In a 2/28 ARM, the “2” represents the number of initial years over which the mortgage rate remains fixed, while the “28” represents the number of years the interest rate paid on the mortgage will be floating. Similarly, the interest rate on a 3/27 ARM is fixed for three years after which time it floats for the remaining 27-year amortization. The margin that is charged over the reference rate depends on the borrower’s credit risk as well as prevailing market margins for other borrowers with similar credit risks.20 These mortgages are known as “hybrids” because they incorporate both fixed- and adjustable-rate features. The initial monthly payment is based on a “teaser” interest rate that is fixed for the first two years (for the 2/28) or three years (for the 3/27). Two important points are noteworthy about 2/28s and 3/27s. First, the fixed rate for

08 Book.indb 148

2/13/09 3:58:26 PM

The Panic of 2007

149

the first 2 or 3 years, the teaser rate, was not particularly low compared to prime mortgages. For example, the national average rate on a 2006 subprime 2/28 mortgage was 8.5 percent, and would reset on average to 6.1 percent over the benchmark LIBOR. (See Rosengren, 2007.) These high initial rates are not surprising because most of these mortgages were refinanced or the homes were sold prior to the mortgage being reset. As an example, on a 2/28 mortgage originated in 2006, the initial interest rate might have been 8.64 percent. After the initial period comes the rate “reset” (or step-up date), which is a higher interest rate, say LIBOR plus 6.22 percent. At the time of origination, LIBOR could have been 5.4 percent. So, the new interest rate at the reset would have been 11.62 percent. This rate floats, so it changes if LIBOR changes. The interest rate is updated every six months, subject to limits called adjustment caps. There is a cap on each subsequent adjustment called the “periodic cap” and a cap on the interest rate over the life of the loan called the “lifetime cap.” The reset rate is significantly higher, but potentially affordable. The above discussion emphasizes why the reset date on a hybrid ARM is so important. The higher payment for the borrower at the reset date comes from the significantly higher monthly mortgage payment that occurs at reset. Borrowers, thus, have an incentive to refinance their mortgage before the reset date. This is what I meant above by the term “essentially force” a refinancing. Another important characteristic of subprime mortgages is the size and prevalence of the prepayment penalties. See, e.g., Farris and Richardson (2004). Fannie Mae estimates that 80 percent of subprime mortgages have prepayment penalties, while only 2 percent of prime mortgages have prepayment penalties (see Zigas, Parry, and Weech, 2002). Further evidence for this comes from the prevalence of net interest margin securities (NIMs) in subprime securitizations. NIMs are securitizations of the early excess cash flows and prepayment penalties in subprime RMBS transactions. They are interestonly strips that derive their cash flow from the excess or residual cash flows, including significantly the prepayment penalties. See Bear Stearns (September 2006B); Frankel (2008); Zelmanovich, et

08 Book.indb 149

2/13/09 3:58:26 PM

150

Gary B.Gorton

al. (2007); and McDermott, Albergo, and Abrams (2001). I discuss NIMs further below. It is worth briefly contrasting a subprime mortgage with a standard, prime, 30-year, fixed-rate mortgage. With a prime mortgage, the borrower repays principal over time, and the mortgage matures after 30 years. The borrower may prepay the mortgage, typically without penalty. The borrower may benefit from house price appreciation, but the lender does not (directly) benefit. In effect, the lenders are not long house prices. I now turn to a simple, stylized example to try to understand how the design of the subprime mortgage addressed the riskiness of the borrowers. III.B. A Simple, Stylized Example The standard, prime mortgage is typically a fixed-rate 30-year loan. The usual way of thinking of mortgage design and pricing is to recognize the embedded optionality in these mortgages: The borrower has the right to prepay the mortgage (a call option to refinance) and the right to default (a put option).21 That is, the mortgage can be purchased from the lender at par, via prepayment, which is a call option, depending on interest rates. Or, the mortgage can be sold by the borrower to the lender for the value of the house, via default, amounting to a put option. The literature on this is voluminous. See Kau and Keenan (1995) for a review. A subprime mortgage is very different. Of course, borrowers can always prepay (but, subject to the prepayment penalty), and they can always default. But, as mentioned above, one important difference is that subprime mortgages typically have significantly higher prepayment penalties than prime mortgages (where it is typically zero). But, that is not the only important difference. The example below is intended to illustrate that a subprime mortgage contains an implicit embedded option on house prices for the lender. To the extent that this option is valuable, lenders may be willing to lend to riskier borrowers. The intuition is as follows. If house prices rise, and borrowers build up equity in their homes, they will become less risky, ceteris paribus. But, lenders are unwilling to speculate on house prices and borrower

08 Book.indb 150

2/13/09 3:58:26 PM

The Panic of 2007

151

repayment behavior for long periods, so they want the right to end the mortgage early, because foreclosure is costly. If borrowers “extract equity” through refinancing, after house prices have risen, then the plan of the lenders may not work. So, lenders incorporate high prepayment fees to try to prevent this. I develop these ideas with the example below. In my example, mortgages to prime borrowers would be made for two periods, but the candidate borrower that I will consider is rated “subprime,” and so the lender is unwilling to make a traditional two-period mortgage. The prospective borrower has a given income, which perhaps cannot be documented, and lacks money for a down payment. So, this mortgage, if made, would be to a borrower with no collateral. It is simply too risky to make a standard prime mortgage. To see how a subprime mortgage works, consider a lender who operates in a competitive market and faces a financing cost of rB per period. Let rM,t be the mortgage rate that the lender may offer for a subprime mortgage during period t. The amount of the mortgage is $L. Over period t the probability of borrower default is p(rM,tiL, LTVt), where the probability of default is increasing in the mortgage payment, rM,tL (implicitly relative to the borrower’s income), and in the loan-to-value (LTVt) ratio, which measures the equity stake the borrower has in the home.22 Borrowers work harder if they have an equity stake. To summarize, a higher mortgage payment and more debt relative to the home value increase the chance of defaulting. If there is a default, the recovery rate on the home value, Vt, at the end of period t is 50 percent, so for a mortgage of size $L, the lender would recover Rt=min[ 0.5Vt, L] if there is a default at the end of period t. Call Rt the “recovery amount” for period t.23 The subprime candidate borrower is applying for a mortgage of size $L for a home worth $L, so the LTV would be 100 percent. On a one-period mortgage, the lender breaks even if the mortgage rate, rM,1, is such that: (1+rM,1)(1- p(rM,1L, LTV1))L + R1 p(rM,1L, LTV1) – (1+rB)L = 0. (1) Of course, there may be no mortgage rate that satisfies (1). The lender cannot simply increase the mortgage interest rate because this increases

08 Book.indb 151

2/13/09 3:58:26 PM

152

Gary B.Gorton

the likelihood of default, as it becomes less likely that the borrower can make the higher mortgage payment. In any case, since, by assumption, the first period is rather short, realistically the borrower would have to refinance at the end of the first period, or default would be certain to occur. But, I have already ruled out granting long-term (two-period) mortgages to subprime borrowers as too risky. Suppose a subprime mortgage, as follows. The lender offers to extend a mortgage loan for the full two periods (imagine that period 1 is two years and period 2 is 28 years, though I omit the technicalities of discounting and so on), with an initial mortgage rate of rM,1 for the first period. Assume that the mortgage rate for the second period (the “step-up” rate) is prohibitively high so that the borrower must refinance the mortgage or default at that time. This is by design. Also, I will assume that the prepayment penalty is high. Suppose now that during any period there is a γ percent chance that house prices rise by Φ percent and a 1 – γ percent chance that they fall by Φ percent. During the first period, house prices will either rise or fall. For simplicity, assume that the house price change occurs an instant before the end of the first period, so that it does not affect the initial LTV ratio or the probability of default during the first period. Then, at the start of the second period, if house prices have risen, the LTV will have fallen to LTVD (the “D” subscript is for “down”). This corresponds to the borrower having positive equity in the home. On the other hand, if during the first period house prices have fallen, then the LTV will be higher, LTVU (“U” is for “up”), corresponding to the borrower having a negative equity position in the home. The assumed evolution of home prices affects the first period outcome—default or refinance. The evolution of house prices does not affect the probability of default (by assumption), but it does affect the recovery amount. If there is a default at the end of the first period, then the value of the house is different in the two cases of whether home price appreciation occurred or did not. Following the notation shown in Chart 1, the expected value of the first period mortgage, E(L1) is: (1+rM,1)(1- p(rM,1L, LTV0))L + γRD,1 p(rM,1L, LTV0) + (1– γ)RU,1 p(rM,1L, LTV0) – (1+rB,1)L

08 Book.indb 152

(2)

2/13/09 3:58:26 PM

The Panic of 2007

153

Chart 1 The Evolution of House Prices and the Loan-to-Value Ratio γ

LTVDD

LTVD γ

1−γ

LTVDU

LTV0 γ

1−γ

LTVUD

LTVU 1−γ

LTVUU

where RD,1=min[0.5(1+ Φ)Vt, L], in the case of house prices rising and LTV going down, and RU,1=min[ 0.5(1– Φ)Vt, L]; note that the subscripts on “R” refer to the LTV going down (D) since house prices went up and house prices rising corresponding to the LTV going up (U). If house prices fall at the end of the first period, assume that the initial lender will not refinance the mortgage (and neither will any other lender). The borrower now has negative equity and the likelihood of default going forward is (by assumption) too high for any lender. If home prices rise at the end of the first period, then the initial lender will be willing to refinance the mortgage. A rise in home prices over the first period has two effects: (1) the borrower has positive equity in the house, which is collateral from the point of view of the lender; this makes the lender’s recovery amount higher; (2) with a lower LTV going forward, the probability of default is lower, ceteris paribus, so the mortgage rate for the next period, rM,2, may be lower, making the payment lower, which also

08 Book.indb 153

2/13/09 3:58:26 PM

154

Gary B.Gorton

reduces the default likelihood. (Of course, as I discuss below, the borrower may extract the equity for consumption.) House prices may rise or fall over the second period. As before, I assume that house prices change an instant before the end of the period, and so the change does not affect the probability of default during the period. It does affect the recovery amount at the end of the second period. The expected value of the second-period mortgage (conditional on it being made), E(L2), is: (1+rM,2)(1- p(rM,2L, LTVD))L + γRDD,2 p(rM,2L, LTVD) + (1– γ) RDU,2 p(rM,2L, LTVD) – (1+rB,2)L. (2) Note that the second-period mortgage rate, rM,2 (and lender borrowing rate, rB,2), may be different than the first-period rate, and that the LTV ratio at the start of the period is now LTVD as house prices have risen. At the end of the second period, if house prices fell and the borrower defaults, the bank will recover RDU,2 ; the bank will recover RDD,2 if house prices rose. The expected payoff to the lender over the two periods (omitting discounting and the prepayment penalty) is: E(L1) + γE(L2). Note that the second-period mortgage is only made if prices have risen during the first period. This occurs with probability γ. At the end of the first period, the borrower is in a difficult spot because he either defaults or must refinance. The lender faces a choice, which depends on house prices. If house prices have risen (LTV goes down), the lender chooses max[RD,1, E(L2)] = E(L2). If house prices have fallen (LTV goes up), the lender chooses max[RU,1, E(L2)] = RU,1. In other words, the lender decides whether to refinance or take the recovery value. This is the optionality in the mortgage for the lender. It is an implicit option, as the strike price is the recovery amount, which depends on what house prices did over the second period. The lender does not take into account costs to the borrower from defaulting, if there are such costs.

08 Book.indb 154

2/13/09 3:58:26 PM

The Panic of 2007

155

The example makes the following points: 1. The key design features of a subprime mortgage are: (1) it is short term, making refinancing important; (2) there is a step-up mortgage rate that applies at the end of the first period, creating a strong incentive to refinance; and (3) there is a prepayment penalty, creating an incentive not to refinance early. If the step-up rate and the prepayment penalty are both sufficiently high so that without refinancing from the lender, the borrower will default, then the lender is in a position to decide what happens. The lender is essentially long the house, exposing the lender to house prices more sensitively than conventional mortgages. 2. In an important sense, the decision to default has effectively been transferred from the borrower to the lender. The step-up interest rate forces the borrower to come back to the lender after the first period, and the lender decides whether to extend another loan or not. Instead of the borrower having an option to default, the lender has an option to extend. 3. The design of the subprime mortgage creates the refinancing option. But, the borrower can refinance at the reset date with any originator. It may be that the subprime market is competitive with respect to initial mortgages, but not with respect to refinancing; borrowers are largely tied to their initial lenders.24 In that case, the original lender can benefit from any home price appreciation. 4. If E(L1)<0, i.e., the expected profit to the lender from the firstperiod loan is negative, then the refinancing must be tied to the original lender. The subprime mortgage, including the possible second-period refinancing, may be expected to be profitable if the probability of a house price increase, γ, is perceived to be sufficiently high. This happens if the borrower is tied to the original lender for refinancing. In fact, r­M,1, the first-period mortgage rate, may be set low (relative to the risk of loss due

08 Book.indb 155

2/13/09 3:58:26 PM

156

Gary B.Gorton

to default), as a teaser rate, making E(L1) negative, and still the overall loan may have a positive expected value if the probability of a house price increase, γ, is perceived to be sufficiently high. This may be viewed as “predatory” lending; the borrower is attracted to borrow, but may not understand that effectively it is the lender who makes the choice to refinance or not at the end of the first period. Refinancing does not mean that the borrower receives a long-term mortgage. The borrower could be rolled into another subprime loan. In fact, a borrower could receive a sequence of subprime loans, as house prices rise, each time building up equity and obtaining increasingly lower interest rates.25 But, in such a sequence, the lender effectively has the right to opt out by not refinancing and taking the recovery amount. In other words, a sequence of refinancings into subprime mortgages corresponds to a compound option for the lender. The borrower always has the right to prepay the mortgage, but with the higher prepayment fee. So far, I have assumed that this was prohibitively high. But, in practice, we do observe prepayments. In prime mortgages, this is usually the result of mortgage rates going down, as with prime mortgages. But, here there is another motivation as well. The borrower may want to extract equity value if house prices have risen. In my example above, one can imagine that this corresponds to the borrower and lender agreeing to refinance the loan at the end of period 1, but that the new mortgage allows the borrower to extract equity in the process. At the end of the first period, the borrower owes $L to the bank. If house prices have risen, the house is now worth (1+Φ)L. If the lender is willing to make the same subprime mortgage that was made at the start of period 1, then the borrower can extract $ ΦL. Such equity extraction is common in the prime market, but also very common—possibly more common, depending on the year—in the subprime market. In survey data, home equity extractions are often used for consumption. See Chomsisengphet and Pennington-Cross (2006, 2007) and Greenspan and Kennedy (2005, 2007). This is discussed further below.

08 Book.indb 156

2/13/09 3:58:26 PM

The Panic of 2007

157

III.C. Refinancing and Equity Extraction Between 1998 and 2006, subprime mortgages worked as they were supposed to. During this period, house prices rose and prepayment speeds were high; at least half of these mortgages (of all types) were refinanced within five years, and up to 80 percent of some types were refinanced within five years. See Bhardwaj and Sengupta (2008A). In other words, the bulk of the “originations” in the subprime market were refinancings of existing mortgages. Who got the benefit of the option on house prices? To the extent that lenders are willing to refinance the house even with equity extraction, there is a split of the capital gain. In that case, the borrower gets cash. Lenders only face a possibly safer borrower if equity is built up. Note that if E(L1)<0, then the lender will not want to allow equity extraction at the end of period 1 unless there is a large fee to compensate the lender for the foregone γE(L2). The benefits of refinancing were divided between lenders and borrowers, but we do not know the split. Greenspan and Kennedy (2007) estimate that during the period 1991-2005, $520 billion was extracted on average annually from all mortgages. Chart 2 shows the Greenspan and Kennedy estimates of net equity extraction and extraction as a percentage of personal disposable income.26 These data do not distinguish between prime and subprime mortgage extractions, and so just convey a sense of the magnitudes. Bhardwaj and Sengupta (2008B) report that the fraction of subprime refinancings that involved some equity extraction ranged from 51.3 percent to 58.6 percent over the period 1998-2007, with no trend.27 Chomsisengphet and Pennington-Cross (2006) examine the early period of the subprime market, prior to 2002, and show that a higher proportion of subprime refinancing involves equity extraction compared to prime refinancings. III.D. Summary To reiterate, no other consumer loan has the design feature that the borrower’s ability to repay is so sensitively linked to appreciation of an underlying asset. This sensitivity to the market price, the house price, will have far-reaching implications. But, if this was the end of the story, there would not have been a systemic banking crisis

08 Book.indb 157

2/13/09 3:58:26 PM

158

Gary B.Gorton

Chart 2 Net Equity Extraction and as a Percent of PDI 12

240

Net Equity Extracted

10

As a % of PDI

190 8

140 $bns

6

4 90 2 40 0

-10

1991Q1

1992Q3

1994Q1

1995Q3

1997Q1

1998Q3

2000Q1

2001Q3

2003Q1

2004Q3

2006Q1

-2

Source: Greenspan and Kennedy (2005, 2007)

(although obviously there would be the problem with foreclosures for many people).28 IV.

The Design and Complexity of Subprime RMBS Bonds

The next link in the chain concerns how the subprime mortgages were financed. This too will require a unique security design, quite different from traditional securitizations.29 The originators of subprime mortgages were largely new entrants into mortgage lending, including many of the names that later became well-known, such as Countrywide Financial, New Century, Option One and Ameriquest. The main financing method for subprime originators was securitization. This will be important not only because the risk will be spread but also because the structure of the securitization will have special features reflecting the design of the subprime mortgages. This latter point means that there will be additional complexity.

08 Book.indb 158

2/13/09 3:58:27 PM

The Panic of 2007

159

Table 5 Mortgage Originations and Subprime Securitization



Total Mortgage Originations (Billions)

Subprime Originations (Billions)

Subprime Share in Total Originations (% of dollar value)

Subprime Mortgage Backed Securities (Billions)

Percent Subprime Securitized (% of dollar value)

2001

$2 ,215

$190

8.6%

$95

50.4%

2002

$2,885

$231

8.0%

$121

52.7%

2003

$3,945

$335

8.5%

$202

60.5%

2004

$2,920

$540

18.5%

$401

74.3%

2005

$3,120

$625

20.0%

$507

81.2%

2006

$2,980

$600

20.1%

$483

80.5%

Sources: Inside Mortgage Finance, The 2007 Mortgage Market Statistical Annual, Key Data (2006), Joint Economic Committee (October 2007)

IV.A. Financing Subprime Mortgages via Securitization Table 5 shows the extent to which lenders relied on securitization for the financing of the mortgages. The table provides a snapshot of the quantitative importance of subprime securitizations. The table shows that subprime mortgage origination in 2005 and 2006 was about $1.2 trillion, of which 80 percent was securitized. IV.B. The Design of Subprime RMBS Subprime RMBS bonds are quite different from other securitizations because of the unique features that differentiate subprime mortgages from other mortgages. Like other securitizations, subprime RMBS bonds of a given transaction differ by seniority, but unlike other securitizations, the amounts of credit enhancement for each tranche and the size of each tranche depend on the cash flow coming into the deal in a very significant way. The cash flow comes largely from prepayment of the underlying mortgages through refinancing. What happens to the cash coming into the deal depends on triggers which measure (prepayment and default) performance of the underlying pools of subprime mortgages. The triggers can potentially divert cash flows within the structure. In some cases, this can lead to a leakage of protection for higher-rated tranches. Time tranching in subprime transactions is contingent on these triggers. The structure

08 Book.indb 159

2/13/09 3:58:27 PM

160

Gary B.Gorton

makes the degree of credit enhancement dynamic and dependent on the cash flows coming into the deal. In this section, I briefly explain the structural features of subprime bonds. The credit risk of the underlying mortgages is one important factor to understand in assessing the relative value of a particular subprime RMBS. Later, I will focus on the characteristics of the mortgages themselves, but here I focus on the securitization structure. However, the credit risk of the borrowers is intimately linked to the structure of the bond and, indeed, the structure of the particular transaction to which the bond is a part. Chart 3 shows the basic structure of a subprime RMBS transaction.30 Overwhelmingly, asset-backed securities (ABS) and mortgagebacked securities (MBS) use one or both of the following structures: • A senior/subordinate shifting of interest structure (“senior/sub”), sometimes called the “6-pack” structure (because there are 3 mezzanine bonds and 3 subordinate bonds junior to the AAA bonds), or • An excess spread/overcollateralization (“XS/OC”) structure. Overcollateralization means that the collateral balance exceeds the bond balance, that is, deal assets exceed deal liabilities. Because credit risk is the primary risk factor, subprime RMBS bonds have a senior/sub structure, like prime RMBS, but also have an additional layer of support that comes from the excess spread, i.e., the interest paid into the deal from the underlying mortgages minus the spread paid out on the RMBS bonds issued by the deal.31 Another important feature is overcollateralization, that is, there are initially more assets (collateral) than liabilities (bonds). (The overcollateralization reverts to an equity claim if it remains at the end of the transaction.) In a prime deal with a senior/sub structure, basically the total amount of credit enhancement that will ever be present is in place at the start of the deal. The tranche sizes are fixed. In this setting, assuming that defaults and losses are bunched near the start of the deal is conservative. This assumption erodes the credit enhancement early

08 Book.indb 160

2/13/09 3:58:27 PM

The Panic of 2007

161

Chart 3 Sample Subprime MBS Structure Mortgage Pools

Individual Mortgages

REMIC Trust

RMBS Bonds

M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 M14 M15 M16 M17 M18 M19 M20 M21 M22 M23 M24 M25 M26 M27 M28 M29 M30 M31 M32 M33 M34 M35 M36 M37 M38 M39 M40 M41 M42 M43 M44 M45 M46 M47 M48 M49 M50

2/28 Hybrid ARM Mortgage Pool

M51 M52 M53 M54 M55 M56 M57 M58 M59 M60 M61 M62 M63 M64 M65 M66 M67 M68 M69 M70 M71 M72 M73 M74 M75 M76 M77 M78 . . .

‘A’ RMBS

M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 M14 M15 M16 M17 M18 M19 M20

M31 M32 M33 M34 M35 M36 M37 M38 . . . M 1000

Special Purpose Vehicle (RMBS Trust) ‘AA’ RMBS

M 2000

M21 M22 M23 M24 M25 M26 M27 M28 M29 M30

‘AAA’ RMBS

Fixed Rate Mortgage

‘BBB’ RMBS ‘BBB-’ RMBS Residual

Source: Kevin Kendra, Fitch, “Tranche ABX and Basis Risk in Subprime RMBS Structured Portfolios,” February 20, 2007

on, which cannot be replaced. Because of sequential amortization, senior tranches are being paid down over time in this structure. Subprime transactions are different because the XS/OC feature results in a buildup of credit enhancement from the collateral itself during the life of the transaction. The allocation of the credit enhancement over time depends on triggers that reflect the credit condition of the underlying portfolio. Excess spread is built up over time to reach a target level of credit enhancement. Once the OC target is reached, excess spread can be paid out of the transaction (to the residual holder), and is no longer available to cover losses. Later, I discuss the triggers in more detail. There are several key features of RMBS structures to be mentioned. First, there is a lockout period. Mezzanine and subordinate bonds are locked out of receiving prepayments for a period of time after deal settlement. In other words, during the lockout period, amortization is sequential. The period of time of the lockout, and other details, differ depending on the type of collateral in the deal. Second, there may be cross-collateralization. That is, some transactions contain

08 Book.indb 161

2/13/09 3:58:27 PM

162

Gary B.Gorton

multiple loan groups. After interest payments are made on bonds in one group, available remaining funds can be used to pay interest to bonds in another group.32 Chart 4 displays the two types of transaction structures: senior/sub structure and the OC structure. These transactions are quite complicated, so as a prelude to discussing XS/OC structures, I will very briefly start with the typical prime and Alt-A deal structure. I emphasize that what follows is a brief overview only. IV.C. Prime and Alt-A Deals Most prime jumbo and Alt-A transactions use a 6-pack structure, and most subprime, and a few Alt-A deals, use the XS/OC structure. Choice of structure is mostly a function of the amount of excess spread in the deal. Excess spread is the difference between the weighted average coupon on the collateral and the weighted average bond coupons. In an XS/OC structure the excess spread is typically between 300 and 400 basis points. There is no overcollateralization in a 6-pack structure. In a 6-pack deal, the mortgage collateral is tranched into a senior (AAA) tranche, mezzanine tranches (AA, A, BBB), and subordinated tranches (BB, B, and unrated). The most junior bond, essentially equity, is unrated because it is the “first loss” piece, meaning that it will absorb the first dollar of loss on the underlying pool of mortgages. In a senior/sub, or 6-pack, structure, the mezzanine (“mezz”) bonds and subordinate bonds are tranched to be thick enough to absorb collateral losses to ensure that the senior bonds have a probability of loss sufficiently low to justify a triple-A rating. This is accomplished by reversing the order of the priority of cash flow payments and losses in the transaction. In the early years of the transaction, prepaid principal is allocated from top down (“sequential amortization”), that is, only the senior bonds are paid, while the mezz bonds and sub bonds are “locked out” from receiving prepaid principal. Losses are allocated from the bottom up, that is, the lowest-rated class outstanding at the time will absorb any principal losses.

08 Book.indb 162

2/13/09 3:58:27 PM

The Panic of 2007

163

Chart 4 Senior/Sub 6-Pack Structure vs. the XS/OC Structure Deal with 6-pack Structure

Collateral

Deal with XS/OC

AAAs

AA “M1” A “M2” BBB “M3” BB “B1” B “B2” N.R. “B3”

AA “M1” Excess-Spread O/C-based Credit Enhancement

Classic “6-pack” Credit Enhancement

Deal Collateral Face Value—Total Principal Payments

AAAs

A “M2”

BBB “M3”

XS. OC

Interest Payments

IO

Residual Interest on the Bonds

Interest on the Bonds

Note: The scale in Chart 1 does not accurately reflect relative size of bonds, IO or interest flow. Source: UBS

08 Book.indb 163

2/13/09 3:58:28 PM

164

Gary B.Gorton

By using sequential amortization, the senior bonds are paid down first, and there is an increase in the percentage of the remaining collateral that is covered by the mezz and sub bonds. This continues during the lock-out period, which may be the first five years, in a fixed-rate transaction, or for as long as ten years in a prime ARM transaction. In ARM deals, there may be triggers that allow for a reduction in the length of the lock-out period if certain performance metrics are satisfied. The two most common metrics in prime ARM senior/sub structures are (1) a Step-down Test and (2) the Double-down Test. A Step-down Test refers to when prepaid principal switches from sequential pay to pro rata amortization. Typically, prepaid principal switches from sequential pay to pro rata for all outstanding classes if: (a) the senior credit enhancement (CE) is twice the original percentage; and (b) the average 60+-day delinquency percentage for the prior six months is less than 50 percent of the current balance; and (c) cumulative losses are under a specified percentage of the original balance. The Double-down Test means that prior to the initial threeyear period, 50 percent of prepaid principal can be allocated to the mezz and sub bonds if the above three criteria, (a)–(c), are satisfied. IV.D. Subprime Deals XS/OC deals are much more complex than straight senior/sub deals (which I have only briefly described above). As an overview, in contrast to a 6-pack deal in a, say, $600 million XS/OC transaction, the underlying mortgage pool might have collateral worth $612 million, a 2 percent overcollateralization. The $12 million of overcollateralization can be created in either of two ways: (1) it can be accumulated over time using excess spread; or (2) it is part of the deal from the beginning, when the face value of the bonds issued is less than the notional amount of the collateral. XS/OC structures involve the following features (see, e.g., Bear Stearns, September 2006A): • Excess Spread: Like senior/sub deals, the excess spread is used to increase the overcollateralization (OC) by accelerating the payment of principal on senior bonds via sequential amortization;

08 Book.indb 164

2/13/09 3:58:28 PM

The Panic of 2007

165

this process is called “turboing.” Once the OC target has been reached, and subject to certain performance tests, excess spread can be released for other purposes, including payment to the residual holder. • The OC Target: The OC target is set as a percent of the original balance and is designed to be in the second loss position against collateral losses. The interest-only strip (IO) is first. Typically, the initial OC amount is less than 100 percent of the OC target, and it is then increased over time via the excess spread until the target is reached. When the target is reached, the OC is said to be “fully funded.” When the deal is fully funded, NIMs can begin to receive cash flows from the deal. Subject to passing certain performance tests, OC can be released to the residual holder. • Step-down Date: The step-down date in an XS/OC deal is the later of a specified month (e.g., month 36) and the date at which the senior credit enhancement reaches a specified level (e.g., 51 percent). Prior to the step-down date, the senior bonds receive 100 percent of the principal prepayments. When the senior bonds are completely amortized away, prepaid principal continues to sequentially amortize, with the next class being the outstanding mezzanine bonds. • Performance Triggers: Transactions are structured to include performance triggers that, under certain circumstances, will cause a reallocation of principal to protect or increase subordination levels. Generally speaking, there are two types of triggers: delinquency triggers and loss triggers. A trigger is said to “pass” if the collateral does not breach the specified conditions, and to “fail” if those conditions are hit or breached. If a trigger fails, principal payments to the mezzanine and subordinate bonds are delayed or stopped, preventing a reduction or credit enhancement for the senior bonds.33 Loss triggers are target levels of cumulative losses as of specific dates after deal start. For example, the loss trigger in months 1-48 might be 3.5 percent, rise to 5.25 percent in months 49-60, 6.75 percent in months 61-72, and stay flat at 7.75 percent thereafter.

08 Book.indb 165

2/13/09 3:58:28 PM

166

Gary B.Gorton

• Available Funds Cap (AFC): Generally, bonds in XS/OC deals pay a floating coupon. The underlying mortgages typically pay a fixed rate until the reset date on hybrid ARMs. This creates the risk that the interest paid in to the deal from the underlying collateral is not sufficient to make the coupon payments to the deal bondholders–“available funds cap risk.” To prevent this situation, the deal is subject to an AFC. Investors receive interest as the minimum of index (e.g., 1-month LIBOR) plus margin or the weighted average AFC. There are many nuances to these triggers. See, e.g., Moody’s (November 22, 2002; May 30, 2003; September 26, 2006). The structure can be summarized with a series of diagrams due to Fitch (2007). Then I will briefly present a sample transaction. Following that, I will show two other transactions to illustrate the cash flow dynamics and credit enhancement buildup. As shown in Chart 5, principal waterfalls are sequential-pay typically for the first three years. That is, all scheduled principal and prepayments go to repay the senior bondholders first, until they are paid in full. Then, principal payments go to the next senior note holder, until they are paid in full, and so on. As discussed, after the first three years (scenario 1, Chart 6), credit enhancement (CE) “steps down,” if certain performance tests have been met (scenario 2, Chart 6). For example, if overcollateralization (OC) targets have been met, the CE steps down by repaying subordinate bondholders. OC targets are set to double the original subordination. Interest waterfalls involve regular interest that is paid sequentially to bonds, capped at the weighted average mortgage rate net of expenses (net weighted average coupon, WAC) or available funds cap (AFC), as discussed above. “Excess interest” is the remaining interest (which goes into the interest collection account) after paying bondholders regular interest. Excess interest (or “excess spread”) is first used to cover realized collateral losses. Second, excess interest is used to cover any interest shortfalls due to the net WAC being lower than the stated bond coupon. Lastly, the

08 Book.indb 166

2/13/09 3:58:28 PM

08 Book.indb 167

Scheduled Principal & Prepayments

M 1000

‘BBB-’ Residual

Residual Excess Interest

‘BBB’ ‘BBB- ’ L + % or Net WAC

‘BBB ’ L + % or Net WAC

Source: Kevin Kendra, Fitch, “Tranche ABX and Basis Risk in Subprime RMBS Structured Portfolios,” February 20, 2007

M31 M32 M33 M34 M35 M36 M37 M38 . . .

M21 M22 M23 M24 M25 M26 M27 M28 M29 M30

M11 M12 M13 M14 M15 M16 M17 M18 M19 M20

‘A’

‘AAA’

Scheduled Principal & Prepayments

M1 M2 M3 M4 M5 M6 M7 M8 M9 M10

‘AAA’ L + % or Net WAC

Principal Payments

‘AA’

$P

Interest

Interest Payments

‘A’ L + % or Net WAC

Servicer

$I

Accounts

‘AA ’ L + % or Net WAC

$

$

Trust

REMIC

M M71 M72 M73 M74 M75 M76 M77 M78 . . . 2000

M61 M62 M63 M64 M65 M66 M67 M68 M69 M70

M51 M52 M53 M54 M55 M56 M57 M58 M59 M60

M41 M42 M43 M44 M45 M46 M47 M48 M49 M50

M31 M32 M33 M34 M35 M36 M37 M38 M39 M40

M21 M22 M23 M24 M25 M26 M27 M28 M29 M30

M11 M12 M13 M14 M15 M16 M17 M18 M19 M20

M1 M2 M3 M4 M5 M6 M7 M8 M9 M10

Monthly Mortgage Payments

Chart 5 Sample Subprime RMBS Payments

The Panic of 2007 167

2/13/09 3:58:28 PM

08 Book.indb 168

Payments Before Step Down

Scheduled Principal & Prepayments

Payments Before Step Down

Scheduled Principal & Prepayments

Residual

‘BBB-’ After Step Down

‘BBB’

‘A’

‘AA’

‘AAA’

Principal Payments

$P Accounts

Source: Kevin Kendra, Fitch, “Tranche ABX and Basis Risk in Subprime RMBS Structured Portfolios,” February 20, 2007

Residual

‘BBB-’

‘BBB’

‘A’

‘AA’

‘AAA’

Principal Payments

Accounts

$P

Scenario 1: Sequential Principal Repayment

Scenario 2: Performance Test Passed the Credit Enhancement “Steps Down” by Paying Principal to Subordinated Notes

Chart 6 Sample RMBS Interest Waterfall 168 Gary B.Gorton

2/13/09 3:58:28 PM

The Panic of 2007

169

remaining excess interest goes to the holder of the residual bond, typically the originator of the mortgages. (See Chart 7.) The lock-out and step-down provisions are common structural features of subprime deals. To reiterate, the “lock-out” provision locks out the subordinate bonds from receiving principal payments for a period of time. After the lock-out period, deals are allowed to “stepdown,” that is, principal payments can be distributed to the subordinated bonds, provided that the credit enhancement limits are twice the original levels and the deal passes other performance tests, measured by triggers. IV.E. Example of a Subprime RMBS Deal As a typical example of a subprime mortgage securitization, I briefly look at the Structured Asset Investment Loan Trust 2005-6, issued in July 2005. The capital structure of the bond is shown in Table 6.34 Note how much of this deal is rated investment-grade and how much is AAA. The certificates consist of the classes of certificates listed in the table, together with the Class P, Class X and Class R certificates. Only the classes of certificates listed in the table were offered publicly by the prospectus supplement. Note the structure of the transaction. There are four mortgage pools, with only limited cross-collateralization. Principal payments on the senior certificates will depend, for the most part, on collections on the mortgage loans in the related mortgage pool. However, the senior certificates will have the benefit of credit enhancement in the form of overcollateralization and subordination from each mortgage pool. That means that even if the rate of loss mortgage pool related to any class of senior certificates is low, losses in the unrelated mortgage pools may reduce the loss protection for those certificates. Note the thinness of the mezzanine tranches at inception; they are almost digital with respect to defaults, unless the amount of prepayment cash coming into the deal is quite significant in the early life of the transaction. For example, the M9 tranche thickness is only 50 basis points, and yet it is rated BBB-, an investment-grade rating. It is

08 Book.indb 169

2/13/09 3:58:28 PM

08 Book.indb 170 ‘BBB ’ ‘BBB- ’ Residual Losses

‘BBB- ’ L + % or Net WAC Residual Excess Interest

‘A’

‘BBB ’ L + % or Net WAC

‘AA ’

‘A’ L + % or Net WAC

‘AAA ’

Scheduled Principal & Prepayments

Principal Payments

‘AA ’ L + % or Net WAC

‘AAA ’ L + % or Net WAC

Interest Payments

Step 2 – Excess Interest to Cover Collateral Losses

Source: Kevin Kendra, Fitch, “Tranche ABX and Basis Risk in Subprime RMBS Structured Portfolios,” February 20, 2007

Interest

Accounts

$I

Step 1 – Interest Paid Sequentially to Bonds, Capped at AFC

Chart 7 Allocation of Interest

L + % - Net WAC

L + % - Net WAC

Interest Shortfalls

Step 3 – Remaining Excess Interest to Pay AFC Shortfalls

170 Gary B.Gorton

2/13/09 3:58:29 PM

The Panic of 2007

171

Table 6 Structured Asset Investment Loan Trust 2005-6 Capital Structure Class

Related Mortgage Pool(s)

Principal Type

Principal Amount

Tranche Thickness

Moody’s

S&P

Fitch

A1

1

Senior*

455,596,000

20.18%

Aaa

AAA

AAA

A2

1

Senior*

50,622,000

2.24%

Aaa

AAA

AAA

A3

2

Senior

506,116,000

22.42%

Aaa

AAA

AAA

A4

3

Senior, Sequential Pay

96,977,000

4.30%

Aaa

AAA

AAA

A5

3

Senior, Sequential Pay

45,050,000

2.00%

Aaa

AAA

AAA

A6

3

Senior, Sequential Pay

23,226,000

1.03%

Aaa

AAA

AAA

A7

4

Senior, Sequential Pay

432,141,000

19.14%

Aaa

AAA

AAA

A8

4

Senior, Sequential Pay

209,009,000

9.26%

Aaa

AAA

AAA

A9

4

Senior, Sequential Pay

95,235,000

4.22%

Aaa

AAA

AAA

M1

1, 2, 3, 4

Subordinated

68,073,000

3.02%

Aa1

AA+

AA+

M2

1, 2, 3, 4

Subordinated

63,534,000

2.81%

Aa2

AA

AA

M3

1, 2, 3, 4

Subordinated

38,574,000

1.71%

Aa3

AA-

AA-

M4

1, 2, 3, 4

Subordinated

34,036,000

1.51%

A1

A+

A+

M5

1, 2, 3, 4

Subordinated

34,036,000

1.51%

A2

A

A

M6

1, 2, 3, 4

Subordinated

26,094,000

1.16%

A3

A-

A-

M7

1, 2, 3, 4

Subordinated

34,036,000

1.51%

Baa2

BBB

BBB

M8

1, 2, 3, 4

Subordinated

22,691,000

1.01%

Baa3

BBB-

BBB-

M9

1, 2, 3, 4

Subordinated

11,346,000

0.50%

N/R

BBB-

BBB-

M10-A

1, 2, 3, 4

Subordinated

5,673,000

0.25%

N/R

BBB-

BB+

M10-F

1, 2, 3, 4

Subordinated

5,673,000

0.25%

N/R

BBB-

BB+

The Class A1 and Class A2 certificates will receive payments of principal concurrently, on a pro rata basis, unless cumulative realized losses or delinquencies on the mortgage loans exceed specified levels, in which case these classes will be treated as senior, sequential pay classes.

not that this rating is necessarily inaccurate, but that it assumes that the deal’s cash flow mechanics have a reasonable chance of working. Some of the characteristics of the pools are shown in Table 7. The prospectus gives an overview of the triggers for this deal, as follows (italicized terms in original, which means they are defined elsewhere in the document): The manner of allocating payments of principal on the mortgage loans will differ, as described in this prospectus

08 Book.indb 171

2/13/09 3:58:29 PM

172

Gary B.Gorton

Table 7 Summary of the Pools’ Characteristics Pool 1

Pool 2

Pool 3

Pool 4

% First Lien

94.12%

98.88%

100.00%

93.96%

% 2/28 ARMS

59.79%

46.68%

75.42%

37.66%

% 3/27 ARMS

20.82%

19.14%

19.36%

9.96%

% Fixed Rate

13.00%

8.17%

2.16%

11.46%

% Full Doc

59.98%

56.74%

44.05%

35.46%

% Stated Doc

39.99%

37.47%

34.30%

33.17%

% Primary Residence

90.12%

90.12%

80.61%

82.59%

WA FICO

636

615

673

635

supplement, depending upon the occurrence of several different events or triggers: • whether a distribution date occurs before or on or after the “stepdown date,” which is the later of (1) the distribution date in July 2008 and (2) the first distribution date on which the ratio of (a) the total principal balance of the subordinate certificates plus any overcollateralization amount to (b) the total principal balance of the mortgage loans in the trust fund equals or exceeds the percentage specified in this prospectus supplement; • a “cumulative loss trigger event” occurs when cumulative losses on the mortgage loans are higher than certain levels specified in this prospectus supplement; • a “delinquency event” occurs when the rate of delinquencies of the mortgage loans over any three-month period is higher than certain levels set forth in this prospectus supplement; and • in the case of pool 1, a “sequential trigger event” occurs if (a) before the distribution date in July 2008, a cumulative loss trigger event occurs or (b) on or after the distribution date in July 2008, a cumulative loss trigger event or a delinquency event occurs (p. S-7 emphasis in original). This is the structure that was discussed above.

08 Book.indb 172

2/13/09 3:58:29 PM

The Panic of 2007

173

IV.F. How Subprime Bonds Work—Why Does the Detail Matter? In this subsection, I briefly look at two subprime securitization deals; one is a 2005 transaction and the other is a 2006 transaction. The two examples are Ameriquest Mortgage Securities Inc. 2005-R2 (AMSI 2005-R2) and Structured Asset Investment Loan Trust 2006-2 (SAIL 2006-2). The point of the comparison is to show how these two transactions fared; one is 2005 vintage mortgages and the other is 2006 vintage mortgages. The 2006 vintage subprime mortgages have not fared well, as house prices started to turn down, as discussed further below. The examples show how the refinancing or lack of refinancing of the underlying mortgages impacts these securitizations. Both AMSI 2005-R2 and SAIL 2006-2 have the basic structures discussed above, with overcollateralization and various triggers determining the dynamics of credit enhancement. AMSI 2005-R2 consists of three portfolios. Both deals have overcollateralization. Tables 8 and 9 show the structure of each deal, what the deals looked like at inception with respect to tranche sizes and ratings, and then what the tranche sizes and ratings looked like in the first quarter of 2007. The BBB tranches are highlighted. Note the tranche sizes of the BBB tranches, as a percentage of collateral, at inception. They are very thin, almost unbelievably thin. Normally, the rating agencies would not allow such thin tranches, but these tranches are expected to build up as the more senior tranches amortize due to refinancing and sequential amortization. Also, note the subordination percentages for the BBB tranches at inception. For example, the M9 tranche of AMSI 2005-R2 has only 1.1 percent of subordination, unbelievably small. But, again, the dynamics of the transaction mean that this should grow as time passes, amortization occurs, and credit enhancement builds up. These features, the thin tranches and low initial subordination levels, are acceptable if the underlying mortgages refinance as expected. In that case, the deals shrink as amortization occurs. Credit enhancement will build up, and after the step-down date, the BBB tranches will look acceptable. Of course, this depends on house prices.

08 Book.indb 173

2/13/09 3:58:29 PM

08 Book.indb 174

258,089,000

64,523,000

258,048,000

64,511,000

124,645,000

139,369,000

26,352,000

32,263,000

31,200,000

49,800,000

16,800,000

28,800,000

16,800,000

12,000,000

19,200,000

A-1A

A-1B

A-2A

A-2B

A-3A

A-3B

A-3C

A-3D

M1

M2

M3

M4

M5

M6

M7

Publicly Offered Certificates

Size

I,II,III

I,II,III

I,II,III

I,II,III

I,II,III

I,II,III

I,II,III

III

III

III

III

II

II

I

I

Related Mortgage Pool(s)

BBB+/Baa1/ BBB+

A-/A3/A-

A/A2/A

A+/A1/A+

AA-/Aa3/AA-

AA/Aa2/AA

AA+/Aa1/AA+

AAA/Aaa/NR

AAA/Aaa/AAA

AAA/Aaa/AAA

AAA/Aaa/AAA

AAA/Aaa/NR

AAA/Aaa/AAA

AAA/Aaa/NR

AAA/Aaa/AAA

Ratings (Fitch, Moody’s, S&P)

At Issue in 2005

1.6%

1.0%

1.4%

2.4%

1.4%

4.1%

2.6%

2.7%

2.2%

11.6%

10.4%

5.4%

21.5%

5.4%

21.5%

% of Collateral

4.80%

6.40%

7.40%

8.80%

11.20%

12.60%

16.75%

19.35%

19.35%

19.35%

19.35%

19.35%

35.48%

19.35%

35.48%

Subordination

19,200,000

12,000,000

16,800,000

28,800,000

16,800,000

49,800,000

31,200,000

3,994,403

26,352,000

9,597,506

-

10,801,936

43,208,414

7,523,047

30,091,837

Size

B/Baa1/BBB+

BBB/A3/A-

A/A2/A

A+/A1/A+

AA-/Aa3/AA-

AA/Aa2/AA

AA+/Aa1/AA+

AAA/Aaa/AAA

AAA/Aaa/AAA

AAA/Aaa/AAA

PIF/WR/NR*

AAA/Aaa/NA

AAA/Aaa/AAA

AAA/Aaa/NA

AAA/Aaa/AAA

Ratings (April 25, 2008)

Table 8 Ameriquest Mortgage Securities Inc. 2005-R2 (AMSI 2005-R2) 2007Q1

5.3%

3.3%

4.7%

8.0%

4.7%

13.8%

8.6%

1.1%

7.3%

2.7%

0.0%

3.0%

12.0%

2.1%

8.3%

Percent of Collateral

15.21%

20.53%

23.85%

28.50%

36.48%

41.13%

54.92%

63.56%

63.56%

63.56%

63.56%

63.56%

77.62%

89.58%

91.67%

Subordination

174 Gary B.Gorton

2/13/09 3:58:29 PM

08 Book.indb 175

13,200,000

M9

I,II,III

I,II,III

12,000,000

15,600,000

1,200,000,000

1,200,000,147

M11

CE

Total

Collateral

I,II,III

I,II,III

NR/NR/NR

BB/Ba2/BB

BB+/Ba1/BB+

BBB/Baa2/BBB-

BBB/Baa2/BBB

1.3%

1.0%

1.1%

0.7%

Prospectus dated March 22, 2005. AMSI 2005-R2 closed March 24, 2005.

* PIF = tranche “paid-in-full”; WR= “withdrawn rating”; NR= “no rating.”

7,800,000

M10

Not Publicly Offered Certificates

9,000,000

M8

0.00%

1.30%

2.95%

4.05%

361,097,430.00

361,097,331.00

12,928,188

12,000,000

7,800,000

13,200,000

9,000,000

NR/NR/NR

CCC/Ba2/BB

CCC/Ba1/BB+

B/Baa3/BBB-

B/Baa2/BBB

3.6%

3.3%

2.2%

3.7%

2.5%

Table 8 Ameriquest Mortgage Securities Inc. 2005-R2 (AMSI 2005-R2) (continued)

0.00%

3.58%

6.90%

9.06%

12.72%

The Panic of 2007 175

2/13/09 3:58:29 PM

08 Book.indb 176

Aaa/AAA/AAA

11,404,000

10,733,000

M7

M8

A3/A-/A-

15,428,000

15,428,000

M5

M6

A2/A/A

20,124,000

20,124,000

M3

M4

Aa2/AA/AA

Baa3/BBB-/BBB-

Baa2/BBB/BBB

Baa1/BBB+/BBB+

A1/A+/A+

Aa3/AA-/AA-

84,875,000

25,136,000

M1

Aaa/AAA/AAA

Aaa/AAA/AAA

Aaa/AAA/AAA

0.8%

0.9%

1.1%

1.1%

1.5%

1.5%

1.9%

6.3%

8.6%

18.2%

11.2%

45.3%

% of Collateral

At Issue 2006

Ratings+(Moody’s, S&P, Fitch)

M2

244,580,000

114,835,000

A3

A4

607,391,000

150,075,000

A1

A2

Publicly Offered Certificates

Size

1.60%

2.40%

3.25%

4.40%

5.55%

7.05%

8.55%

10.42%

16.75%

16.75%

16.75%

16.75%

Subordination

10,733,000

11,404,000

15,428,000

15,428,000

20,124,000

20,124,000

25,136,000

84,875,000

114,835,000

244,580,000

150,075,000

89,285,238

Size

C/D/C

C/CC/C

C/CC/CC

Ca/CC/CC

Caa3/CC/CC

Caa2/CCC/CCC

B3/CCC/CCC

Ba3/CCC/B

Aaa/A/A

Aaa/AAA/AAA

Aaa/AAA/AAA

Aaa/AAA/AAA

Ratings (April 25, 2008)

1.3%

1.4%

1.9%

1.9%

2.5%

2.5%

3.1%

10.5%

14.2%

30.2%

18.5%

11.0%

% of Collateral

2007Q1

Table 9 Structured Asset Investment Loan Trust 2006-2 (SAIL 2006-2)

1.10%

2.42%

3.83%

5.73%

7.63%

10.12%

12.60%

15.70%

26.16%

26.16%

26.16%

26.16%

Subordination

176 Gary B.Gorton

2/13/09 3:58:29 PM

08 Book.indb 177

Ba1/?/?

Ba2/?/?

0.6% 0.6%

1.05% 0.50%

7,379,000

810,940,982.00

98

1,534,646

C/D/C WR/NR/NR

0.9%

11.0%

0.2%

0.19%

88.99%

0.00%

There are also Class P, Class X, Class LT-R and Class R certificates. The Class X Certificates will be entitled to Monthly Excess Cash flow, if any, remaining after required distributions are made to the Offered Certificates and the Class B1 and Class B2 Certificates and to pay certain expenses of the Trust Fund (including any payments to the Swap Counterparty) and, on and after the Distribution Date in April 2016, to deposit any Final Maturity Reserve Amount in the Final Maturity Reserve Account. The Class P Certificates will solely be entitled to receive all Prepayment Premiums received in respect of the Mortgage Loans and, accordingly, such amounts will not be available for distribution to the holders of the other classes of Certificates or to the Servicers as additional servicing compensation. The Class LT-R and Class R Certificates will represent the remaining interest in the assets of the Trust Fund after the required distributions are made to all other classes of Certificates and will evidence the residual interests in the REMICs.

Prospectus dated September 26, 2005.

6,708,733

1,341,599,733.00

CE

Total

7,379,000

7,379,000

B1

B2

Not Publicly Offered Certificates

Table 9 Structured Asset Investment Loan Trust 2006-2 (SAIL 2006-2) (continued)

The Panic of 2007 177

2/13/09 3:58:29 PM

178

Gary B.Gorton

What happened? Looking at 2007Q1, things are very different for the two deals. AMSI 2005-R2 is, of course, older. By 2007Q1, AMSI 2005-R2 has passed its triggers. Note that the tranche thicknesses, measured as a percentage of collateral, have increased. And, very significantly, note the subordination level percentages have built up. For example, initially M9 had 1.1 percent subordination. In 2007Q1 its subordination percent is 9.06 percent. (Still, Fitch—ever conservative—has downgraded the BBB tranches to B!!) Things are much different for SAIL 2006-2. Being a 2006 deal, it is younger. But, it is also a transaction that occurred during the period where house prices did not rise and refinancing was harder to accomplish. Neither the tranche size nor the subordination has increased significantly. This deal is in trouble, as reflected in the ratings of the mezzanine tranches. There are also Class P, Class X, Class LT-R and Class R certificates. The Class X certificates will be entitled to monthly excess cashflow, if any, remaining after required distributions are made to the offered certificates and the Class B1 and Class B2 certificates and to pay certain expenses of the trust fund (including any payments to the swap counterparty) and, on and after the distribution date in April 2016, to deposit any final maturity reserve amount in the final maturity reserve account. The Class P certificates will solely be entitled to receive all prepayment premiums received in respect of the mortgage loans and, accordingly, such amounts will not be available for distribution to the holders of the other classes of certificates or to the servicers as additional servicing compensation. The Class LT-R and Class R certificates will represent the remaining interest in the assets of the trust fund after the required distributions are made to all other classes of certificates and will evidence the residual interests in the REMICs. Standard securitizations have fixed tranche sizes; that is, tranche thickness does not vary over time. To some extent, excess spread is used to create credit enhancement through reserve fund buildup, but this is not the main credit enhancement. See Gorton and Souleles (2007) for a description of standard securitization.

08 Book.indb 178

2/13/09 3:58:29 PM

The Panic of 2007

179

The above examples of subprime securitization show a very different story. They are not at all like standard securitization transactions. In particular, the difference illustrates how the “option” on house prices implicitly embedded in the subprime mortgages has resulted in very house price-sensitive behavior of the subprime RMBS. Unlike standard securitization transactions, here the tranche thickness and the extent of credit enhancement depend on the cash flow coming into the deal from prepayments on the subprime mortgages via refinancing. This depends on house prices. This point about the link to house prices is dramatically illustrated by these two bonds. The 2005 bond passed its triggers and has achieved the levels of credit enhancement and subordination envisioned by the original structure. It has benefited from the refinancing and prepayments of the underlying mortgages. The 2006 bond has not. In 2006 subprime borrowers had not built up enough equity to refinance. They could not prepay, and the 2006 bond has not been able to pass its triggers. (This does not mean that the 2006 bond would be a bad buy. At fire sale prices it may well be a good buy.) If this was the end of the story, it is not clear whether there would have been a systemic problem when the house price bubble burst. I suspect not, but in any case, it is not the end of the story. V.

Collateralized Debt Obligations (CDOs)

The next link in the chain is collateralized debt obligations (CDOs), SPVs that issue long-dated liabilities in the form of rated tranches in the capital markets and use the proceeds to purchase structured products for assets. In particular, ABS CDOs purchased significant amounts of subprime RMBS bonds. This section proceeds as follows. In subsection A, I start with a very brief description of how cash CDOs work (as opposed to synthetic or hybrid CDOs). In subsection B, I describe the amounts of CDOs issued. Subsection C concerns the question of how much subprime RMBS went into CDOs. Subsection D looks at synthetic subprime risk. Subsection E discusses the issue of the final location of the CDO tranches with subprime risk. This involves a discussion of some off-balance sheet

08 Book.indb 179

2/13/09 3:58:30 PM

180

Gary B.Gorton

vehicles that purchased CDO tranches—another link in the chain. The final subsection, F, summarizes. V.A. The Design of CDOs A cash CDO is an SPV which buys a portfolio of fixed-income assets and finances the purchase of the portfolio via issuing different tranches of risk in the capital markets. These tranches are senior tranches (rated Aaa/AAA), mezzanine tranches (rated Aa/AA to Ba/BB), and equity tranches (unrated). Of particular interest are ABS CDOs, CDOs which have underlying portfolios consisting of ABS, including RMBS and commercial mortgage-backed securities (CMBS). CDO portfolios typically included tranches of subprime and Alt-A deals, sometimes quite significant amounts. The interlinking of subprime mortgages, the subprime RMBS, and the CDOs is portrayed in Chart 8 (due to UBS). To the left of the chart is a representation of the creation of a subprime RMBS deal. Some of the bonds issued in this subprime deal go into ABS CDOs. In particular, as shown on the right-hand side of the chart, RMBS bonds rated AAA, AA, and A form part of a “High Grade” CDO portfolio, so called because the portfolio bonds have these ratings. The BBB bonds from the RMBS deal go into a “Mezz CDO,” so named because its portfolio consists entirely, or almost entirely, of BBB-rated ABS and RMBS tranches. If bonds issued by Mezz CDOs are put into another CDO portfolio, then the new CDO­—now holding Mezz CDO tranches—is called a “CDO squared” or “CDO2.” There are some important features to ABS CDOs that make their design more complicated in ways which play a role later. Perhaps most importantly, many cash ABS CDOs are managed, which means that there is a manager (a firm) that oversees the CDO portfolio. In particular, this manager is allowed to trade­—buy and sell—bonds, to a limited extent (say 10 percent of the notional amount per year) over a limited period of time (say the first three years of the transaction). The putative reason for this is that structured products amortize, so to achieve a longer maturity for the CDO, managers need to be allowed to reinvest. They can take cash that is paid to the CDO

08 Book.indb 180

2/13/09 3:58:30 PM

08 Book.indb 181

LOW

Borrower Down Payment

Borrower Credit

GOOD

BAD

Subprime Mortgage Loans

Source: UBS, “Market Commentary,” December 13, 2007

HIGH

‘A’

1% not in all deals

Other credit support. Excess Spread, Over-collateralization

BB, NR

3%

4%

‘BBB’

11%

‘AA’

3% 2%

‘BBB’ NR

6% 4%

‘BBB’ NR

Other credit support: Excess Spread

Other credit support: Excess Spread

3% ‘A’

6% ‘A’

‘AA’

4%

14%

Junior ‘AAA’

‘AA’

60%

8%

Senior ‘AAA’

CDO2

27%

62%

1%

Junior ‘AAA’

Senior ‘AAA’

Mezz ABS CDO

NR

‘BBB’

‘AAA’ 1%

2%

‘A’ 81%

3%

‘AA’

5%

Junior ‘AAA’ Subprime Mortgage Bonds

88%

Senior ‘AAA’

High Grade ABS CDO

Chart 8 Risk Profile of Subprime Mortgage Loans

The Panic of 2007 181

2/13/09 3:58:30 PM

182

Gary B.Gorton

from amortization and reinvest it, and with limitations, as mentioned, they can sell bonds in the portfolio and buy other bonds. There are restrictions on the portfolio that must be maintained, however. CDO managers typically owned part or all of the CDO equity, so they would benefit from higher yielding assets for a given liability structure. Essentially, think of a managed fund with term financing and some constraints on the manager in terms of trading and the portfolio composition. The restrictions on the portfolio composition would limit structured product asset categories to certain maximum amounts of the portfolio. Other restrictions would include maximums and minimum by rating category, restrictions on weighted average life (WAL), correlation factors, weighted average rating factor (WARF), numbers of obligors, etc.35 Table 10 is a very simplified summary example. Priority of cash flows in CDOs is first of all based on seniority, for allocating losses. Credit enhancement is also provided via other mechanisms such as sequential amortization. Finally, there are also coverage tests and triggers which divert cash flows from subordinate tranches, prevent reinvestment of new assets, and cause amortization to be sequential, if the tests are not met. Two common tests are overcollateralization (OC) tests and interest coverage tests. Roughly speaking, an OC test is the ratio of CDO assets at par to the par value of the A tranche, the most senior tranche (in the Tranche A overCDO Assets at Par

collateralization test): Tranche A Par Amount . The Tranche B OC test CDO Assets at Par

is similar: Tranche A and B Par Amount , and so on. There are also interest coverage tests. For example, the Tranche A Interest Coverage Test is a ratio:

CDO Assets' Coupon , and other interest coverage ratios are Tranche A Coupon

analogous. If coverage tests are not met, cash is diverted, and trading limited, until the tests are passed. For purposes here, I do not need to go into all the details of how CDOs work.

08 Book.indb 182

2/13/09 3:58:31 PM

The Panic of 2007

183

Table 10 Sample ABS CDO Portfolio Criteria Correlation Factor/10-year WARF

23 max / 465 max

Collateral Items rated A3 or better

12.5% min

Collateral Items rated Baa3 or better

95.0% min

Collateral Items rated < Ba3

0.0%

Obligor Concentration Limit

1.5% max

Obligor Concentration of > 1.0% and ≤ 1.5% Number of Obligors Obligations with WALs > 10 years

15 obligors max 93 min 0.0%

Obligations with WALs of > 9.0 and ≤ 10.0 years

5.0% max, must be RMBS/CMBS

Obligations with WALs of > 6.5 and ≤ 10.0 years

25.0% max

Obligations with WALs of > 6.0 and ≤ 10.0 years

57.5% max

Obligations with WALs of > 5.5 and ≤ 10.0 years

70.0% max

Portfolio WAL in Years

5.65 max

CDO Securities

20.0% max

CLO Securities (subset of CDO Securities)

5.0% max

Portfolio restrictions are far from standardized.

Many CDOs are structured to experience an event of default (EOD) when a minimum OC ratio for senior liabilities is not maintained. This means that if the par value of assets falls below the face value of senior liabilities, an EOD occurs, allowing the senior investors (the controlling class) to take control of the CDO. Senior investors may choose to liquidate the assets.36 Also, many CDO transactions that have OC-linked EODs also include ratings-based par haircuts in the calculation of the aggregate outstanding par amount of the underlying assets. As a result, downgrades of underlying collateral assets such as RMBS and ABS CDO tranches trigger EODs. In the EODs that have occurred to date, the CDO has tripped a trigger that is related to the failure to maintain a minimum ratio of OC, namely, the ratio of the par value of assets to the face value of the CDO’s senior obligations. The EODs that have occurred to date have not been due to the failure of the CDO to make payments to noteholders. Rather, the OC-related EOD triggers have been hit because their calculation is affected by certain rating-related par “haircuts.”37

08 Book.indb 183

2/13/09 3:58:31 PM

184

Gary B.Gorton

When an EOD occurs, the senior controlling classes of the CDO are in a position to decide what to do. They may: (1) do nothing, and continue to receive payment of principal and interest; (2) accelerate the maturity date of their notes; (3) liquidate the assets of the CDO and use the proceeds to pay off the notes following the order of priority. Currently, some CDOs are liquidating, but it is not clear what will happen in the remaining cases.38 There is no standardization of triggers across CDOs. Some have sequential cash flow triggers, others do not. Some have OC trigger calculations based on ratings changes; others do not. There is no straightforward template. In fact, each ABS CDO must be separately modeled. The above discussion provides a much abbreviated glimpse at the structure that must be modeled. This will play a role later when I discuss the problems investors face when they attempt a valuation of CDO tranches.39 Why would CDOs buy subprime RMBS bonds? Not surprisingly, it was profitable. With regard to the lower-rated tranches, the BBB tranches of subprime RMBS were difficult to sell. Perhaps this was because they were so thin when first issued (see the above examples), so that at first glance they seemed unreasonable. But, this would not be so obvious if they were purchased by a CDO. By 2005, spreads on subprime BBB tranches appeared to be wider than other structured products with the same rating, creating an incentive to arbitrage the ratings between the ratings on the subprime and on the CDO tranches.40 CDO portfolios increasingly were dominated by subprime, suggesting that the market was pricing this risk inconsistently with the ratings. This was not common knowledge. Also, concerning the higher-rated tranches, CDOs may have been motivated to buy large amounts of structured assets because their AAA tranches would be used as fodder for profitable negative basis trades. This may have increased the appetite of CDOs and of dealer banks underwriting the CDOs. In a negative basis trade, a bank buys the AAA-rated CDO tranche while simultaneously purchasing protection on the tranche under a physically settled CDS. From the bank’s viewpoint, this is the simultaneous purchase and sale of a CDO security, which meant (for a while) that the bank could book

08 Book.indb 184

2/13/09 3:58:31 PM

The Panic of 2007

185

the NPV of the excess yield on the CDO tranche over the protection payment on the CDS. If the CDS spread is less than the bond spread, the basis is negative. Here’s an example. Suppose a bank borrows at LIBOR + 5 and buys a AAA-rated CDO tranche which pays LIBOR + 30. Simultaneously, the bank buys protection (possibly from a monoline insurer) for 15 basis points. So the bank makes 25 bps over LIBOR net on the asset, and they have 15 bps in costs for protection, for a 10 bps profit.41 Note that a negative basis trade swaps the risk of the AAA tranche to a CDS protection writer. Now, the subprime-related risk has been separated from the cash host. Consequently, even if we were able to locate the AAA CDO tranches, this would not be the same as finding out the location of the risk. We do not know the extent of negative basis trades.42 V.B. CDO Issuance Table 11 shows CDO issuance. The first column of the table shows total issuance of CDOs. The next column shows total issuance of structured finance CDOs (also called ABS CDOs); these CDOs have RMBS, CMBS, CMOs, ABS, CDOs, CDS, and other securitized/ structured products as collateral. This is the category of CDO that would include subprime mortgages.43 Structured finance CDOs have consistently been the modal category. Another way to divide CDOs is by their structure. Cash flow CDOs have assets and liabilities that are entirely cash instruments (i.e., physical bonds). Liabilities are paid with the interest and principal payments (cash flows) of the underlying cash collateral. Hybrid CDOs combine the funding structures of cash and synthetic CDOs. Synthetic CDOs sell credit protection via CDS rather than purchase cash assets.44 The liability side is partially synthetic, in which case some protection is purchased on tranches from investors, on the most senior tranches. Mezzanine tranches are not synthetic, but paid-in in cash which is deposited in an SPV and used to collateralize the SPV’s credit swap obligations, namely, potential losses resulting in writedowns of the issued notes. Note that synthetic funded CDOs would

08 Book.indb 185

2/13/09 3:58:31 PM

08 Book.indb 186

98,735.4

271,803.3

2005Q4

2005 Total

66,220.2

65,019.6

108,012.7

124,977.9

138,628.7

2006 Q1

2006 Q2

2006 Q3

89,190.2

65.0%

% of Total

176,639.1

67,224.2

34,517.5

102,167.4

97,260.3

83,790.1

75.9%

206,225.9

71,604.3

44,253.1

49,524.6

52,007.2

46,720.3

71,450.5

2005 Q3

119,531.3

2005 Q2

28,177.1

NA

38,829.9

40, 843.9

157,418.5

2004 Total

NA

36,106.9

49,610.2

47,487.8

2004 Q4

NA

25,786.7

18,807.8

2005 Q1

42,086.6

2004 Q3

NA

NA

Cash Flow and Hybrid

75.9%

42,864.6

2004 Q2

Structured Finance

% of Total

24,982.5

2004 Q1

Total Issuance

14,703.8

24,808.4

24,222.6

23.9%

64,957.4

26,741.1

7,754.1

21,695.9

8,766.3

23.7%

37,237.2

8,657.9

5,329.7

17,074.9

6,174.7

Synthetic Funded

Table 11 Global CDO Issuance ($ millions)

125,945.2

102,564.6

101,153.6

83.7%

227,403.6

71,957.6

49,636.7

62,050.5

43,758.8

93.4%

146,998.5

45,917.8

38,207.7

39,715.5

23,157.5

Arbitrage

12,683.5

22,413.3

6,859.1

16.3%

44,399.7

26,777.8

2,370.5

9,400.0

5,851.4

6.6%

10,419.8

1,569.9

3,878.8

3,146.1

1,825.0

Balance Sheet

186 Gary B.Gorton

2/13/09 3:58:31 PM

08 Book.indb 187

93,663.2

47,508.2

502,978.8

2007 Q3

2007 Q4

2007 Total

2008 Q1

40.4%

4,736.1

52.4%

263,455.9

23,500.1

40,136.8

98,744.1

Source: Securities Industry and Financial Markets Association

% of Total

11,710.1

93,063.6

2007 Q2

% of Total

186,467.6

175,939.4

2007 Q1

101,074.9

314,093.2

56.9%

551,709.6

% of Total

180,090.3

2006 Q4

2006 Total

91.2%

10,673.9

72.1%

362,651.7

31,257.9

56,053.3

135,021.4

140,319.1

75.2%

414,742.9

131,525.1

1.6%

186.0

9.1%

46,230.4

5,202.3

5,198.9

8,403.0

27,426.2

16.1%

89,042.7

25,307.9

89.4%

10,468.4

86.8%

436,102.5

39,593.7

86,331.4

153,385.4

156,792.0

85.6%

472,197.7

142,534.3

Table 11 Global CDO Issuance ($ millions) (continued)

10.6%

1,241.7

13.3%

66,876.3

7,914.5

6,732.2

22,554.0

29,675.6

14.4%

79,511.9

37,556.0

The Panic of 2007 187

2/13/09 3:58:31 PM

188

Gary B.Gorton

be the location of synthetic subprime risk in the form of credit protection written on a subprime index (the ABX index).45 Finally, we can think of categorizing CDOs by the motivation for the transaction. As the name suggests, arbitrage CDOs are motivated by the spread difference between higher yielding assets and the lower yields paid as financing costs. This is often viewed as a rating agencycreated arbitrage. Another motivation is regulatory bank capital relief or risk management. Balance sheet CDOs remove the risk of assets off the balance sheet of the originator, typically synthetically. Looking at the table, the first point to note is that CDO issuance has been significant—and the bulk of it has been CDOs with structured products as collateral. The issuance volume that involves synthetically creating risk is also significant. As noted, the motivation has primarily been arbitrage. It is also notable what data are missing. There is no data on the amount of subprime exposure in CDOs, whether cash or synthetic. This is a glimpse of part of the information problem. To figure out the subprime exposure in a CDO requires a “look through” to the subprime RMBS bonds in the portfolio of the CDO and then looking through those bonds individually to determine what subprime mortgages are associated with each RMBS bond in the portfolio. V.C.

Subprime RMBS Bonds and ABS CDOs

Issuance of ABS CDOs roughly tripled over the period 2005–07, and ABS CDO portfolios became increasingly concentrated in U.S. subprime RMBS. Table 12 shows estimates of the typical collateral composition of high-grade and mezzanine ABS CDOs. As the volumes of origination in the subprime mortgage market increased, subprime RMBS increased, and so did CDO issuance (Table 13). How pervasive is subprime collateral in ABS CDOs? Looking through the CDO portfolios for a sample of CDOs gives a sense of how many real estate-related bonds are in the CDO portfolios. UBS

08 Book.indb 188

2/13/09 3:58:31 PM

The Panic of 2007

189

undertook this exercise for a sample of 420 ABS CDOs. The results are shown in Table 14. The important point of this analysis is that the amount of subprime RMBS bonds in ABS CDOs is very significant. V.D. Synthetic Subprime Risk Subprime risk can be traded via credit derivatives referencing individual subprime cash bonds, or via an index linked to a basket of such bonds. Dealer banks launched the ABX.HE (ABX) index in January 2006. The ABX Index is a credit derivative that references twenty equally-weighted RMBS tranches. There are also indices comprising sub-indices linked to a basket of subprime bonds with specific ratings: AAA, AA, A BBB and BBB-. Each subindex references twenty subprime RMBS bonds with the rating level of the subindex. Every six months the indices are reconstituted based on a pre-identified set of rules. The index is overseen by Markit Partners. The dealers provide Markit Partners with daily and monthly marks.46 For our purposes here, the main point is that subprime risk can be traded synthetically with credit derivatives. Risk cannot be created on net because these are derivatives, but the identities of the longs and shorts are not known as this market is over-the-counter. Table 15 shows approximations of the amount of BBB-rated subprime RMBS issuance over 2004–07 and the exposures of mezzanine CDOs issued in 2005–07 to those vintages of BBB-rated subprime RMBS. Note that the mezzanine CDOs issued in 2005–07 used CDS to take on significantly greater exposure to the 2005 and 2006 vintages of subprime BBB-rated RMBS than were actually issued. This suggests that the demand for exposure to riskier tranches of subprime RMBS exceeded supply by a wide margin. The additional risk exposure was created synthetically. (Though, on net, there is no new risk.) In addition, synthetic CDOs, relying completely on derivatives, became increasingly important. Prior to 2005, the portfolios of ABS CDOs were mainly made up of cash securities. After 2005, CDO managers and underwriters began using CDS referencing individual ABS, creating synthetic exposures. “Synthetic CDOs” are CDOs

08 Book.indb 189

2/13/09 3:58:31 PM

190

Gary B.Gorton

Table 12 Typical Collateral Composition of ABS CDOs (percent) High-Grade ABS CDO

Mezzanine ABS CDO

Subprime RMBS Tranches

50%

77%

Other RMBS Tranches

25

12

CDO Tranches

19

6

Other

6

5

Source: Citigroup, cited by Basel Committee on Banking Supervision (BIS) (April 2008)

Table 13 Subprime-Related CDO Volumes Vintage

Mezz ABS CDOs ($ billions)

High-Grade ABS CDOs ($ billions)

All CDOs ($ billions)

2005

27

50

290

2006

50

100

468

Yr to 9/2007

30

70

330

Source: UBS, “Mortgage Strategist,” November 13, 2007.

Table 14 Residential Mortgage Deals in 420 ABS CDOs Number of Deals by Vintage and Mortgage Loan Type Vintage

Subprime

Alt-A

Seconds

Prime

Total

2003

215

63

7

144

429

2004

371

252

25

188

836

2005

488

452

62

209

1,211

2006

522

487

69

142

1,220

2007

150

113

21

28

312

Total

1,746

1,367

184

711

4,008

Source: UBS, “Mortgage and ABS CDO Losses,” December 13, 2007

08 Book.indb 190

2/13/09 3:58:31 PM

The Panic of 2007

191

Table 15 BBB-Rated Subprime RMBS Issuance and Exposure in Mezzanine ABS CDOs Issued in 2005-2007 to BBB-Rated Subprime RMBS ($ billions) 2004

2005

2006

2007

BBB-rated Subprime RMBS Issuance

12.3

15.8

15.7

6.2

Exposure of Mezzanine ABS CDOs issued in 2005-2007

8.0

25.3

30.3

2.9

Exposure as a Percent of Issuance

65

160

193

48

Source: Federal Reserve calculations, cited by Basel Committee on Banking Supervision (BIS) (April 2008)

with entirely synthetic portfolios; the portfolio of a “hybrid CDO” consists of a mix of cash positions and CDS. CDO managers and underwriters used synthetic exposures to meet the growing investor demand for ABS CDOs and to cater to investors’ preferences to have particular exposures in the portfolio that may not have been available in the cash market. CDO managers and underwriters were able to use CDS to fill out an ABS CDO’s portfolio when cash ABS, particularly mezzanine ABS CDO tranches, were difficult to obtain. So far, the subprime mortgages have been securitized and tranches of these securitizations have been sold, in large part, to CDOs, and tranches of the CDOs have been sold to investors. Additional subprime risk has been traded via derivatives. I now turn to the question of the identity of the investors in these risks. Who were these investors? Where did the risk go? V.E. Where Did the CDO Tranches Go? The short answer is that we do not know for sure. Investors around the world purchased rated tranches of CDOs. Lehman Brothers has estimates of the location of the AAA-rated CDO tranches (see Chart 9). Investors in the AAA CDO tranche risk (synthetic, if not cash) include bond insurers, insurance companies, and other categories of institutional investors. The category labeled “ABCP/SIV” refers to Asset Backed Commercial Paper Conduits (ABCPs) and structured investment vehicles (SIVs), which I discuss briefly below.

08 Book.indb 191

2/13/09 3:58:31 PM

192

Gary B.Gorton

Chart 9 Estimated Holdings of AAA CDO Tranches 30%

30%

25%

25%

20%

20%

15%

15%

10%

10%

5%

5%

0%

0% Bond Insurers

Insurance Cos.

CDO CP Put Providers

ABCP/SIV

Others

Source: Lehman Brothers estimates, as of November 13, 2007, based on the 10-Qs of AMBAC, MBIA, ACA, XLCA, FGIC, and rating agency reports on bond insurers

The remaining category, “CDO CP Put Providers,” refers to structures which transform long-dated CDO tranche paper into money market mutual fund eligible investments. This is accomplished by shortening the maturity of the CDO tranche via a liquidity put provider, sometimes called a 2A-7 put, after the part of the Investment Company Act that restricts money market funds to instruments that are 365 days or less in maturity.47 Longer–term bonds are shortened by attaching a put option or tender feature allowing or requiring the investor to sell the security to the put provider, with a stated notice period. Rule 2A-7 allows the money market fund to treat the put notice period as being the maturity of the bond. Note that in the crisis, money market funds exercised their puts, forcing put writers to buy the notes, putting further stress on their liquid resources. One significant category of investors, shown Chart 9, consists of certain kinds of off-balance sheet vehicles, known as structured investment vehicles (SIVs), ABCPs, and SIV-lites. The nuances of the differences between these vehicles do not concern us here (see Moody’s, February 3, 2003; Moody’s, January 25, 2002; Standard

08 Book.indb 192

2/13/09 3:58:32 PM

The Panic of 2007

193

and Poor’s, September 4, 2003). I provide the briefest of overviews to highlight one structural feature that is important. An SIV is a limited-purpose operating company that undertakes arbitrage activities by purchasing mostly highly rated medium- and long-term fixed-income assets and funding itself with cheaper, mostly short-term, highly rated CP and MTNs. An SIV is a leveraged investment company that raises capital by issuing capital market securities (capital notes and medium-term notes) as well as ABCP. ABCP typically comprises around 20 percent of the total liabilities for the biggest SIVs.48 A variant of an SIV is a so-called SIV-lite. SIV-lites share some similarities with collateralized debt obligations (CDOs) in that they are closed-end investments. SIV-lites issue a greater proportion of their liabilities as ABCP than SIVs (around 80 percent–90 percent), are typically more highly leveraged, and seem to have invested almost exclusively in U.S. RMBS. As a consequence, several SIV-lites have restructured their liabilities following the recent turmoil in U.S. mortgage markets. Appendix B lists the larger SIVs and their outcomes. Unlike conduits that issue only ABCP, SIVs and SIV-lites tend not to have committed liquidity lines from banks that cover 100 percent of their ABCP. Rather, they use capital and liquidity models, approved by ratings agencies, to manage liquidity risk. The lack of a full commercial bank guarantee has reportedly led to discrimination against SIV paper by ABCP investors. The important point is that these vehicles are very different from the SPVs used in securitization. Standard securitization SPVs are not managed; they are robot companies that are not marked-to-market; they simply follow a set of prespecified rules. See Gorton and Souleles (2007). Unlike securitization vehicles, these are managed and they are market value vehicles. They raise funds by issuing commercial paper and medium-term notes, and they use the proceeds to buy high-grade assets to form diversified portfolios. They borrow short and purchase long assets. They are required by rating agencies to mark portfolios to market on a frequent basis (daily or weekly), and based on the marks, they are allowed to lever more or required to delever. On SIVs, see Moody’s (January 25, 2002), and on ABCPs, see Moody’s (February 3, 2003).

08 Book.indb 193

2/13/09 3:58:32 PM

194

Gary B.Gorton

Money market mutual funds apparently not only purchased various structured assets, via liquidity (or 2A–7) puts (as discussed above), but also sometimes invested in SIVs. Later, these money market mutual funds had to be bailed out by their sponsors to keep them from “breaking the buck.” See the chronology in Appendix A. V.F. Summary Investors purchased CDO tranches based on ratings, portfolio criteria, and the identity of the CDO manager. Purchasers of CDO bonds receive trustee reports detailing the portfolio of the CDO, which changes over time as the manager trades. CDOs are not market value structures.49 It is literally not possible for a buyer of a CDO tranche to do the double look-through to determine, say, the extent of subprime exposure. That would require looking through each of the bonds in the CDO portfolio, and if the CDO owns other CDO tranches, looking through those as well. Imagine also an investor in an SIV. The SIV has a portfolio of structured assets, which may include CDO tranches. The investor cannot answer the question: Is my SIV investment sensitive to 2006 subprime mortgages? VI.

Complexity, the Loss of Information, and the Current Crisis

Now we come to the first information issue. What is the loss of information? The information problem is that the location and extent of the (2006 and 2007 Q1-2 vintage) subprime risk is unknown to anyone. It is very hard to determine the location of the risk, partly because of the chain of interlinked securities, which does not allow the final resting place of the risk to be determined. But also, because of derivatives, it is even harder: Negative basis trades moved CDO risk, and credit derivatives created additional long exposure to subprime mortgages. Determining the extent of the risk is also difficult because the effects on expected losses depend on house prices as the first–order risk factor. Simulating the effects of that through the chain of interlinked securities is basically impossible. In this section I start by illustrating this last point with a very simple description of the payoffs to the interlinked securities. I then discuss the implications.

08 Book.indb 194

2/13/09 3:58:32 PM

The Panic of 2007

195

VI.A. A Simple, Stylized Example of the Interlinking of Security Designs As before, I will give an extremely simplified example to (hopefully) convey the essence of the complexity problem and the loss of information. I will ignore the dynamic aspects of subprime RMBS transactions. I will consider extremely simple tranching: a subordinated (or, synonymously, junior or first loss) tranche (called the “sub” tranche) and a senior tranche. The subprime RMBS deal will securitize a single subprime mortgage. There are three financial instruments: (1) a subprime mortgage; (2) a senior/sub tranche RMBS securitization of the single subprime mortgage; (3) a senior/sub tranche CDO, which has purchased the senior tranche of the RMBS. I omit a fourth step, of an SIV buying the senior CDO tranche or a CDO tranche having a 2A-7 put attached, and so on. The transactions all last for one period and all payoffs are at the end of the period. I will ignore discounting. The mortgage has a face value of 100. At the end of the period, the mortgage has a step-up rate and will be refinanced, or not. If it is not refinanced, then it defaults, in which case the lender will recover $R. So, the loss is 100–R ≡ Loss if there is a default. If it is refinanced, then the new mortgage is worth M (in expected value), to the lender. Ignoring, for a moment, the dependence of R and M on home prices, the payoff to the lender at the end of the period on the current mortgage is: Max(R, M). If the new mortgage is worth less than the recovery value of the home, then the lender does not refinance (nor will any other lender), and the homeowner defaults. The lender finances the mortgage by securitizing it. It is sold at par of 100. The lender retains the refinancing option as discussed above, and the securitization will either receive par or R at the end of the period. The subprime RMBS transaction has two tranches: The first tranche attaches at 0 and detaches at $N; the second tranche attaches at $N (e.g., N=30 means that the first $30 of loss are absorbed by the sub piece) and goes to the end, 100. The par value of the senior

08 Book.indb 195

2/13/09 3:58:32 PM

196

Gary B.Gorton

tranche is, therefore, 100-N. In other words, the first $N of loss will be borne by the sub piece. Looking at the senior tranche, the loss on this tranche at the end of the period, LS, is given by:

LS = Max[Loss – N, 0].

The payoff or redeemed amount, V, on this senior tranche at the end of the period is:

 Max[100 − N , 0 ], V = Min  .  100 − N − LS

Since 100 – N is always greater than 0, Max[100-N, 0] is always equal to 100 – N. L­S is always greater than or equal to 0, so 100 – N – LS is always less than or equal to 100 – N. Therefore, V = 100 – N – LS . Substituting in for LS: V = Min[100 – N, 100 – Loss]. So, for example, if Loss=50 and N=30, then if the mortgage is not refinanced and defaults, then the senior tranche will have a $20 loss since the first loss tranche only absorbs the first $30 of loss. The final value, V, of the senior tranche is $50. The senior tranche of the subprime RMBS is sold to a CDO, which has two tranches: the first tranche attaches at 0 and detaches at $NCDO; the second tranche attaches at $NCDO and goes to the end, 100-N. Note that the size of the CDO is 100-N (=70 in the example), since it only purchases the senior tranche of the subprime RMBS. Note that NCDO will be less (in dollars) than N because the CDO portfolio is smaller; the sub tranche of the CDO may be larger in percentage terms though. Looking at the senior tranche, the loss on this tranche at the end of the period, LCDO, is given by:

LCDO=Max[Min(LS, 100-N) – NCDO, 0].

At the end of the period, the payoff on the senior tranche of the CDO, VCDO, is given by:

08 Book.indb 196

2/13/09 3:58:32 PM

The Panic of 2007



197

 Max[(100 − N − N CDO ), 0 ], VCDO = Min  .  100 − N − N CDO − LCDO

Substituting for LCDO:

 Max[(100 − N − N CDO ), 0 ], VCDO = Min   N − N CDO − Max[ Min( LS ,100 − N ) − N CDO

,

and substituting now for LS:  Max[(100 − N − N CDO ), 0 ] VCDO = Min  N − N − Max [ Min ( M a x[ Min(100 − Loss,100 ) − N , 0 ],100 − N ) − N CDO , 0 ].  CDO

Looking at this final expression, we can see the dependence of the senior CDO tranche on the structure of the securitization, i.e., the tranching (N), and on the underlying single subprime mortgage, namely, its loss, Loss. And keep in mind that Loss depends on house price appreciation. Nowhere does M appear, because if the loan is refinanced at the end of the period, then it is paid off and there are no losses. M is the expected value of the new loan. In the simple formulation above, the dependence on house prices only appears in terms of the recovery value of the house if there is a default. In the real structure, the refinancing results in M being paid into the securitization which is cash that would be allocated following the priority rules and the triggers, which determine the amortization. So, that aspect is lost in the simplified example. Here’s a very simple numerical version of the example. Assume that the subprime mortgage par amount is 100; assume the size of RMBS sub tranche is $20, so, the size of the senior RMBS tranche is $80. The senior RMBS tranche is sold to a CDO, which only buys this tranche, so, the size of the CDO is $80. The size of CDO sub tranche is $15 and so the senior tranche size is $65. I maintain these parameters and vary the recovery amount in Table 16. The table shows the loss the senior RMBS tranche, the payoff to the senior RMBS tranche, the loss on the senior CDO tranche, and the payoff on the senior CDO tranche—all at the end of the period.50 The example is not realistic because it is too simplified, but it does convey the intuition for a few points. What does the example show?

08 Book.indb 197

2/13/09 3:58:33 PM

198

Gary B.Gorton

Table 16 Parameters Recovery Amount ($)

90

70

Loss on Senior RMBS Tranche (LS)

0

10

Payoff on Senior RMBS Tranche (V)

80

Payoff on Sr. Tranche as % of par

60

50

40

30

20

10

20

30

40

50

60

70

70

60

50

40

30

20

10

100%

87.5%

75%

62.5%

50%

37.5%

25%

12.5%

Loss on Senior CDO Tranche (LCDO)

0

0

5

15

25

35

45

55

Payoff on Senior CDO Tranche (VCDO)

65

65

60

50

40

30

20

10

Payoff on Sr. Tranche as % of par

100%

100%

92.3%

76.9%

61.5%

46.2%

30.8%

15.4%

Outcomes

First, the effects of tranching are apparent. The sub tranche of the RMBS absorbs the first loss. Since the “inner” RMBS tranche (i.e., the one in the CDO) is a senior tranche, the losses on the senior CDO tranche are always less than (or, in the extreme, equal to) the losses on senior RMBS tranche.51 However, conversely, if the CDO had purchased a mezzanine tranche, say going from 10 to 20, then the example would be very different. A senior CDO tranche could easily be at risk of loss if the portfolio consisted of mezzanine RMBS tranches. Obviously, the example could be extended to include an SIV which purchases the senior tranche of the CDO. VI.B. Discussion Valuation of VCDO requires integrating the above expression over a distribution of house prices. There are two practical problems with this. First, as a practical matter, the dependence on house prices creates a practical valuation problem—even if one takes a stand on the distribution of house prices. Imagine, for example, that the subprime securitization has four portfolios, each with thousands of mortgages, as in the above examples. The CDO has purchased 100 tranches of different securitizations, including, say, twenty senior subprime tranches from different deals. In principle, the issue is how to evaluate the senior CDO tranche (even ignoring all the OC tests and other complications

08 Book.indb 198

2/13/09 3:58:33 PM

The Panic of 2007

199

of the CDO structure). Not only is that valuation very difficult to do, but even linking the three structures together in a meaningful way is nigh impossible. An investor who actually purchased a particular CDO tranche or a particular subprime RMBS tranche would receive trustee reports and would, therefore, know the underlying portfolio.52 The subprime RMBS investor could, with some difficulty, look through to the underlying mortgages and try to determine the value of his tranche.53 The computational complexity is very high. The second problem is taking all of the structure into account. There are vendor-provided packages that model the structure of structured products, but the valuation is based on (point estimate) assumptions that are input by the user, rather than simulation of the performance of the underlying portfolios. VII.

The Panic

A bank is…a manufacturer of credit. The cornerstone of credit is confidence—confidence of men in men. A panic is a collapse of credit. It is an intensely human affair, and many of the determining influences are of a personal and confidential character, and very inadequately reflected in the cold figures of the bank statement.

— E.W. Kemmerer (1911)

Like Tolstoy’s family, economic good times are all alike, but every crisis is bad in its own way.54 What triggered the Panic of 2007? How did it develop? The Appendix contains a brief chronology of the events of the Panic. I argued above that a complex chain of securities, derivatives, and SPVs resulted in asymmetric information and a loss of information: The structurers understood the chain, but investors did not. But, valuation is difficult for all parties. The chain began to unravel when house prices did not rise and foreclosures began. In this section I begin by briefly documenting these developments. House price declines and foreclosures do not explain the Panic. I argue that the information story is more complicated. Dealer banks had the information about the subprime-related structures, and about the placement of the various bonds. But, there was no way to learn the consensus value of these bonds and structures. There was

08 Book.indb 199

2/13/09 3:58:33 PM

200

Gary B.Gorton

no mechanism for the revelation and aggregation of diverse information about the effects of the house price decline and the foreclosures. This created a pivotal role for the ABX index, which started trading in early 2006 around the time that house prices began to fall. I review the role of this index in creating common knowledge that the situation of subprime borrowers was deteriorating quickly and that the value of subprime-related bonds and structures was going down. By 2007 the ABX indices had become the focal point of the crisis. I discuss the role of the ABX index in revealing information. This is followed by a brief discussion of the runs on SIVs—the Panic itself. Finally, I try to summarize the information argument of the paper. VII.A. House Prices Do Not Rise House prices were supposed to always go up. Between 2001 and 2005 homeowners enjoyed an average increase of 54.4 percent in the value of their houses, as measured by the Office of Federal Housing Enterprise Oversight (OFHEO).55 In terms of the two-year fixed-rate part of a 2/28 subprime mortgage, from January 1997 to July 2007 every rolling two-year period showed positive house price appreciation, according to the S&P/Case-Shiller (U.S. National) Index. In fact, from March 1998 to March 2007, every rolling two-year period displayed double digit house price appreciation. There was no appreciation or depreciation in August 2007, and starting in September 2007 house price appreciation has been negative. Chart 10 shows a plot of the lagging two-year house price appreciation. But, then house prices declined. In fact, the S&P/Case-Shiller (U.S. National) quarterly home price index declined by 4.5 percent in Q3 2007 versus Q3 2006—the largest drop since the index started recording data in 1988.56 Home prices, as measured in the 20 U.S. metropolitan areas, declined by 4.9 percent, the largest drop since the index was started in 2001, with 15 of the 20 cities showing year-onyear declines in prices. The two largest declines occurred in Tampa (-11.12 percent Y-o-Y) and Miami (-9.96 percent Y-o-Y). U.S. home prices declined 6.7 percent in October from a year earlier, a record drop for the ten-city S&P/Case-Shiller index (Chart 10).57

08 Book.indb 200

2/13/09 3:58:33 PM

The Panic of 2007

201

Chart 10 Lagging Two-Year House Price Appreciation (%) 50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

-10%

-20%

08 20

07 20

06 20

05 20

04 20

03 20

02 20

01 20

00 20

99 19

98 19

97

0% 19

0%

-10%

-20%

Source: S&P

The ability of subprime and Alt-A borrowers to sustain their mortgage payments depends heavily on house price appreciation because of the need for refinancing. When house prices did not appreciate to the same extent as in the past, and in many areas they have recently gone down, the ability of borrowers to refinance has been reduced. In fact, now because of the crisis, underwriting standards have become much tougher, and many lenders are in bankruptcy, meaning that the mortgage market for these borrowers to refinance has effectively closed. Currently, almost all the major issuers of subprime mortgages are either out of business or have stopped making subprime loans unless they conform to government sponsored enterprise (GSE) underwriting criteria. Problems in the Alt-A market are still mostly in the future, and it is likely that this market will also shut down. The unwillingness to originate subprime mortgages is significantly driven by the impossibility of a securitization take-out of the loans. This shutdown means that borrowers in the subprime and Alt-A mortgages will have a very difficult time refinancing when their hybrid ARMs are reset. The shutdown of the subprime mortgage market is very important because of the number of borrowers who will soon reach their reset date, that is, the date at which the initial fixed teaser rate ends and the

08 Book.indb 201

2/13/09 3:58:34 PM

202

Gary B.Gorton

mortgage rate resets to a significantly higher floating rate. Evidence of the shutdown in the refinancing market comes from remittance data. Remittance data shows that the shutdown is dramatically reducing subprime prepayment speeds.58 A decline in prepayment speed means that borrowers cannot refinance either because they no longer can find a lender or because they have no equity built up on their houses. Delinquencies and foreclosures are the result. (Table 17.) We now turn to the issue of how the information about house prices and delinquencies and foreclosures was linked to valuations of the various parts of the chain. Keep in mind that house price and mortgage performance information arrives with a lag, not in real time. VII.B. Information and Common Knowledge It was widely understood that the structures along the chain were sensitive to house prices, that house prices were likely a “bubble.” Not everyone had the same view on whether house prices would continue to rise, or if they were to stop rising, on when this would occur. Or what the effects would be. Different parties made different bets on this. But, they did this without knowing the views of other participants. That is, there was a lack of common knowledge about the effects and timing of house price changes and about the appearance of increases in delinquencies. This explains why the interlinked chain of securities, structures, and derivatives did not unravel for a while. In an important way, this changed with the introduction of the ABX indices at the start of 2006. As mentioned earlier, the ABX Index is a credit derivative that references twenty equally weighted RMBS tranches. There are also indices that comprise subindices: AAA, AA, A, BBB, and BBB-. Each subindex includes twenty subprime home equity bonds. The reference obligations in each subindex comprise bonds at the rating level of the subindex. Every six months the indices are reconstituted based on a pre-identified set of rules. The ABX.HE indices that reference lower-rated RMBS tranches typically carry higher coupons than those referencing higher-rated tranches due to the higher expected likelihood of default. (Table 18.)

08 Book.indb 202

2/13/09 3:58:34 PM

The Panic of 2007

203

Table 17 Delinquency Rates (%) Home Mortgage Delinquency Rate: Total (%)

Delinquency Rate: Prime Borrowers (%)

Delinquency Rate: Subprime Borrowers (%)

2003Q1

4.92

2.62

13.04

2003Q2

4.97

2.60

12.35

2003Q3

4.65

2.44

11.74

2003Q4

4.49

2.37

11.53

2004Q1

4.46

2.26

11.66

2004Q2

4.56

2.40

10.47

2004Q3

4.54

2.32

10.74

2004Q4

4.38

2.22

10.33

2005Q1

4.31

2.17

10.62

2005Q2

4.34

2.20

10.33

2005Q3

4.44

2.34

10.76

2005Q4

4.70

2.47

11.63

2006Q1

4.41

2.25

11.50

2006Q2

4.39

2.29

11.70

2006Q3

4.67

2.44

12.56

2006Q4

4.95

2.57

13.33

2007Q1

4.84

2.58

13.77

2007Q2

5.12

2.73

14.82

2007Q3

5.59

3.12

16.31

2007Q4

5.82

3.24

17.31

Source: Mortgage Bankers Association

Chart 11 portrays the creation of a vintage of the ABX Index and the subindices for the different ratings: AAA, AA, A, BBB, and BBB-. Each subindex includes twenty subprime home equity bonds. The introduction of these indices is important for two reasons. First, they provided a transparent price of subprime risk, albeit with liquidity problems (see Gorton, 2008). Second, it allowed for efficiently shorting of the subprime market. In addition to outright shorting, parties with long positions could hedge. The common knowledge problem concerning the value of subprime bonds may have been solved, but not the location problem. This is, of course, conjecture.60 As with CDS generally, entering into an ABX index contract is analogous to buying or selling insurance on basket of the underlying

08 Book.indb 203

2/13/09 3:58:34 PM

204

Gary B.Gorton

Table 18 ABX.HE Index Overview ABX.HE Portfolio

20 deals in basket, with a new ABX.HE series expected to be launched approximately every 6 months

Credit Score

Each deal must have a maximum average FICO equal to 660

Age

Each tranche must have settled within 6 months of the roll date

Weighting

Reference obligations equally weighted by initial par amount, with subsequent weightings evolving as a function of prepayment and credit experience of underlying transactions

Lien Type

The pool must consist of at least 90% first lien loans

Diversification

–Limits same originator to 4 deals –Limits master servicer to 6 deals

Minimum Deal Size

$500mm

Average Life

Each tranche must have a weighted average life of 4-6 years as of the issuance date (except AAAs which must be greater than 5 years)

Credit Events

Failure to Pay Principal, Write-down

Settlement

Pay-as-you-go (PAUG)59

Chart 11 ABX.HE Indices RMBS 1

RMBS 2

RMBS 3

RMBS 4

RMBS 5

RMBS 6

RMBS 7

RMBS 8

RMBS 9

RMBS 10

RMBS 11

...

RMBS 20

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

‘AAA’ RMBS

...

‘AAA’ RMBS

ABX.HE.AAA

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

‘AA’ RMBS

...

‘AA’ RMBS

ABX.HE.AA

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

‘A’ RMBS

...

‘A’ RMBS

ABX.HE.A

‘BBB’ RMBS

‘BBB’ RMBS

‘BBB’ RMBS

‘BBB’ RMBS

‘BBB’ RMBS

‘BBB’ RMBS

‘BBB’ RMBS

‘BBB’ RMBS

‘BBB’ RMBS

...

‘BBB’ RMBS

ABX.HE.BBB

‘BBB-’ RMBS

‘BBB-’ RMBS

‘BBB-’ RMBS

...

‘BBB-’ RMBS

ABX.HE.BBB-

Residual Residual Residual

...

Residual

‘BBB’ RMBS ‘BBB-’ RMBS

‘BBB’ RMBS ‘BBB-’ RMBS

Residual Residual

‘BBB-’ RMBS

‘BBB-’ RMBS

‘BBB-’ RMBS

‘BBB-’ RMBS

Residual Residual Residual Residual

‘BBB-’ RMBS

‘BBB-’ RMBS

Residual Residual

Source: Kevin Kendra, Fitch, “Tranche ABX and Basis Risk in Subprime RMBS Structured Portfolios,” February 20, 2007

08 Book.indb 204

2/13/09 3:58:34 PM

The Panic of 2007

205

RMBS tranches. An investor wanting to hedge an existing position, or otherwise establish a short credit position using the index (known as the “protection buyer”), is required to pay a monthly coupon to the other party (the “protection seller”). The payment is calculated based on the outstanding notional amount of the index and the fixed coupon. In exchange for the payment, the protection buyer in an ABX index contract is compensated by the protection seller when any interest or principal shortfalls or write-downs on the underlying mortgages affect the constituent RMBS. Unlike with conventional “single name” CDS, the index contract does not terminate when these credit events occur; rather it continues with a reduced notional amount until maturity. If credit events are subsequently reversed —for example, a principal shortfall is made up—then the protection buyer reimburses the protection seller. The ABX tranche coupon is determined on the initiation date. Subsequently, trades require an upfront exchange of premium/discount. In a typical transaction, a protection buyer pays the protection seller a fixed coupon at a monthly rate on an amount determined by the buyer. When a credit event occurs, the protection seller makes a payment to the protection buyer in an amount equal to the loss. Credit events include the shortfall of interest or principal as well as the write-down of the tranche due to losses on the underlying mortgage loans. The initial coupon is determined at the launch of each ABX.HE index based on an average quote from a survey of the market makers, the dealer banks. Knowledge about the structure of the subprime RMBS, CDOs, and off-balance sheet vehicles is held by the dealer banks, who structure these transactions. They are the ones polled to determine the initial coupons on the ABX indices. The polling process works as follows: At or about 9:00 a.m. on the Business Day immediately prior to the Roll Date (the “Fixed Rate Determination Date”), the fixed rate for each sub-index for the new ABX. HE Index for purposes of the ABX Transactions Standard Terms Supplement will be determined by the Administrator by soliciting each ABX.HE Participant to submit an

08 Book.indb 205

2/13/09 3:58:34 PM

206

Gary B.Gorton

average mid-market spread for each sub-index (in increments of 1 basis point). The Administrator will re-solicit ABX.HE Participants until at least two-thirds of the ABX.HE Participants (rounded down) have submitted such spreads. The Administrator shall rank such submissions for each sub-index from lowest to highest spread and discard the top and bottom quartiles thereof (the number of submissions q in each discarded quartile will be given by q=int(NS/4) where NS is the total number of submissions). The fixed rate for each sub-index shall be the lesser of (i) average of the remaining submissions for such sub-index (rounded up to the nearest basis point), as determined by the Administrator and (ii) 500 basis points.

—Markit, ABX Index Rules

The ABX.HE 06-1 (this is the official name for the 2006 first vintage) began trading on January 19, 2006. So, unfortunately, there are no observations on early index subprime product, such as the 2005 vintage. Moreover, the company administering the ABX, Markit, announced that the roll of the new ABX.HE, ABX.HE 08-1, would be postponed for three months from the date it was scheduled to launch, January 19, 2008. Markit said that: “The decision to postpone its launch was taken following extensive consultation with the dealer community. It follows a lack of RMBS deals issued in the second half of 2007 and eligible for inclusion in the forthcoming Markit ABX. HE roll. The Markit ABX.HE 07-2 remains the on-the-run series until further notice.” See http://www.markit.com/information/products/abx/contentParagraphs/04/document/20071219%20Markit%20 ABX.HE.pdf. No subsequent vintage has been issued. Chart 12 shows the prices of the 2006-1, 2006-2, 2007-1, and 2007-2 vintages of the index for the BBB- tranche. These are the only vintages available. In three of the four cases, the index starts trading at par of 100. In the case of the 2007-2 index, it opened at a price significantly below par.61 The time pattern of prices in this chart is very interesting. The first vintage ABX 2006-01 trades near par, as does the 2006-02 vintage

08 Book.indb 206

2/13/09 3:58:34 PM

The Panic of 2007

207

Chart 12 ABX BBB– Prices 120

100 40Y ABX.HE.BBB- 06-1 80

40Y ABX.HE.BBB- 06-2

30Y ABX.HE.BBB- 07-1

60

30Y ABX.HE.BBB- 07-2 40

20

19

-O

ct-

07

7 l-0 -Ju 19

19

-A pr-

07

7 n-0 -Ja 19

19

-O

ct-

06

6 l-0 -Ju 19

06 -A pr19

19

-Ja

n-0

6

0

initially. During 2006, there is little evidence of a major crisis. But, the 2007-01 BBB- ABX nosedives upon issuance, and the 2007-02 vintage opens trading below 60. The dealers got the coupons badly wrong. One interpretation of this is that the fundamentals of subprime were weakening during 2006, as the ABX drifted down somewhat in the second half of 2006. But, starting in 2007 it seems clear that there were major problems. I view the ABX indices as revealing hitherto unknown information, namely, the aggregated view that subprime was worth significantly less. In fact, some of the dealer banks themselves, we now know, were shorting the index to hedge their long positions—of course so was everyone else.62 The ABX indices also allow all parties, e.g., hedge funds, to express their views on the value of subprime RMBS bonds. Kiet Tran (no date) of Markit put it this way: The sub-prime debacle in the U.S. brought about a global credit crunch this summer with the ABX leading the charge. Subordinate tranches of the 06-2 and 07-1 series have lost over 75% of their value since the end of May. Even with the Fed rate cuts, the ABX free fall continues, particularly for

08 Book.indb 207

2/13/09 3:58:35 PM

208

Gary B.Gorton

the lower rated tranches. Early signals were seen in February 2007, a month where prices of the ABX BBB- tranches plunged more than 20%. Shareholder values of sub-prime mortgage lenders deteriorated in the following weeks, with the stock price of Accredited Home Lenders Holding Corporation dropping just over 80% between February month-end and the mid-March low. ABX.HE acts as a vehicle for investors to hedge their subprime exposure and to express their views on the sub-prime market using a liquid and transparent instrument. The recent performance of the ABX does not bode well for the outlook for the sub-prime mortgage market but time will tell how far losses will extend. For the time being, the ABX.HE index is the acting weatherman of the sub-prime mortgage market, predicting a rough storm ahead. It is not clear whether the housing price bubble was burst by the ability to short the subprime housing market or whether house prices were going down and the implications of this were aggregated and revealed by the ABX indices. As discussed below, the indices were the sole source of information for marking-to-market. It seems that the indices played a central informational role. VII.C. The Run on the SIVs The runs began on ABCP conduits and SIVs. These vehicles were funded with short maturity paper and the “run” amounted to investors not rolling over the paper. Following the implicit (state dependent) contract, discussed below, SIVs were absorbed back onto the balance sheet of their sponsors. The SIV sector essentially disappeared during the panic. See Appendix B. As of December 2007, ABCP had declined by $404 billion from a peak of $1.2 trillion—a decline of about 34 percent (See Chart 13). How much of this decline is due to SIVs unwinding? According to UBS: …in August, SIV outstandings were $400 billion ($130 billion ABCP + $270 billion MTNs). Current SIV outstandings are $300 billion ($75 ABCP + $225 billion MTNs).

08 Book.indb 208

2/13/09 3:58:35 PM

The Panic of 2007

209

This is, however, illusory; a large percentage of the $75 billion current outstanding SIV ABCP is no longer held by the intended investors (such as money market funds), but rather by bank sponsors themselves (which, of course, also ties up bank balance sheets), and to a lesser extent, by ABCP dealers and capital note holders. (UBS, “Mortgage Strategist,” December 18, 2007, page 10.) Appendix B describes the outcomes for the major SIVs. Concurrently with the run on these vehicles, prices of subprime-related bonds began to decline. Highly levered hedged funds that held these bonds began to incur write-downs, and face margin calls. A number of hedge funds liquidated. Dealer banks began to announce write-downs. Why were there runs on SIVs? Did they hold massive amounts of subprime-related paper? In August 2007, a few months prior to the runs, S&P reported on the portfolio composition of SIVs: We reviewed the portfolios specifically with an eye toward mortgage assets and CDO of ABS assets, which have recently experienced considerable pricing pressure in the markets. In the aggregate, SIV portfolios remain well diversified. Portfolio exposure to residential mortgage assets and CDOs of ABS average 24%. The exposure to subprime and home equity-backed RMBS assets forms a small proportion of the portfolios. Assets backed by prime RMBS form the largest proportion of the portfolios. On average, portfolios hold approximately 21% exposure to the U.S. RMBS prime markets, of which the vast majority is ‘AAA’ rated prime assets. Two vehicles have significant above-average exposure to home equity and subprime assets. On Aug. 28, Standard & Poor’s took a rating action on Cheyne. The other vehicle, Rhinebridge, recently received an infusion of capital. In aggregate, across the portfolios of all rated SIVs, the weighted averages of the portfolio rating exposures are rounded to approximately 61% invested in ‘AAA’ rated assets, 27% invested in ‘AA’ rated assets, 12% invested in ‘A’ rated assets, and a residual of less than 1% in lower-rated assets. These numbers exclude Eaton

08 Book.indb 209

2/13/09 3:58:35 PM

210

Gary B.Gorton

1,200,000

1,200,000

1,100,000

1,100,000

1,000,000

1,000,000

7 -0

-0

07 20

07 20

1

-0 06 20

-0

1

-0 06 20

05 20

-0 05 20

-0

7

-0

04 20

04 20

-0 03 20

-0 03 20

-0 02

-0 02 20

1

600,000 7

600,000 7

700,000

7

700,000

1

800,000

1

800,000

7

900,000

1

900,000

20

$ Millions

Chart 13 Asset-Backed Commercial Paper Outstanding

Source: Federal Reserve

Vance because it focuses on the non-investment-grade corporate market and has lower leverage guidelines. The financial sector comprises a weighted average of 41.5% of SIV portfolios. Chart 14 shows the average asset distribution by sector across all SIV portfolios. SIVs did not have significant exposure to subprime in aggregate. Home equity loans and subprime were 2.01 percent. CDOs of ABS amounted to 0.28 percent. Perhaps the problem was the exposure to the financial sector, 41.50 percent. The basic problem was that investors could not penetrate the portfolios far enough to make the determination. There was asymmetric information. The run on SIVs does resemble pre-Federal Reserve panics, and it is not surprising that the “super SIV” was a proposed solution. That resembled the 19th century clearinghouses.63 VII.D. Summary Overview I have written throughout about information being “lost” due to complexity, while at other times I have described a situation as

08 Book.indb 210

2/13/09 3:58:35 PM

The Panic of 2007

211

Chart 14 Composition of SIV Portfolios 45%

45%

40%

40%

35%

35%

30%

30%

25%

25%

20%

20%

15%

15%

10%

10% 5%

5% 0%

BS

e

im

r bp

RM

u /S

EL

H

BS

s

BS

an

CM

n

de

Stu

o tL

O

CD

A of

O

CD

S

AB

Fi

l

cia

n na

e

rat

Co

o rp

er th O

0%

Source: S&P (August 2007)

being characterized as “asymmetric information.” Finally, I argued that the introduction of the ABV revealed and aggregated information. I said it created “common knowledge” about subprime risk pricing. I also argued that no one knows the location of subprime risk. In this subsection I try to clarify and focus what I mean by these terms, and in the process, summarize the information story. Table 19 may help organize these thoughts. Prior to the introduction of the ABX, there was no liquid, publicly visible market where subprime risk was directly priced. Individual transactions were priced, but these prices were not widely seen. Only the direct participants saw the prices. Moreover, parties wishing to hedge or short subprime had no easy way of doing this. To the extent that there was hedging and shorting, again the prices were not seen by a wider audience. The value of subprime mortgages, and subprime-related instruments, was not common knowledge. The ABX started trading in 2006, and started drifting downwards in the second half of that year. In 2007 all the indices showed a distinctly negative view. This negative view became known, and it became known that everyone knew

08 Book.indb 211

2/13/09 3:58:35 PM

212

Gary B.Gorton

Table 19 Summary of the Chain of Subprime Risk Step in the Chain

Information Created

Parties Involved

Origination of mortgages

Underwriting Standards: Risk characteristics of mortgages

Mortgage originators; brokers

Securitization of mortgages

Portfolio of mortgages selected and RMBS Structured

Dealer banks; servicers; rating agencies; investors buy the deal

Securitization of ABS, RMBS, CMBs in CDOs of ABS

Portfolio of ABS selected, manager selected, and CDO Structured

Dealer Banks; CDO managers; Rating Agencies; Investors buy the deal

Possible transfer of CDO risk via CDS in Negative Basis Trade

CDO and tranche selected; counterparty risk introduced

Dealer banks; banks with balance sheets; CDO

Possible sale of CDO tranches to SIVs and other such vehicles

CDO and tranche selected for SIV portfolio

SIV manager; SIV investors buy SIV liabilities

Possible investment in SIV liabilities by money market funds

Choice of SIV and seniority

Only the parties directly involved: buyer and seller

Possible sale of CDO tranches to money market funds via liquidity puts

CDO and tranche selected

Dealer banks; money market funds; put writer

Final resting place of the cash RMBS tranches, cash CDO tranches, and synthetic risk

Location of risk

Only the parties directly involved: buyer and seller

this. Once the ABX indices started to drift downwards, accountants required market participants to use these indices for mark-to-market purposes, which may have led to a feedback effect, discussed later. “Asymmetric information” is a familiar term, referring simply to a situation where one side of a transaction knows more relevant information than the other side about the object being traded, potentially leading to well-known agency problems. Referring to the table above, investors purchased tranches of RMBS, CDOs, SIV liabilities, money market funds, and so on, and did so without knowing everything known by the structurers of the securities they were purchasing. These investors likely relied on repeated relationships with bankers and on ratings. Essentially, investors do not have the resources to individually analyze such complicated structures and, in the end, rely to a lesser extent on the information about the structure and the fundamentals and more on the relationship with the product seller. Agency relationships are substituted for the actual information. To emphasize this is not surprising, and it is not unique to structured

08 Book.indb 212

2/13/09 3:58:35 PM

The Panic of 2007

213

products. But, in this case the chain is quite long. Below I discuss whether incentives were aligned in these agency relationships along the chain. No one knows where the subprime risk ultimately ended up, except that the final buyers and sellers of the risk of a particular transaction know. The final investor is invariably an agent acting as a delegated portfolio manager. Even if the final investor is a regulated entity, the entity may not report in a way which would make the risk clear to outsiders or regulators. In economics we often think of information as being exogenous payoff relevant information, such as the distribution of payoffs or the type of a manager, which affects the distribution of payoffs. Economists think of information as a “signal” about the future payoff of a security. Agents obtain signals by expending resources. If they expend resources, they learn the signal plus noise. The costs of learning the signal are recovered by trading on this private information. In the process the price aggregates the information. This is the gist of Grossman and Stiglitz’s (1980) paper. There is also information about the actions of other agents, that is, the strategies of other agents can affect payoffs, and so agents must form beliefs about what other agents are going to do. These are all familiar notions. I have argued that one problem leading to the current crisis was the loss of information. What does it mean for information to be “lost” due to “complexity”? “Lost” implies that the information was known at one point, and then it became “lost.” By “lost” I mean that for CDO investors and investors in other instruments that have CDO tranches in their portfolios, it is not possible to penetrate the chain backwards and value the chain based on the underlying mortgages. The structure itself does not allow for valuation based on the underlying mortgages, as a practical matter. There are (at least) two layers of structured products in CDOs. Information is lost because of the difficulty of penetrating to the core assets. Nor is it possible for those at the start of the chain to use their information to value the chain “upwards” so to speak.

08 Book.indb 213

2/13/09 3:58:35 PM

214

Gary B.Gorton

To be a bit more precise, the Grossman-Stiglitz story is about secondary market security trading. But, the securities and derivatives relevant to the subprime panic are not traded in secondary markets. The chain is a sequence of primary markets. In this chain, how are the signals propagated? The initial “signal” concerns the underwriting standards for the mortgages. At each step of the way, signals are somehow combined, as different portfolios are formed, each requiring multiple signals. Economists simply have no theories about the aggregation and transmission of “signals” in this context. Essentially incentive-compatible arrangements are substituted for the actual signals, which are too complex to be transmitted. VIII. The Panic Continued: Liquidity, Accounting, and Collateral Calls The Panic was rooted in the fear of losses, the location and extent of which can’t be determined. But there was also a virulent knock-on effect, which is a significant force in its own right: Liquidity dried up. With no liquidity and no market prices, the accounting practice of “marking-to-market” became highly problematic and resulted in massive write-downs based on fire sale prices and estimates. Collateral calls, also based on “marking-to-market” were massive, creating liquidity problems for some and windfall funding for others. Finally, there was an inability to raise cash because of a refusal to lend, especially in terms of repurchase agreements (repo).64 I review these issues in this section. VIII.A.

Liquidity

Aside from actual experiences of watching the repo market disappear, the evidence for the liquidity crisis is the sharp increase in spreads in important short-term funding markets, such as the interbank market. A number of observers point to the spread between Libor and the overnight indexed swap (OIS) rate of the same maturity.65 The OIS rate embeds the expectation of the overnight rate at that maturity but does not reflect credit and liquidity risks, so the idea is that the spread takes account of interest rate expectations. The increase in the spread is viewed as evidence of the stress in the interbank market, though

08 Book.indb 214

2/13/09 3:58:35 PM

The Panic of 2007

215

whether it is “liquidity” or counterparty risk, to the extent that these are different, is less clear. See Mishkin (2008), Taylor and Williams (2008A, B), Michaud and Upper (2008).66 The 3-month Libor-OIS spread is shown in Chart 15. This spread had a multi-year average of 11 basis points, and was 15 basis points on August 8, 2007. On August 10 it was over 50 basis points, and it was over 90 basis points by mid-September. This liquidity crisis was magnified by several factors, which I discuss next. The first is the accounting practice of marking-to-market. I briefly discuss this issue; a thorough discussion is beyond the scope of this paper.67 A second factor in the liquidity crisis is collateral calls.68 Illiquidity causes “mark-to-market” losses to differ significantly from expected losses based on credit fundamentals. The difference is the liquidity premium. Of course, the problem is that we have no sure measure of the illiquidity discount, nor do we have a sure measure of the expected losses based on fundamentals. The Bank of England (2008) estimated, based on actuarial methods, that the realized subprime-related losses would eventually reach $170 billion. On the other hand, an estimate based on the usual market value method gives an expected loss of $380 billion. See Bank of England (2008). This result is hardly unique: Every comparison between market–price–based measures and actuarial measures gives the same result, namely, that write-downs calculated with market-price-based measures are significantly higher than expected losses calculated using any other approach. This is no surprise—it is exactly the effect of illiquidity on prices. VIII.B.

The Impact of Accounting

The relevant accounting rule (in the U.S.) is the U.S. Financial Accounting Standards Board Rule 157, which was introduced in September 2006, to become effective for fiscal years that began after November 15, 2007. So, the rule was coming into effect essentially in the middle of the Panic.69 The rule requires that (most) positions be “marked-to-market” under FASB 157.70 The logic follows from the idea that if markets are efficient, that is, if prices aggregate the information and beliefs of market participants, then this is the best

08 Book.indb 215

2/13/09 3:58:36 PM

216

Gary B.Gorton

Chart 15 3-Month Libor-OIS Spread (bps) 120

120

100

100

6/2/08

5/2/08

4/2/08

3/2/08

2/2/08

1/2/08

12/2/07

11/2/07

10/2/07

9/2/07

0

8/2/07

20

7/2/07

20

6/2/07

40

5/2/07

40

4/2/07

60

3/2/07

60

2/2/07

80

1/2/07

80

0

Source: Bloomberg

estimate of “value.”71 I leave aside the issue of whether “efficient markets” is an accurate description of any market other than perhaps stock markets. This accounting view creates an obvious problem during a banking panic when market participants withdraw from markets, a problem which has been much commented on. See, e.g., Fitch (2008), Euromoney (March 3, 2008), Norris (2007), and Standard and Poor’s (2007, 2008), to name just a few.72 The accounting rules put the accountants at the forefront of decision-making about the valuation of complex financial instruments. While the accounting outcome is basically negotiated, the rules put management at a bargaining disadvantage. As Pollock (2008) put it: There is no doubt that the move to FAS 157 and similar rules has resulted in a shift of power toward accounting firms and away from corporate management, a shift that only adds to the change put in place by Sarbanes-Oxley. At the same time, we have this perverse situation where the accountant has to opine on accounting treatment, but they cannot provide advice to the client because that would violate their “independence.”

08 Book.indb 216

2/13/09 3:58:36 PM

The Panic of 2007

217

Accounting is supposed to produce information.73 How can that happen in a panic? In a panic, no one wants to trade; there are no markets. And hence there are no market prices. Think of a 19th century banking panic. In a 19th century banking panic, the banking system was insolvent; the system could not honor depositor demands for withdrawal. There is no place to sell the assets of the banking system. Obviously, “marking-to-market” would confirm this. In the U.S. during the 19th century this problem was solved by clearinghouses (something the short-lived “super SIV” attempted to imitate). During the 19th century, the institution of the clearinghouse evolved to the point where banks’ response to panics was fairly effective. In the face of the insolvency of the banking system, the banks suspended convertibility and issued clearinghouse loan certificates. Clearinghouse loan certificates created a market price, one which valued the assets of the banking system. These certificates traded at a discount to par initially. When the discount to par disappeared, corresponding to the market’s view that the banking system was solvent, suspension was lifted. In other words, it took time for the asymmetric information to dissipate, and when it did, suspension was lifted. This system was abandoned with the founding of the Fed and the subsequent adoption of deposit insurance. These were institutions aimed at preventing a panic from happening. But, they are not equipped to solve the information problem that arises if a panic does happen. Clearinghouse loan certificates served an important function in producing information about the aggregate banking system. There is no modern equivalent to clearinghouses. There is no informationproducing mechanism that is implemented during panics. Accountants follow rules. So, accountants enforced “marking.” Accountants initially seized on the ABX indices as the “price,” even for earlier vintages, but later were willing to recognize the difficulties of using the ABX indices. Marking-to-market, however implemented during a panic, has very real effects because regulatory capital and capital for rating agency purposes is based on GAAP. There are no sizable platforms that can operate ignoring GAAP capital. In the current situation, partly as a result of GAAP capital declines, banks are selling assets or are

08 Book.indb 217

2/13/09 3:58:36 PM

218

Gary B.Gorton

attempting to sell assets—billions of dollars of assets—to “clean up their balance sheets,” raising cash and delevering. This pushes down prices, and another round of marking down occurs, and so on. This downward spiral of prices­—marking down, selling, marking down again­—is a problem when there is no other side of the market, as has been often noted of late.74 VIII.C. The Scramble for Cash—Collateral Calls A scramble for cash ensued, not just from delevering and hoarding balance sheet, but also from collateral calls.75 E.g., Bear Stearns Form 10-K, November 30, 2007: … investors lost confidence in commercial paper conduits and SIVs causing concerns over large potential liquidations of AAA collateral. The lack of liquidity and transparency regarding the underlying assets in securitizations, CDOs and SIVs resulted in significant price declines across all mortgagerelated products in fiscal 2007. Price declines were further driven by forced sales of assets in order to meet demands by investors for the return of their collateral and collateral calls by lenders. (p. 16) Accredited Home Lenders Holding Co. SEC filing Schedule 14D-9, June 19, 2007:76 … these events with the continued heavy repurchase demands from whole loan purchasers experienced during this period created a cycle beginning with a significant increase in the amount of distressed loans for sale in the market. This increase in loan supply reduced whole loan prices, providing a basis for warehouse line providers to mark down the collateral value of loans held in inventory and, as a result, to place margin calls on non-prime lenders. These increased margin calls resulted in more distressed sales which, in turn, put further downward pressure on whole loan sale prices, regenerating the cycle with escalating negative results. (p. 8) There are many examples like this.

08 Book.indb 218

2/13/09 3:58:36 PM

The Panic of 2007

219

Collateral usage in derivative transactions has increased significantly. Collateral usage in derivative transactions is governed by the Credit Support Annex (CSA) to the ISDA Master Agreement. A CSA provides credit protection by setting forth the rules governing the mutual posting of collateral.77 The ISDA Margin Survey of 2007 estimates that the gross amount of collateral in use at the end of 2006 was $1.335 trillion, an increase of 0.4 percent over the previous year. The 2007 survey reported a 10 percent increase. The number of collateral agreements in 2007 was 133,000, compared to 110,000 in 2006. Cash is the most common kind of collateral. In the credit derivatives market, buyers of protection can make collateral calls when spreads increase, that is, when marks suggest an increase in the likelihood that protection seller will have to pay. (The mechanics of this are governed by the CSA.) Dealer banks, which have written and purchased protection, will both make collateral calls and face collateral calls. Collateral typically earns Libor, so a collateral call means paying Libor in an environment where the bank will have to pay much more than Libor to borrow. So, there is a lot at stake in collateral calls. This issue cannot be underestimated. The credit derivatives market is sizeable, indeed, and is based on collateral provisions in ISDA CSAs. The British Bankers’ Association 2006 survey estimated the total market notional at the end of 2006 to be $20.207 trillion. The ISDA mid-2007 survey estimated the size of the credit derivatives market to be $45.25 trillion. In the June 2007 survey, the U.S. Office of the Comptroller of the Currency found that the total notional amount of credit derivatives held by U.S. commercial banks was $10.2 trillion. To put these numbers in a broader perspective, keep in mind that the U.S. corporate bond market is currently $5.7 trillion, and that the U.S. Treasury market is currently $4.3 trillion.78 For the party calling for collateral, collateral becomes a form of funding. Because Libor is paid on collateral, firms receiving collateral can fund themselves at Libor, when issuing debt in the market would cost them much more. This is one reason that the scramble for cash in the form of collateral calls is very important. In fact, it is difficult to convey the ferocity of the fights over collateral.

08 Book.indb 219

2/13/09 3:58:36 PM

220

Gary B.Gorton

VIII.D. Panic in the Repo Market Aside from collateral calls creating a scramble for cash, the basic form of lending, repo, disappeared. The most important part of the panic occurred in the repo market. Repos are essentially secured loans, so counterparty risk is not an issue. All general collateral (GC) repos have the same rate, the GC repo rates, or simply the repo rate. Typically, repos can be rolled over easily and indefinitely, though the repo rate may change. Repo is integral to intermediation by dealer banks because when assets are purchased for sale later the assets are financed by repo. Repo is likely one of the largest financial markets, though there are no official statistics on the size of the market. Tripartite repo was $2.5 trillion in 2007 (see Geithner, 2008).79 Tripartite repo is estimated to be about 15–20 percent of the repo market. 80 With respect to the financing activities of primary dealers, reporting to the New York Fed, the average daily outstanding repo and reverse repo contracts totaled $7.06 trillion in the first quarter of 2008, a 21.5 percent increase over the same period in 2007. See the Securities Industry and Financial Markets Association (2008, p. 9). The Bond Market Association (since renamed the “Securities Industry and Financial Markets Association”) (2005) conducted a dealer survey in September 2004 to determine the size of the repo market. As of June 30, 2004, the repo and securities lending market was $7.84 trillion. It is generally believed that this market has grown at around 10 percent per year, making it about $11.5 trillion today. The repo market virtually disappeared in August 2007, and the drought has lasted for months. The repo market dried up because dealer banks would not accept collateral because they rightly believed that if they had to seize the collateral, there would be no market in which to sell it. This is due to the absence of prices. The amount lent depends on the perceived market value of the asset offered as security. If that value cannot be determined, because there is no market—no liquidity—or there is the concern that if the asset is seized by the lender, it will not be saleable at all, then lender will not engage in repo.

08 Book.indb 220

2/13/09 3:58:36 PM

The Panic of 2007

221

Why did the repo market disappear, if the problem was uncertainty about the valuation of subprime bonds? One can understand that dealers would not want to take subprime RMBS as collateral in repo, but what about ABS, RMBS, and CMBS generally? Repo traders report that there was uncertainty about whether to believe the ratings on these structured products, and in a very fast moving environment, the response was to pull back from accepting anything structured. If no one would accept structured products for repo, then these bonds could not be traded­—and then no one would want to accept them in a repo transaction. This externality is reminiscent of Pagano (1989). Without repo, assets cannot change hands, because the intermediaries cannot function. The only way to sell assets is at extremely low prices. But low prices then have a feedback affect, as they cause the mark-to-market value of all assets to fall, making it even less likely that repo can be done. Like repo, collateral calls, against credit derivative positions for example, are also based on marks. That leads to fights over collateral due to disagreements about prices (such fights are ultimately governed by the Credit Support Annex). E.g., the VCG Special Opportunities hedge fund sued Wachovia after the fund was asked to post $750,000 of collateral, but then was asked for an increase to $8.2 million. The fund refused the final call of $1.49 million and Wachovia foreclosed on the fund (see Wall Street Journal, March 4, 2008, p. C1). IX.

Explaining the Panic: A Competing Hypothesis

I have argued that the design of subprime mortgages and subprime securitizations are unique in that they are particularly sensitive to declines in house prices, leading to an information problem for investors when the house price bubble burst, particularly due to the distribution methods, including CDOs, off-balance sheet vehicles, and derivatives. In my view, it is precisely the particularity of the underlying subprime mortgage design and its transmission through the chain of structures that explains the problem. There is a specific

08 Book.indb 221

2/13/09 3:58:36 PM

222

Gary B.Gorton

sensitivity to house prices embedded in the design of these securities, structures, and derivatives. There are no such issues with securitization generally, or with the use of off-balance sheet vehicles for the securitization of those asset classes. Other securitizations are not so sensitive to the prices of the underlying assets and so they are not so susceptible to bubbles. So, my claim is that a very specific set of interlinked security designs made the chain susceptible to a house price decline. House prices stopped increasing in 2006, and the effects were revealed by ABX prices. There is, however, another hypothesis about the panic, and in this section, I very briefly discuss this competing hypothesis. The dominant explanation for the Panic is the “originate-to-distribute” view, which is the idea that banking has changed in such a way that the incentives have been fundamentally altered as a general matter. It is argued that originators and underwriters of loans no longer have an incentive to pay attention to the risks of loans they originate, since they are not residual claimants on these loans. In this view, investors apparently do not understand this and have been fooled (fingers point to the rating agencies). The “originate-to-distribute” viewpoint has been described by The Joint Forum (which includes the Basel Committee on Banking Supervision, the International Organization of Securities Commissions, and the International Association of Insurance Supervisors) as follows: …under the “originate-to-distribute” model, banks frequently no longer have significant retained exposures, nor have they necessarily retained the personnel specializing in workouts who can steer creditor negotiations. (Credit Risk Transfer, April 2008, p. 20) Since 2005, the growth of CRT [Credit Risk Transfer] continues to provide banks and securities firms with opportunities to profit from originating, structuring and underwriting CRT products. They can earn fees while not having to hold the associated credit risk or fund positions over an extended

08 Book.indb 222

2/13/09 3:58:36 PM

The Panic of 2007

223

time period. This has been termed the “originate-to-distribute” model. (Credit Risk Transfer, April 2008, p. 41) Here is a slightly fuller articulation of the view, by Mishkin (2008): The originate-to-distribute model, unfortunately, created some severe incentive problems, which are referred to as principal-agent problems, or more simply as agency problems, in which the agent (the originator of the loans) did not have the incentives to act fully in the interest of the principal (the ultimate holder of the loan).  Originators had every incentive to maintain origination volume, because that would allow them to earn substantial fees, but they had weak incentives to maintain loan quality. All major bank regulators and central bankers appear to subscribe to this view, though their views have some differences and nuances.81 There is no question that banking has changed, and that these changes are very significant.82 Chart 16 conveys a sense of the magnitudes of these changes. Issuance of asset-backed securities, excluding mortgagebacked securities, has exceeded the issuance of corporate debt in the U.S. in the past few years. Broadly, “originate-to-distribute” refers to this change. Twenty-five years ago, there were no asset-backed securities. In addition, banks sell loans. The syndicated loan market was $1.5 trillion in 2005 for non-financial corporations. Secondary loan trading in 2005 had a market volume of $176.3 billion. See Drucker and Puri (2007). Clearly, the old model of the bank, in which loans were held to maturity, does not exist as it used to. The issue is whether these changes somehow explain the panic. The “originate-to-distribute” seems to refer to the general trend in banking that has been going on for at least twenty years, possibly starting with the junk bond market becoming a major competitor for bank loans.83 In response to this, and other competition, banks began selling loans and securitizing assets.84 The originate-to-distribute view proposes nothing specific to explain why problems arose with the securitization of subprime mortgages, as opposed to any other category

08 Book.indb 223

2/13/09 3:58:36 PM

224

Gary B.Gorton

Chart 16 Issuance of Various Securities ($ billions) 3,500

3,000 Mortgage–Related 2,500 Corporate Debt 2,000 Asset–Backed Securities 1,500

1,000

500

0

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

of assets that are securitized. In fact, in securitization generally, there does not seem to have been the same problems as with subprime mortgages. The “severe incentive problems” and “principal-agent problems” would seem to be present in all securitizations. IX.A. Were Incentives Aligned? The “originate-to-distribute” view argues that the risks of loans were passed along to investors, leaving the originators with no risk. But, this can be immediately rejected. Significant losses have been suffered by many up and down the subprime chain. Originators, securitization structurers, and underwriters—firms and individuals—have suffered. The subprime originators/underwriters that went bankrupt include, e.g., Option One, Ameriquest, New Century, and to the likes of Citibank, UBS, and Merrill Lynch, with billions of write-downs.85 The following “agents” were fired: Chuck Prince, Ken Thompson, Marcel Ospel, James Cayne, Huw Jenkins, Stanley O’Neal, and a host of others. Thousands of other employees up and down the chain have lost their jobs. If these firms and individuals took excessive risk, they have realized losses. The fact there have been losses on subprime mortgages is not ipso facto evidence of a lack of incentives.

08 Book.indb 224

2/13/09 3:58:36 PM

The Panic of 2007

225

How are interests aligned in securitization? There is direct exposure to the originated risk, and there are implicit contracts making the arrangements incentive-compatible. I very briefly review these points. Originators of subprime mortgages face a number of direct risks. The mortgages must be warehoused by the originator prior to securitization. In other words, loans must be held before they are securitized. See Gordon (2008). When the pool of mortgages is large enough, they are transferred to the underwriter, who will assemble the securitization. The underwriters of the securitizations then must warehouse the RMBS tranches. In later stages, securitization tranches will be warehoused by the dealer banks, who underwrite the CDOs. In 2006 and early 2007 some banks kept the most senior portions of CDOs on their balance sheets. Along this chain, these firms have significant risks in warehousing the different securities. Much of the write-downs by banks came from such warehousing. For example, UBS “Shareholder Report on UBS’s Write-Downs,” April 18, 2008: UBS acquired its exposure to CDO Warehouse positions through its CDO origination and underwriting business. In the initial stage of a CDO securitization, the desk would typically enter into an agreement with a collateral manager. UBS sourced residential mortgage backed securities (“RMBS”) and other securities on behalf of the manager. These positions were held in a CDO Warehouse in anticipation of securitization into CDOs. Generally, while in the Warehouse, these positions would be on UBS’s books with exposure to market risk. Upon completion of the Warehouse, the securities were transferred to a CDO special-purpose vehicle, and structured into tranches. (p. 13) The CDO Warehouse was a significant contributor to Value at Risk (“VaR”) and Stress limits applicable to this business relative to other parts of the CDO securitization process and warehoused collateral was identified as one of the main sources of market risk in reviews by IB Market Risk Control (“MRC”) conducted in Q4 2005 and again in Q3 2006. (p. 13)

08 Book.indb 225

2/13/09 3:58:37 PM

226

Gary B.Gorton

Similarly, the CFO of Bear Stearns, during the Earnings Conference Call of December 20, 2007: “…of the $1.9 billion in writedowns… about $1 billion of that came from the writedowns of CDOs and the unwinding of the CDO warehouse.” Warehousing is not the only risk. Originators of mortgages retain significant interests in the mortgages they originate due to servicing rights and retained interests. Mortgage servicing rights are valuable, and retained interests are also significant. When loans are sold in the secondary market, the mortgage servicing rights that are created are typically not sold.86 An example of the value of mortgage servicing rights is provided by Countrywide. Countrywide Form 10-K, December 31, 2007: When we sell or securitize mortgage loans, we generally retain the rights to service these loans. In servicing mortgage loans, we collect and remit loan payments, respond to customer inquiries, account for principal and interest, hold custodial (impound) funds for payment of property taxes and insurance premiums, counsel delinquent mortgagors and supervise foreclosures and property dispositions. We receive servicing fees and other remuneration in return for performing these functions. (p. 7) In October 2007 Countrywide recorded write-downs of $830.9 million in the value of mortgage servicing rights. As of March 31, 2008, Countrywide had an estimated value of mortgage servicing rights of $17 billion and a total assets of $199 billion, about 9 percent of total assets (see SEC Form 10-K, April 29, 2008). More formally, see Kohlbeck and Warfield (2002), calculate the present value of mortgage servicing rights for a sample of banks and show its relation to abnormal earnings. They find that the present value of mortgage servicing rights, as a percentage of equity, ranges from 2.7 percent to 3.5 percent. Other financial interests are often retained as well, including, for example, interest-only securities, principal-only securities and residual securities. These retained financial interests are also significant.

08 Book.indb 226

2/13/09 3:58:37 PM

The Panic of 2007

227

Missal (2008): “New Century’s residual interests were large assets of the Company (worth hundreds of millions of dollars)” (p. 234). The overcollateralization gives the sponsor a Credit Enhancement Security­—a claim on the OC. These could be securitized in NIMs. Then the sponsor of the NIMs would retain a residual interest in the NIMs trust, which would remain on the balance sheet. Perhaps a more detailed example can summarize this point. The information below and Table 20 are from page 35 of the 2007 Merrill Lynch Annual Report: Residuals: We retain and purchase mortgage residual interests which represent the subordinated classes and equity/first-loss tranche from our residential mortgage-backed securitization activity. We have retained residuals from the securitizations of third-party whole loans we have purchased as well as from our First Franklin loan originations. Residential mortgage-backed securities (“RMBS”): We retain and purchase securities from the securitizations of loans, including sub-prime residential mortgages. Warehouse lending: Warehouse loans represent collateralized revolving loan facilities to originators of financial assets, such as sub-prime residential mortgages. These mortgages typically serve as collateral for the facility. Table 20 provides a summary of our residential mortgagerelated net exposures and losses, excluding net exposures to residential mortgage-backed securities held in our U.S. banks for investment purposes. Note the sizes of “Warehouse Lending,” “Residuals,” and “Mortgage Servicing Rights” (the numbers are in millions of dollars). The losses are clearly significant.87 All along the chain, from originators to underwriters, there are very significant risks involved in creating and maintaining securitized products. There are also implicit contractual arrangements in securitization, between the investors in the securitized assets—buyers of tranches—

08 Book.indb 227

2/13/09 3:58:37 PM

228

Gary B.Gorton

Table 20 Residential Mortgage-Related Net Exposures and Losses ($ millions) Net Exposure as of Dec. 29, 2007

Net Losses for the Year ended Dec. 28, 2007

Warehouse Lending

$137

$(31)

Whole Loans

994

(1,243)

Residuals

855

(1,582)

Residential MBS

723

(332)

Total U.S. Subprime

2,709

$(3,188)

U.S. Alt-A

2,687

(542)

U.S. Prime

28,189

N/A

Non-U.S.

9,582

(465)

Mortgage Servicing Rights

389

N/A

Total

$43,556

$(4,195)

U.S. Subprime

Source: Merrill Lynch Annual Report, 2007, p. 357

and the sponsors of the deals.88 Gorton and Souleles (2007) argue that there is an implicit contract between the sponsor and investors in the liabilities of the SPVs used for securitization. The implicit contract exists precisely to address the agency problems that could arise when assets are sold, essentially is that the sponsor of the securitization guarantees it. How do we know that such implicit contracts exist? Gorton and Souleles, empirically analyzing credit card securitizations, argue that this implicit contract is understood by investors and provide evidence that it is priced. Implicit contractual arrangements have also been argued to explain loan sales. Loan sales are not supposed to happen according to the traditional theories of banking, but following the advent of the junk bond market, banks began to sell loans. Although not required to retain part of the loan, banks in fact do retain pieces, more so for riskier borrowers. Also, loan covenants are tighter for riskier borrowers, whose loans are sold. See, e.g., Gorton and Pennacchi (1995, 1989); Calomiris and Mason (2004); Drucker and Puri (2007); and Chen, Liu and Ryan (2007). Jiangli and Pritsker (2008) “find that banks use mortgage securitization to reduce insolvency risk.”

08 Book.indb 228

2/13/09 3:58:37 PM

The Panic of 2007

229

With respect to subprime specifically, the implicit contractual arrangement between SIV sponsors and investors led sponsoring banks to take the off-balance sheet SIVs back onto their balance sheets, when there was no explicit obligation to do so, consistent with the arguments of Gorton and Souleles (2007). See Appendix B. IX.B. Did Underwriting Standards Decline? The evidence cited for the alleged “originate-to-distribute” agency problems is the deterioration of the 2006 and early 2007 subprime mortgages. Subprime performance during the period 2001-2005 was good by historical (subprime) standards. While delinquency and foreclosure rates for subprime mortgages were higher than for prime mortgages, their experience was as expected, i.e., delinquencies and foreclosures rose during the recession of the early 2000s. But, the 2006 vintage of subprime mortgages is much worse. The extreme deterioration of the 2006 vintage has been attributed to a decline in underwriting standards and to outright fraud. For example, the President’s Working Group on Financial Markets (March 2008) concluded that “The turmoil in financial markets was triggered by a dramatic weakening of underwriting standards for U.S. subprime mortgages, beginning in late 2004, and extending into early 2007” (p. 1; emphasis in original). Or, another example, according to Fitch (2007): Fitch attributes a significant portion of this [2006] early default performance to the rapid growth of high-risk “affordability” features in subprime mortgages…. In addition to the inherent risk of these products, evidence is mounting that in many instances these risks were not controlled through sound underwriting practices. Moreover, in the absence of effective underwriting, products such as “no money down” and “stated income” mortgages appear to have become vehicles for misrepresentation or fraud ….(p. 1) The evidence often cited are statistics like those in Table 21, which shows the time profile of some subprime mortgage characteristics.

08 Book.indb 229

2/13/09 3:58:37 PM

230

Gary B.Gorton

Table 21 Underwriting Standards for Subprime Mortgages



2001

ARM Share

Interest Only Share

Low/No Doc Share

Debt-to-Income Ratio

Average Loan-toValue Ratio

73.%

0.0%

28.5%

39.7

84.0

2002

80.0%

2.3%

38.6%

40.1

84.4

2003

80.1%

8.6%

42.8%

40.5

86.1

2004

89.4%

27.2%

45.2%

41.2

84.7

2005

93.3%

37.8%

50.7%

41.8

83.2

2006

91.3%

22.8%

50.8%

42.4

83.4

Source: Freddie Mac; see Joint Economic Committee (October 2007).

Looking separately at these characteristics, it seems that standards were lowered. Note, for example, the increase in the percentage of mortgages with less than full documentation. But, such statements are problematic because there are many dimensions to borrower risk and there are trade-offs between them. For a given aggregation of risk, there is a trade-off between risk and return. So, it seems difficult to define a “decline in lending standards.” Bhardwaj and Sengupta (2008B) attempt to address the multidimensional nature of lending standards. Before getting to econometric tests, however, they point out the difficulties of casual observation. For example, “…borrowers with lower documentation have on average higher FICO scores” (p. 12). Or, “For a given vintage, mortgages with a smaller LTV have a lower FICO score on average” (p. 14). FICO scores trend gradually up over the period 1998-2006 (see Bhardwaj and Sengupta, 2008B). Their final conclusion is: “Noticeably, there is little to suggest anything particularly remarkable about underwriting standards for mortgages of 2005-2007 vintages…” (p. 16). So, what does explain the performance of the post-2005 vintage subprime mortgages? House price appreciation (HPA), or more specifically depreciation, is the biggest single factor explaining defaults. For example, according to UBS: “…HPA alone is able to explain ~60% of the credit performance variance across states. Combined with combined LTV and percentage Full Doc, the three variables account for ~74% of credit performance variance. Also, interestingly, FICO score is statistically insignificant in interpreting the credit

08 Book.indb 230

2/13/09 3:58:37 PM

The Panic of 2007

231

performance.” See UBS, “Mortgage Strategist,” Nov. 13, 2007, p. 33. The conclusion that HPA is the most important factor explaining default and loan severity is confirmed with econometric evidence. See Bhardwaj and Sengupta (2007A): Using a competing risk hazard model, we show that an appreciation in house price had a positive and significant effect on the likelihood of prepayment but a negative and significant effect on the likelihood of default. In a regime of rising house prices, a financially distressed borrower could avoid default by prepaying the loan (either through a refinance or a property sale). Conversely, a sudden reversal in prices increased default in this market because it made this prepayment exit option cost-prohibitive. The conclusion that HPA is the most important factor explaining default and loan severity is evidenced by Demyanyk and Van Hemert (2007). If underwriting standards were declining, then “first payment default” mortgages would increase. These are mortgages where the borrower defaults right away, missing the very first monthly payments. But, most securitization contracts stipulate that if there is an early payment default, or some other defect in the mortgage (e.g., incorrect documentation), then the mortgage originator must repurchase the mortgage from the SPV. Because it is a defective mortgage, its value declines, so the originator incurs a realized loss; it has repurchased a loan for the same amount at which it was sold to the SPV, but has received back a mortgage worth less. It is difficult to see how a dramatic decline in underwriting would not result in a large number of first payment defaults that the originators would have to absorb. Since the originators would, in fact, absorb these mortgages, they have no incentive to make them in the first place. Finally, it is worth noting that evidence of a decline in lending standards is only a piece of the puzzle. The argument must be that, if this did occur, it was not reflected in the structure of the RMBS bonds. Somehow, the structurers would have to have been fooled into not increasing the credit enhancement to reflect this decline. This has never been systematically examined.

08 Book.indb 231

2/13/09 3:58:37 PM

232

Gary B.Gorton

IX.C. Summary Securitization is an efficient, incentive-compatible response to bankruptcy costs and capital requirements. Although there are only a few studies, the evidence to date is consistent with the experience of a quarter century of securitization working very well. The assertions of the “originate-to-distribute” view simply are not consistent with what we know. The idea that there is a moral hazard due to the alleged ability of originators to sell loans without fear of recourse, and with no residual risk, also assumes that the buyers of these loans are irrational. That may be, but the irrationality, it turns out, had to do with the belief that house prices would not fall. X.

Concluding Remarks It might very properly be urged that the present is too early a date for us to draw wise conclusions from the lessons of the recent financial crisis. Indeed, one can hardly speak of it, as I did just now, as the recent crisis. It is the present crisis…. Domestic exchanges are still seriously disorganized. After the most heroic measures for relief, taken by the Treasury and by banks generally, we continue to be surrounded by abnormal conditions, and the day is somewhere in the future when we can look back with anything like academic interests and comment with intelligence on the true lessons which have been taught by this extraordinary financial event.

—Frank Vanderlip, Vice-President, National City Bank, New York, speaking of the Panic of 1907; see Vanderlip (1908, p. 2). When I read the numberless projects for our financial well being that fill the newspapers, our book shelves, and the Congressional Record, I ask myself on what do these men base their plans, on observation or actual contact and familiarity with the subject they talk about, and I must conclude that much of it is spun out of their inner consciences.

08 Book.indb 232

2/13/09 3:58:37 PM

The Panic of 2007

233

—William Nash (1908, p. 61), speaking of the Panic of 1907. The Panic of 1907 led to the founding of the Federal Reserve System. In 1908 Congress passed the Aldrich-Vreeland Act, which, among other things, created the National Monetary Commission. This commission published a voluminous report that served as the major impetus for the founding of the Fed. (See Weston, 1922, for a review.) The Federal Reserve Act passed in December 1913. But, it was then followed by a panic in 1914. (See Sprague, 1915, and Silber, 2007.) And, of course, the Great Depression came later, followed by large institutional changes, with the advent of deposit insurance being foremost among them. A century after the Panic of 1907 we again contemplate the causes of a panic. Identifying the causes of the Panic of 2007 will in large part determine the policy response to the crisis. I have argued that the subprime crisis was caused by information problems related to declining house prices, which prevent subprime mortgages from being refinanced. The design of subprime mortgages is unique in that they are linked to house price appreciation. The securitization of subprime mortgages is also unique. Because subprime mortgages are financed through a chain of securities and structures, investors could not easily determine the location and extent of the risk. Information was lost. The house price declines led to a fear of losses that could not be measured because the subprime risk had been spread around the globe opaquely. The available information was on the side of the market that produced the chain of structures; outside investors know much less. The problem is that the magnitude of the structures, and their impenetrability by outsiders, was not completely understood; it was not common knowledge. The introduction of the ABX indices created a set of market prices that aggregated and revealed that subprime-related securities were worth a lot less than had been thought. The ability to short subprime risk may have burst the bubble and, in any case, resulted in the market crowding on the short side to hedge, driving ABX prices very low. The Panic was then on. There is much work to be done to understand the ongoing panic, to formally test my (sometimes admittedly vague) conjectures, and

08 Book.indb 233

2/13/09 3:58:37 PM

234

Gary B.Gorton

it will be surely be some time before researchers can sort through the events. As Mr. Vanderlip wrote, above, the lessons to be learned are likely only going to be known when there is more distance from the events. But, since panics are rare, it may be that we never have the ability to formally test in the way that is acceptable to academic economists. The scholars who studied panics before us, many of whom I have quoted, described the events with narratives. Perhaps that is the best we can do. I have tried to convey the richness of the information and agency setting in which the crisis is taking place. At this point a few tentative conclusions can be drawn. • Subprime mortgages were a financial innovation designed to be profitable by serving a constituency that had previously not had access to mortgage financing and hence could not own homes. This point is very important because the future regulatory response to the crisis will have implications for whether this constituency’s needs will be met in the future or not. The re-regulation of the financial system is intertwined with national housing policy and this should be recognized. The current situation with Fannie Mae and Freddie Mac also stresses this point. • The crisis was caused by house prices not rising and then falling. The introduction of the ABX index revealed that the values of subprime bonds (of the 2006 and 2007 vintage) were falling rapidly in value. But, it was not possible to know where the risk resided, and without this information, market participants rationally worried about the solvency of their counterparties. This led to a general freeze of intra-bank markets, write-downs, and a spiral downwards of the prices of structured products as banks were forced to dump assets. • The crisis illustrates and emphasizes the extent to which the traditional banking system is no longer as central to the savingsinvestment process as it once was. The capital markets, through the sale of intermediary-originated loans via securitization and the distribution of risk through derivatives, highlight the centrality of capital markets and illustrate the flexibility of structured

08 Book.indb 234

2/13/09 3:58:37 PM

The Panic of 2007

235

products. This evolution has been going on for at least 25 years and should be viewed favorably. • Securitization generally is not the problem currently. It is not the cause of the crisis. Securitization is an efficient form of financing, and there is no evidence that there is a systematic agency problem in its functioning. Rather, the particular form of the design of subprime mortgages is at the root of the problem. It was highly sensitive to house prices, and this sensitivity was passed through to a variety of other financial structures. • Structured products and derivatives allow for the distribution of risks globally in a way which is opaque. In principle, this is not different than in, say, the 19th century U.S. banking system, in which the particular loans that banks held was also opaque. Opaqueness and innovation probably go together, and there is a danger that innovation will be squelched if we do not recognize that there is likely a trade-off here. The lesson, perhaps, is that we should be looking at the sectors that are very opaque, such as the hedge fund world, more closely. • At the heart of many academic analyses of the functioning of capital markets and crises is the notion of “collateralizable” wealth, roughly the amount of verifiably riskless assets or bonds that an economic agent has available to borrow against. The current crisis shows that in the case of financial collateral, it can be the case that portfolios thought to be safe, collateralizable wealth in all states of the world, ex ante, turn out not to be collateralizable in the crisis. Even agency bonds, for example, were not acceptable as collateral in the repo market in August 2007. What is “collateralizable” is very intimately related to information. There is simply no financial wealth that can be thought of as “collateralizable” in all states of the world. • The crisis also illustrates that “states of the world” may best be viewed as endogenous. The chain of securities and structures created a “state of the world” that many agents did not recognize as existing ex ante. So, the notion of “incomplete markets” may be more complicated than we generally recognize.

08 Book.indb 235

2/13/09 3:58:37 PM

236

Gary B.Gorton

• Institutional investing, in general, must rely on representations from underwriters because each investor cannot afford extensive staffs of analysts. This would be excessively duplicative. As a result, when an investment is at the end of a chain, or perhaps early in the chain, the investment is made on the basis of a repeated agency relationship (with the seller and the rating agency) rather than on extensive and costly production of information. The complexity of the design makes this substitution of agency for information more likely. Information is lost. This is not a problem during normal times, as incentives are aligned. But, in a crisis, it makes it very difficult for investors to understand and value the risk. • Accounting is widely recognized as having a great deal of difficulty measuring firm value in a world of derivatives. The crisis also reveals that marking-to-market, based on the notion of “efficient markets” is flawed and needs to be rethought. At some point, double-entry bookkeeping as a paradigm will be recognized as inappropriate for financial firms, which are already moving aggressively to risk management as a replacement. Risk management, of course, is not perfect by any means, but looking at firms’ behavior the revealed preference of managements is that this is a more informative way of looking at firms. • As Merton Miller (1986) pointed out over twenty years ago, financial innovation is largely driven by regulation and taxes. Regulation means constraints and costs. Imposing capital requirements on banks, for example, that are not consistent with their competitive environment accelerates disintermediation (see Gorton and Winton, 2000). Imposing costs, such as Sarbanes-Oxley, may have led to a competitive disadvantage for U.S. capital markets. See, e.g., Zingales (2007). Entrepreneurs will take risk in some form, somewhere. In a global environment, one where capital is extremely fluid, and risk can be moved quickly with derivatives, it will be difficult for national regulators to constrain entrepreneurs. The trends are already clear. Talent is increasingly moving to the least regulated platform: hedge funds. See, e.g.,

08 Book.indb 236

2/13/09 3:58:37 PM

The Panic of 2007

237

Hester and Burton (June 19, 2008) and Guerrera and Brewster (April 30, 2008). Postscript (written October 4, 2008) Since the above was written, the Panic has continued almost out of control, and the economy has noticeably deteriorated. Lending has stopped almost completely. There is no doubt that we are now in a recession. The financial landscape has been completely altered by failures, mergers, and de facto nationalization. The “Emergency Economic Stabilization Act of 2008” has been passed. The logic of the plan seems to be that by buying distressed assets from banks, the uncertainty about the value of the banks will be removed, possibly enticing new investors to recapitalize the banks. The success or failure will depend on the exact details of how this is implemented. If the money allocated to the Troubled Assets Relief Program is used to try to shore up the weakest bank, the government may quickly use up the money allocated by Congress. That is the danger. If the money shores up banks that are stronger, it may be possible to entice lending again.

Author’s Note: Thanks to Geetesh Bhardwaj, Omer Brav, Adam Budnick, Jared Champion, Kristan Blake Gochee, Itay Goldstein, Ping He, Bengt Holmström, Lixin Huang, Matt Jacobs, Arvind Krishnamurthy, Tom Kushner, Bob McDonald, Hui Ou-Yang, Ashraf Rizvi, Geert Rouwenhorst, Hyun Shin, Marty Wayne, Axel Weber, and to those who wished to remain anonymous, for comments, suggestions, and assistance with data and examples. In the interests of full disclosure, the author has, for the last twelve years, been intimately involved in modeling, structuring, and transacting very significant synthetic credit portfolios, as a consultant to AIG Financial Products Corp. The views expressed are those of the individual author and do not necessarily reflect the official positions of AIG Financial Products Corp.

08 Book.indb 237

2/13/09 3:58:37 PM

238

Gary B.Gorton

Appendix A: A Brief Chronology of the Events of the Panic of 2007 Date Event December 2006: Ownit Mortgage Solutions files for bankruptcy. March 13, 2007: Mortgage Banker Association data for the last three months of 2006 shows late or missed payments on mortgages rose to 4.95 percent, rising to 13.3 percent in the subprime market. Subprime lender Accredited Home Lenders loses 65 percent of its value, having lost 28 percent a day earlier. April 2: Mortgage originator New Century Financial Corporation files for bankruptcy. April 18: Ellington Capital Management, large hedge fund, buys $2.9 billion of nonprime mortgage loans from Fremont General Corp. May 3: UBS closes its hedge fund Dillon Read Capital Management. June 10-12: Moody’s downgrades the ratings of $5 billion worth of subprime RMBS bonds and places 184 CDO tranches on review for downgrade. S&P places $7.3 billion of 2006 vintage RMBS bonds on negative watch and announces a review of CDO deals ex posed to subprime RMBS bonds. June 20: News reports suggest that two Bear Stearns-managed hedge funds invested in securities backed by subprime mortgage loans are close to being shut down. June 22: One of the troubled hedge funds is bailed out through an injection of $3.2 billion in loans. July10: S&P places $7.3 billion worth of 2006 vintage ABSs backed by residential mortgage loans on negative ratings watch and announces a review of CDO deals exposed to such collateral; Moody’s downgrades $5 billion worth of subprime mortgage bonds. July 11: Moody’s places 184 mortgage-backed CDO tranches

08 Book.indb 238

2/13/09 3:58:37 PM

The Panic of 2007

239

on downgrade review; further reviews and downgrades are announced by all major rating agencies in the following days. July 24: U.S. home loan lender Countrywide Financial Corp. reports a drop in earnings and warns of difficult conditions ahead. July 26: The NAHB index indicates that new home sales slid by 6.6 percent year on year in June; DR Horton, the largest homebuilder in the United States, reports an April–June quarter loss. July 30: Germany’s IKB warns of losses related to the fallout in the U.S. subprime mortgage market. Its main shareholder, Kreditanstalt für Wiederaufbau (KfW), assumes its financial obligations from liquidity facilities provided to an asset-backed commercial paper (ABCP) conduit exposed to subprime loans. July 31: American Home Mortgage Investment Corp. announces its inability to fund lending obligations; Moody’s reports that the loss expectations feeding into the ratings for securitizations backed by Alt-A loans will be adjusted. Hedge fund Sowood capital informs investors it will shut down after losing 57 percent during the month (Sowood Alpha Fund). Sowood went from $3 billion to $1.5 billion in less than four weeks. August 1: Further losses exposed at IKB lead to a €3.5 billion rescue fund being put together by KfW and a group of public and private sector banks. August 3-10: Massive deleveraging causes quant hedge funds to suffer losses. August 6: American Home Mortgage Investment Corp. files for Chapter 11 bankruptcy, leading to a term extension on outstanding ABCP by one of its funding conduits. August 9: BNP Paribas freezes redemptions for three investment funds, citing an inability to appropriately value them in the current market environment; the ECB injects €95 billion of liquidity into the inter-

08 Book.indb 239

2/13/09 3:58:37 PM

240

Gary B.Gorton

bank market; other central banks take similar steps. August 13: Coventree, the largest nonbank sponsor of Canada’s asset-backed commercial paper market announces that it is unable to place any asset-backed commercial paper on behalf of its conduits, including Aurora, Comet, Gemini, Planet, Rocket, Slate, SAT, and SIT II. August 16: Countrywide draws its entire $11.5 billion credit line. August 17: The Federal Reserve’s Open Market Committee issues a statement observing that the downside risks to growth have increased appreciably; the Federal Reserve Board approves a 50 basis point reduction in the discount rate and announces that term financing will be provided for up to 30 days; Run on Countrywide: “Anxious customers jammed the phone lines and website of Countrywide Bank and crowded its branch offices to pull out their savings because of concerns about the financial problems of the mortgage lender that owns the bank,” Los Angeles Times, August 17, 2007. August 23: Countrywide gets $2 billion cash injection from Bank of America. September 4: Overnight Libor reaches 6.7975 percent, the highest level since the LTCM crisis. Bank of China reveals $9 billion in subprime losses. September 9: Run on Northern Rock; see Telegraph.co.uk, September 14, 2007. September: Cheyne Finance SIV goes into receivership, the first SIV to do so. September 15: There is a run on British bank Northern Rock, the first in 150 years; £1 billion, amounting to 4-5 percent of retail deposits, are withdrawn (see BBC News: http://news.bbc.co.uk/2/hi/business/6996136.stm). September 18November: Repeated large write-downs by major financial firms, leading to several high-profile CEOs to leave their positions.

08 Book.indb 240

2/13/09 3:58:37 PM

The Panic of 2007

241

October 15: Citigroup writes down additional $5.9 billion. October 18: Rhinebridge Plc, the IKB SIV, suffers a “mandatory acceleration event” after IKB determines that the SIV may be unable to repay its debt. November 13: Bank of America, Legg Mason, SEI Investments, and SunTrust Banks step in to prop up their money market funds against possible losses to debt issued by SIVs. November 26: HSBC takes $41 billion in SIV assets onto its balance sheet. November 27: Citigroup agrees to sell shares worth $7.5 billion to an investment fund owned by Abu Dhabi. November 29: E-Trade, the online brokerage that was teetering at the edge of the subprime mortgage abyss, received a $2.55 billion bailout package led by Citadel Investment Group, a large hedge fund. December 3: West LB and HSH Nordbank bailout $15 billion of their SIVs. December 10: UBS announces a further $10 billion write-down. Bank of America announces it is shutting a $12 billion money-market mutual fund after losses on subprime-related instruments, including investments in SIVs. December 15: Citibank says it will take its seven SIVs back onto its balance sheet, $49 billion. December 19: Morgan Stanley writes off $9.4 billion. ACA, a financial guarantor rated A, is downgraded to CCC by S&P, triggering collateral calls from its counterparties. January 3, 2008: Peloton Partners, a $3 billion hedge fund, forced to liquidate. January 15: Citigroup announces a fourth quarter loss, partly due to $18 billion of additional write-downs on mortgage-related exposure. February 27: Hedge fund Sailfish Capital Partners announces it is liquidating. Sailfish had managed $1.9 billion in the

08 Book.indb 241

2/13/09 3:58:37 PM

242

Gary B.Gorton

previous year. March 3: Thornburg Mortgage Asset Corp. announces that it could not meet margin calls. March 7-16: Fed announces an increase of $40 billion in the size of its new Term Auction Facility, and then expands its securities lending activities through a $200 billion Term Securities Lending Facility that lends Treasuries against a range of eligible assets. March 14: Failure to roll repos causes a liquidity crisis at Bear Stearns. Bear Stearns announces $30 billion in funding provided by JP Morgan and backstopped by the government. March 17: JP Morgan announces purchase of Bear Stearns for $2 a share, a little more than $236 million. April 2: New Century files for bankruptcy. June 5: MBIA and Ambac lose their triple A ratings from S&P. June 9: Lehman says it expects to lose $2.8 billion in the quarter ending May 31. June 30: Legg Mason announces another $240 million in capital contributions to support three money market funds. July 11: IndyMac Bank, a large mortgage lender, is seized by federal regulators. The cost to the Federal Deposit Insurance Corporation is estimated to be between $4 billion and $8 billion, potentially a loss of 10 percent of the FDIC’s insurance fund for banks. Freddie Mac and Fannie Mae lost half their value in the week ending July 11. Moody’s and S&P affirm that the U.S. would retain its AAA rating even if forced to rescue Fannie Mae and Freddie Mac. July 14: Federal Reserve Board grants authority to New York Fed to lend to Fannie Mae and Freddie Mac should the need arise. Sources: Various, including Fender and Hördahl (2007), BIS Annual Report 2007-2008, Bloomberg; Financial Times; The Wall Street Journal; BBC (http://news.bbc.co.uk/2/hi/business/7096845.stm), company press releases.

08 Book.indb 242

2/13/09 3:58:37 PM

08 Book.indb 243

Sept. 17, 1998 Feb. 1, 1999 June 18, 1999 Nov. 15, 1999 Dec. 8, 1999 Sept. 7, 2001

Citibank International PLC

Dresdner Kleinwort

Bank of Montreal

Citibank International PLC

N.S.M. Capital Management/ Emirates Bank

Dorada Corp.

K2 Corp.

Links Finance Corp.

Five Finance Corp.

Abacas Investments Ltd.

May 31, 1996

Standard Chartered Bank

Whistlejacket Capital Ltd.

July 10, 2002

Ceres Capital Partners

Societe Generale

Victoria Finance Ltd.

Premier Asset Collateralized Entity Ltd.

Feb. 4, 2002

Standard Chartered Bank

White Pine Corp. Ltd. (merged with Whistlejacket Capital Ltd.)

July 24, 2002

July 10, 2002

Jan. 11, 2002

Bank of Montreal

WestLB

Parkland Finance Corp.

Harrier Finance Funding Ltd.

Sept. 9, 1996

Eiger Capital Management

Citibank International PLC

Orion Finance Corp.

Centauri Corp.

Feb. 2, 1995

Sept. 8, 1989

Citibank International PLC

Gordian Knot Ltd.

Beta Finance Corp.

Initial rating date

Manager/adviser

Sigma Finance Corp.

SIV

8,844.63

4,312.70

13,243.95

7,854.63

12,343.37

3,414.43

1,007.95

12,843.06

22,301.10

29,056.47

12,484.15

21,838.84

2,298.43

52,641.87

20,175.95

Senior debt (mil. $)*

Appendix B Main SIV Outcomes

Insolvent Feb. 15, 08

Moody’s threatens downgrade: S&P affirms

Chapter 11 in April 08

See below under Whistlejacket

Back on Balance Sheet

Back on Balance Sheet

S&P affirms ratings

Back on Balance Sheet

Back on Balance Sheet

Back on Balance Sheet

Back on Balance Sheet

Back on Balance Sheet

Defaulted

Must refinance $20 bil. by Sept.; S&P and Moody’s downgrade

Back on Balance Sheet

Current Status

Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

SEC 6-K 2/20/08

Reuters 6/17/08

Citi Press Release 12/13/07

SEC 6-K 2/20/08

Reuters 2/21/08

Citi Press Release 12/13/07

Citi Press Release 12/13/07

Reuters 1/16/08

Bloomberg 4/8/08

Citi Press Release 12/13/07

Source

The Panic of 2007 243

2/13/09 3:58:38 PM

08 Book.indb 244

Nov. 26, 2002

HSBC Bank PLC

Asscher Finance Ltd.

*As of July 13, 2007, S&P.

Total

Axon Asset Management Inc.

IKB Credit Asset Management GmbH

Axon Financial Funding Ltd.

Rhinebridge PLC

MBIA

Banque AIG

Hudson-Thames Capital Ltd.

Nightingale Finance Ltd.

Citibank International PLC

Vetra Finance Corp.

Sept. 18, 2006

Citibank International PLC

IXIS/Ontario Teachers

Zela Finance Corp.

Cortland Capital Ltd.

Aug. 2, 2006

WestLB/Brightwater Capital

Kestrel Funding PLC

Sept. 23, 2005

May 11, 2007

April 13, 2007

March 30, 2007

March 15, 2007

Dec. 5, 2006

Nov. 15, 2006

Nov. 1, 2006

June 30, 2006

Eaton Vance

HSH Nordbank

Eaton Vance Variable Leveraged Fund

Aug. 3, 2005

July 18, 2005

June 22, 2004

Carrera Capital Finance Ltd.

HSBC Bank PLC

Cheyne Capital Management Ltd.

Cullinan Finance Ltd.

Cheyne Finance PLC

Rabobank International

Citibank International PLC

Tango Finance Corp.

Sedna Finance Corp.

$274,896.99

7,330.00

2,199.63

11,193.76

2,330.23

1,767.33

2,616.94

1,344.19

4,188.70

3,315.86

4,283.48

542.76

9,726.18

35,142.00

14,415.28

14,039.75

Back on Balance Sheet

Defaulted Oct. 07

S&P cuts rating to default

Back on Balance Sheet

Ceased Operations Dec. 07

Back on Balance Sheet

S&P affirmed ratings in Feb.; now on negative watch

Back on Balance Sheet

Back on Balance Sheet

Restructured

Moody’s cuts ratings

Goldman leads restructuring

Back on Balance Sheet

Back on Balance Sheet

Back on Balance Sheet

Appendix B Main SIV Outcomes (continued) Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

Citi Press Release 12/13/07

Reuters 6/17/08

Citi Press Release 12/13/07

Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

Reuters 6/17/08

Citi Press Release 12/13/07

244 Gary B.Gorton

2/13/09 3:58:38 PM

The Panic of 2007

245

Endnotes 1 My use of the phrase “no trade theorem” is an abuse of its original meaning. The “no trade theorem” is the theoretical result that in most circumstances it is not possible for an agent with superior information to profit from trading on that information. See Grossman and Stiglitz (1980) and Milgrom and Stokey (1982). Here, I mean to imply that counterparties assumed their trading partners were better informed and hence refused to trade. “Every banker knows that if he has to prove that he is worthy of credit, however good may be his arguments, in fact his credit is gone.” Walter Bagehot, Lombard Street (1873, chapter II, paragraph II): http://www.econlib.org/Library/Bagehot/bagLom.html.

See Gorton (1984, 1985, 1988) and Gorton and Mullineaux (1987) for discussion of the clearinghouses issue of their own emergency currency. Gorton and Huang (2006) provide a theory. 2

I have described these changes in banking, with various coauthors, including the rise of loan sales and securitization, the use of derivatives, and the regulatory implications of a declining bank charter values. See Gorton and Pennacchi (1989, 1995), Gorton and Souleles (2006), Gorton and Rosen (1995), Gorton (1994). 3

See Gorton (1988), Gorton and Mullineaux (1987), Calomiris and Gorton (1991), and Gorton and Huang (2006). 4

The details are also important in the study of historical panics generally. Little work has been done. Exceptions include, for example, Kelley and Ó Gráda (2000) and Ó Gráda and White (2003). Ó Gráda and White (2003) conclude: “The outcome is partly at variance with the stylized facts of the theoretical literature on banking panics. Banking panics were not characterized by an immediate mass panic of depositors…” (p. 238). Other examples of empirical work include Calomiris and Schweikart (1991), Moen and Tallman (1992), Calomiris and Mason (1997), Richardson (2005), and Richardson and Troost (2005). 5

6 I do not address the issue of bubbles in this paper. Although I have written about bubbles (see Allen and Gorton, 1993), I don’t think we really understand how they start, or are sustained, or why they end. In any case, others are more capable than I on this topic. See, e.g., Shiller (2007) and Case and Shiller (2003).

As Andrew argued a century ago: “The unique dimensions of the recent panic among the experiences of the present generation render important the preservation for future study of all records concerning its phenomena” (1908A, p. 291). 7

See Inside Mortgage Finance, The 2007 Mortgage Market Statistical Annual, Key Data (2006), Joint Economic Committee (October 2007). 8

See Gorton and Souleles (2006) for a discussion of off-balance sheet vehicles and the implicit contracting between investors and vehicle sponsors. 9

08 Book.indb 245

2/13/09 3:58:38 PM

246

Gary B.Gorton

Andrews (1908A), speaking of the Panic of 1907: “As there was no common market for money, there were no regular quotations …” (p. 292). 10

11 A survey of the Panic is provided by the Bank of England (2008). Appendix A of this paper provides a chronology of events.

By “breakdown” I mean that the arbitrage relations between the ABX indices and the underlying cash bonds broke down, as described in Gorton (2008). 12

In fact, the first subprime crisis occurred in 1998 when a number of subprime originators failed. See Temkin et al. (2002) and Moody’s (October 1998). This first crisis did not result in a systemic problem emanating from subprime mortgages, though it was part of the larger Asian and Long Term Capital Management (LTCM) crises. 13

On automated credit evaluation and other technological change in mortgage underwriting see LaCour-Little (2000); Straka (2000); and Gates, Perry, and Zorn (2002). 14

Smith (1998) is a Bank of America national manager of community lending, who was interviewed for the Listokin, et al. study. The citations in the quotations are to that interview. 15

16 Raiter and Parisi (2004) find a significant, nonlinear relationship between FICO scores and coupon differentials: “We find that risk-based pricing has become more rational since 1998. The data show a trend towards greater differentiation in mortgage coupons over time” (p. 1).

Some borrowers in the subprime market may have been “prime” borrowers but without documented income, for example. 17

FICO is a credit score developed by Fair Isaac & Company (http://www.fairisaac.com/fic/en). FICO scores range from 300 to 850. The higher the score, the better the chances of repayment of a loan. 18

The difference between the original balance and the current balance is the amount that has defaulted or has prepaid. The factor is the percentage remaining (current balance divided by original balance). The factor varies from 65.8 percent to 90.5 percent, reflecting differing speeds of prepayment. 19

There are other types of subprime loans, such as hybrid interest-only, 40-year hybrid ARMs, and piggyback second liens. These types are less important quantitatively. 20

21 There is also an option to delay payment, in which case the mortgage becomes delinquent.

The probability of default is also a function of other factors, but I do not include other variables, to ease notation. 22

To ease notation, I will omit the prepayment penalty.

23

08 Book.indb 246

2/13/09 3:58:38 PM

The Panic of 2007

247

There is no hard evidence on this that I know of, but casually, this seems to be the case. The initial bank may have an information advantage over competitors. Gross and Souleles (2002), for example, show the additional explanatory power of bank internal information, over publicly available information like FICO scores, in predicting consumer defaults in credit card accounts. Other evidence concerns the originating bank waiving prepayment fees. For example: “Some lenders may waive the prepayment penalty if you refinance your loan with them and you have held the mortgage for at least one year.” Pena Lending Group, see http://www. penalending.com/cash-out_refinance.html. Or: Mark Ross, president and CEO of Tucson lender Prime Capital Inc.: Prepayment penalties are most often found on subprime loans made to buyers with less-than-perfect credit histories, Ross said. However, some lenders may be willing to waive prepayment penalties to let borrowers refinance, Ross said. See http://www.azstarnet.com/business/226559. However, if a loan is securitized, then the prepayment fee cannot be waived because there is a claimant on that cash flow stream in the RMBS. 24

25 As far as I know, there is no data set which tracks this. LoanPerformance, the mortgage data set for securitized mortgages, is careful not to allow individual lenders to be identified.

Updated estimates provided by Jim Kennedy of the mortgage system presented in “Estimates of Home Mortgage Originations, Repayments, and Debt On Oneto-Four-Family Residences,” Alan Greenspan and James Kennedy, Federal Reserve Board FEDS working paper no. 2005-41. 26

Their data set does not allow them to determine how much was extracted.

27

An interesting question is whether house price increases in some parts of the country were in part caused by the granting of mortgages. Mayer and Pence (2008) is relevant here. 28

Gorton and Souleles (2006) describe the mechanics of securitization.

29

30 A REMIC (Real Estate Mortgage Investment Conduit), shown in the charts, is an investment vehicle, a legal structure that can hold commercial and residential mortgages in trust, and issue securities representing undivided interests in these mortgages. A REMIC can be a corporation, trust, association, or partnership. REMICs were authorized under the Tax Reform Act of 1986.

This is true of securitization generally; see Gorton and Souleles (2007).

31

Two other features are: (1) the clean-up call and (2) compensating interest. (1) The clean-up call gives the owner of the call, generally the residual owner, the option to purchase the remaining bonds in a deal at a predetermined price, when the collateral factor reaches a certain level, i.e., when the deal has amortized down to a sufficiently low level. Normally, the call is to purchase the bonds at par plus accrued interest, when the factor is at or below 10 percent. (2) The day that 32

08 Book.indb 247

2/13/09 3:58:38 PM

248

Gary B.Gorton

a borrower prepays his loan, interest payments on that loan stop. The mortgage servicer, in a non-agency deal, is normally required to compensate investors for this foregone interest, using funds paid to the service as fees. 33 Delinquency triggers are classified as either “soft” or “hard.” The trigger is hit if serious delinquencies, defined as 60+ days, foreclosure, and REO, are at or exceed certain limits. With a soft trigger, the delinquency limit is defined relative to the current amount of senior credit enhancement: the balance of the mezz and subordinate classes, plus OC, expressed as a percentage of the balance of the collateral, e.g., serious delinquencies exceed 50 percent of the senior credit enhancement). With a hard trigger, the delinquency limit is defined as a specific percentage of the current collateral balance, e.g., if serious delinquencies exceed 12 percent of the current balance.

See the prospectus: http://www.secinfo.com/d12atd.z3e6.htm#1stPage.

34

The weighted average rating factor refers to a weighted average rating where ratings have been converted to numbers by a rating agency (in such a way that the ratings are not equidistant apart). Similarly, “correlation factors” refers to rating agency stated correlation assumptions. The details do not concern us here. 35

During the panic, this will be problematic, as the senior investors may choose to liquidate even though they know that the prices are fire sale prices, and their sale will push prices down further, causing another round of marking down—as discussed later. 36

See Moody’s, “Impact of Subprime Downgrades on OC-linked Events of Default in CDOs,” Special Report, November 1, 2007. 37

38 As of January 10, 2008, about $58 billion worth of CDOs have hit “events of default” (EOD) (see Financial Times, January 10, 2008). Moody’s reported on January 7, 2008, that “more than 50 structured CDOs (‘ABS CDOs’) have experienced an Event of Default (‘EOD’) …” (see Moody’s, “Understanding the Consequences of ABS CDO Events of Default Triggered by Loss of Overcollateralization,” Special Report, January 7, 2008).

When investors indicate an interest in investing in a CDO, and even when they invest, the CDO is not completely “ramped up,” that is, all the ABS bonds for the portfolio have not been purchased yet. Investment will be made based on the criteria restricting the portfolio’s composition. 39

I recognize that this is a causal observation. Though I believe this view is widely held by traders, I know of no formal documentation of this. 40

Gorton (2008) discusses negative basis trades in more detail.

41

We do know that these were a source of write-downs for banks. For example, UBS (2008): “Negative Basis Super Seniors: these were Super Senior positions where the risk of loss was hedged through so-called Negative Basis (or “NegBasis”) 42

08 Book.indb 248

2/13/09 3:58:38 PM

The Panic of 2007

249

trades where a counterparty, such as a monoline insurer, provided 100% loss protection. The hedge resulted in a credit exposure towards the protection seller. As of the end of 2007, write-downs on these positions represented approximately 10% of the total Super Senior losses” (p. 14). 43 The difference between total issuance and structured finance issuance would be other categories such as investment–grade loans, high-yield loans, investmentgrade bonds, high–yield bonds, etc.

Synthetic CDOs are not included in the table.

44

The residual category, which has been excluded, consists of market value CDOs. Fully synthetic CDOs are not included. 45

See http://www.markit.com/information/products/abx.html.

46

The rule also restricts the credit quality of the securities that a money market fund may purchase. 47

48 There was a maximum of 30 SIVs that existed, of which 21 were run by 10 banks, including Citigroup, Dresdner, and Bank of Montreal. The approximate size of the SIV sector at its peak was $400 billion in November 2007, having grown from $200 billion three years earlier. See S&P, transcript of teleconference, “Update on U.S. Subprime and Related Matters,” November 1, 2007, http://www2. standardandpoors.com/spf/pdf/media/teleconference_transcript_110107.pdf.

There were market value CDOs, but they died out.

49

The example is simplified with only one mortgage in the subprime RMBS, and only one RMBS tranche in the CDO. This ignores a number of important issues in practice, which need not concern us here. 50

The example does not display the “cliff ” risk that can occur when the CDO contains many tranches of various ABS, RMBS, and CMBS bonds. “Cliff ” risk refers to the phenomenon of a tranche being wiped out quickly once losses reach it. 51

Though note that the investor in a CDO tranche would know the underlying ABS, RMBS, and CMBS bonds, but would not know the underlying portfolios of those instruments. 52

53 When I say “value” I usually mean to compute an expected loss or expected payoff using historical information. “Marking-to-market” is another matter, briefly discussed later.

Leo Tolstoy, Anna Karenina, “Happy families are all alike; every unhappy family is unhappy in its own way.” 54

The calculation is the percentage change in the seasonally adjusted OFHEO repeat-sales house price index for purchase transactions only between the fourth quarters of 2000 and 2005. See www.ofheo.gov/HPLasp. 55

08 Book.indb 249

2/13/09 3:58:38 PM

250

Gary B.Gorton

There are two indices that measure house price appreciation, S&P/Case-Shiller and the OFHEO House Price Index. Both of these indices are based on repeat sales. The two indices differ in important respects. Case-Shiller does not cover the entire U.S., and the omitted areas seem to be doing better than the included areas. Case-Shiller omits 13 states altogether and has incomplete coverage of 29 other states (see Leventis, 2007). The OFHEO index is not value-weighted and only includes homes with conforming mortgages. 56

57 The United States has not experienced a large, nationwide decline in house prices since the Great Depression of the 1930s. In 1940 the median nonfarm housing value was 48.6 percent below the 1930 median value (based on the 1940 Housing Census). Over the same decade, the Consumer Price Index had fallen 17.4 percent and food prices had fallen 27 percent. In other words, even adjusting for the deflation during the period, housing prices had not recovered to the levels at the beginning of the Depression by 1940. See Fishback, Horrace, and Kantor (2001).

The trustees for transactions make monthly reports known as remittance reports. Remittance reports detail scheduled and unscheduled remittances of principal, servicer advances, loan repurchases, realized losses, delinquencies, and so on. 58

PAUG is a form of settlement used in asset-backed CDS. It allows two-way payments between the protection buyer and protection seller during the life of the contract. If the reference obligation is affected by interest shortfalls or principal write-downs, the protection buyer compensates the protection seller. These amounts are paid back to the protection buyer if the interest shortfalls or principal writedowns are reversed. The protection buyer has the option of physically settling the CDS if there is a principal write-down. 59

This is related to some ideas of Grossman (1988) about the 1987 stock market crash. Grossman argues that portfolio insurance, in synthetically creating a put option, does not reveal to market participants the amount of such puts outstanding, something that would be known if actual put options were traded. 60

The initial coupons for the BBB- and AAA tranches are shown below:

61

ABX-HE BBB -

Coupon (bps)

2006-1

267

2006-2

242

2007-1

389

2007-2

500

ABX-HE AAA 2006-1

18

2006-2

11

2007-1

9

2007-2

76

Source: Markit

08 Book.indb 250

2/13/09 3:58:38 PM

The Panic of 2007

251

“… as the ABX has widened and gone down in price, on some bad fundamental news, we’ve gotten quite a nice mark to market benefit on that move,” Ralph Cioffi, manager of the Bear Stearns hedge funds that subsequently were liquidated; Bear Stearns Investor Conference Call, April 25, 2007. 62

63 The “super SIV” was the Master-Liquidity Enhancement Conduit (M-LEC), which was an attempt to create the incentive-compatible structure of a 19th century clearinghouse, but failed. See The Economist, October 18, 2007, “Curing SIV,” http://www.economist.com/displaystory.cfm?story_id=9993423.

For background on repurchase agreements (repo), see Bank for International Settlements (1999). 64

See Mishkin (February 15, 2008) and Taylor and Williams (2008A, B). Libor stands for “London interbank offered rate.” It is the most widely used benchmark for short-term interest rates in major currencies worldwide. Libor is compiled, for ten currencies over a range of maturities from overnight to twelve months, by the British Bankers’ Association (BBA) and is published daily between 11:00 a.m. and 12 noon London time. Libor rates are averages of interbank rates submitted by a panel of banks. For each currency, panels comprise at least eight contributor banks. Sterling, dollar, euro, and yen panels contain sixteen banks. See http://www.bba. org.uk/bba/jsp/polopoly.jsp?d=141. OISs are interest rate swaps in which the floating leg is linked to a published index of daily overnight rates. The two parties agree to exchange at maturity, on an agreed notional amount, the difference between interest accrued at the agreed fixed rate and interest accrued through the geometric average of the floating index rate. 65

The question of what the spread represents is addressed by Taylor and Williams (2008A, B) and Michaud and Upper (2008). I do not pursue this here. 66

In fact, there is a general question concerning double-entry bookkeeping as a paradigm in a world of derivatives. 67

I know of no direct evidence on either of these issues.

68

Many banks had implemented it earlier, in anticipation of the rule coming into effect. 69

Statement 157 defines “fair value” as: “The price that would be received to sell an asset or paid to transfer a liability in an orderly transaction between market participants at the measurement date.” See Statement of Financial Standards No. 157 http://www.fasb.org/pdf/fas157.pdf . 70

See Gorton, He, and Huang (2008) and Plantin, Sapra, and Shin (2006) for discussions. 71

Haldeman (2007) provides background, dating back to Enron.

72

08 Book.indb 251

2/13/09 3:58:38 PM

252

Gary B.Gorton

See Committee of European Banking Supervisors (2008) and Bond Market Association and the American Securitization Forum (2006) for descriptions of the marking process and the data inputs. 73

74 Obviously, this would not occur if there was another side to this market. But, investors are the very agents facing asymmetric information.

On trends in the use of collateral, also see BIS (2001).

75

See http://www.sec.gov/Archives/edgardata/1174735/000119312507138443/ dsc14d9.htm#rom81455_10. 76

CSAs are used in documenting collateral arrangements between two parties that trade privately negotiated (over-the-counter) derivative securities. The trade is documented under a standard contract called a master agreement, developed by the International Swaps and Derivatives Association (ISDA). The two parties must sign the ISDA master agreement and execute a credit support annex before they trade derivatives with each other. See, also, ISDA “2005 ISDA Collateral Guidelines,” http://www.isda.org/publications/pdf/2005isdacollateralguidelines.pdf. 77

Keep in mind that long credit derivative positions cannot be delivered to the discount window. 78

79 In triparty repo, a custodian bank or clearing organization acts as an intermediary between the two repo parties. There is no data that I know of that quantifies the amount of bilateral repo.

Private communication from a repo trader.

80

See, e.g., Bernanke (2008); Wellink (2007), President of the Netherlands Bank and Chairman of the Basel Committee on Banking Supervision; Knight (2008), General Manager of the BIS; Gieve (2008), Deputy Governor of the Bank of England. 81

See, e.g., Gorton (1988); Berger, Kashyap, and Scalise (1995); and Boyd and Gertler (1994) for some discussion of these trends. 82

See Benveniste, Singh, and Wilhelm (1993) for a description of this competition.

83

For the sake of space I do not review these developments.

84

Eighty subprime mortgage lenders have exited the business since the end of 2006—many going bankrupt (see Worth Civils and Mark Gongloff, “Subprime Shakeout,” Wall Street Journal online, http://online.wsj.com/public/resources/documents/info-subprimeloans0706-sort.html. 85

Mortgage servicing rights may also be securitized.

86

Note that losses can exceed exposures due to the timing of the numbers. Net losses are for the year ending December 28, while net exposure is for December 29. 87

In addition, the sponsors hold the residuals of the securitizations.

88

08 Book.indb 252

2/13/09 3:58:38 PM

The Panic of 2007

253

References Allen, Franklin, and Gary Gorton (1993), “Churning Bubbles,” Review of Economic Studies, Vol. 60(4), p. 813-836. Andrew, A. Piatt (1908A), “Hoarding in the Panic of 1907,” Quarterly Journal of Economics, Vol. 22, No. 4, p. 497-516. Andrew, A. Piatt (1908B), “Substitutes for Cash in the Panic of 1907,” Quarterly Journal of Economics, Vol. 22, No. 2, p. 290-299. Bank of England (2008), Financial Stability Report (April), Issue No. 23. Bank for International Settlements (1999), “Implications of Repo Markets for Central Banks,” Report of a Working Group established by the Committee on the Global Financial System of the Central Banks of the Group of Ten Countries, March 9, 1999. Bank for International Settlements (2001), “Collateral in Wholesale Financial Markets: Recent Trends, Risk Management and Market Dynamics,” Report prepared by the Committee on the Global Financial System Working Group on Collateral (March 2001). Basel Committee on Banking Supervision, Bank for International Settlements (2008), The Joint Forum, “Credit Risk Transfer,” Consultative Document (April). Bear Stearns, “Bear Stearns Quick Guide to Non-Agency Mortgage-Backed Securities,” September 2006A. Bear Stearns, “RMBS Residuals: A Primer,” September 2006B. Benveniste, Lawrence, Manoj Singh, and William Wilhelm (1993), “The Failure of Drexel Burnham Lambert: Evidence on the Implications for Commercial Banks,” Journal of Financial Intermediation, Vol. 3, p. 104-137. Berger, Allen, Anil Kashyap, and Joseph Scalise (1995), “The Transformation of the U.S. Banking Industry: What a Long Strange Trip It’s Been,” Brookings Papers on Economic Activity, Vol. 2, p. 55-218. Bernanke, Ben (2008), speech “Addressing Weaknesses in the Global Financial Markets: The Report of the President’s Working Group on Financial Markets,” at the World Affairs Council of Greater Richmond’s Virginia Global Ambassador Award Luncheon, Richmond, Virginia, April 10, 2008. Bhardwaj, Geetesh, and Rajdeep Sengupta (2008A), “Prepaying Subprime Mortgages,” Federal Reserve Bank of St. Louis, working paper. Bhardwaj, Geetesh, and Rajdeep Sengupta (2008B), “Where’s the Smoking Gun? A Study of Underwriting Standards for U.S. Subprime Mortgages,” Federal Reserve Bank of St. Louis, working paper.

08 Book.indb 253

2/13/09 3:58:38 PM

254

Gary B.Gorton

Bond Market Association and the American Securitization Forum (2006), “An Analysis and Description of Pricing and Information Sources in the Securitized and Structured Finance Markets,” October 2006. Bond Market Association (2005), “Repo & Securities Lending Survey of U.S. Markets Volume and Loss Experience,” Research (January). Bordo, Michael (1986), “Financial Crises, Banking Crises, Stock Market Crashes and the Money Supply: Some International Evidence, 1870-1933,” in Financial Crises and the World Banking System (The MacMillan Press: London), edited by Forrest Capie and Geoffrey Wood, p. 190-248. Bordo, Michael (1985), “The Impact and International Transmission of Financial Crises: Some Historical Evidence, 1870-1933,” Revista di Storia Economica 2, p. 41-78. Boyd, John, and Mark Gertler (1994), “Are Banks Dead? Or Are the Reports Greatly Exaggerated?” in Proceedings of a Conference on Bank Structure and Competition, Federal Reserve Bank of Chicago, p. 85-117. Calomiris, Charles, (1993), “Regulation, Industrial Structure, and Instability in U.S. Banking: An Historical Perspective,” in Structural Change in Banking (Business One Irwin: Homewood, Illinois), edited by Michael Klausner and Lawrence White, 19-116. Calomiris, Charles and Joseph Mason (1997), “Contagion and Bank Failures During the Great Depression: The June 1932 Chicago Banking Panic,” American Economic Review, Vol. 87(5), p. 863-883. Calomiris, Charles, and Joseph Mason (2004), “Credit Card Securitization and Regulatory Arbitrage,” Journal of Financial Services Research, Vol. 26(1), p. 5-27. Calomiris, Charles, and Larry Schweikart (1991) “The Panic of 1857: Origins, Transmission, and Containment,” Journal of Economic History, Vol. 51(4), p. 807-834. Calomiris, Charles, and Gary Gorton (1991), “The Origins of Banking Panics: Models, Facts, and Bank Regulation,” in Financial Markets and Financial Crises, ed. Glenn Hubbard (University of Chicago Press). Canner, Glenn, and Wayne Passmore (1999), “The Role of Specialized Lenders in Extending Mortgages to Lower-Income and Minority Homebuyers,” Federal Reserve Bulletin (November), p. 709-723. Case, Karl, and Robert Shiller (2003), “Is There a Bubble in the Housing Market?” Brookings papers on Economic Activity, Vol. 2, p. 299-362. Chen, Weitzu, Chi-Chun Liu, and Stephen Ryan (2007), “Characteristics of Securitizations that Determine Issuers’ Retention of Risks on the Securitized Assets,” http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1077798.

08 Book.indb 254

2/13/09 3:58:38 PM

The Panic of 2007

255

Chomsisengphet, Souphala, and Anthony Pennington-Cross (2007), “Subprime Refinancing: Equity Extraction and Mortgage Termination,” Real Estate Economics Vol. 35, no. 2, p. 233-263. Chomsisengphet, Souphala, and Anthony Pennington-Cross (2006), “The Evolution of the Subprime Mortgage Market,” Federal Reserve Bank of St. Louis Review, January/February 2006. Citibank, “A Simple Guide to Subprime Mortgages, CDOs, and Securitization,” April 13, 2007. Committee of European Banking Supervisors (2008), “Report on Issues Regarding the Valuation of Complex and Illiquid Financial Instruments,” June 18, 2008. Demyanyk, Yuliya, and Otton Van Hemert (2007), “Understanding the Subprime Mortgage Crisis,” (December 10), Stern School of Business, New York University, working paper. Drucker, Steven, and Manju Puri (2007), “On Loan Sales, Loan Contracting, and Lending Relationships,” Duke University, working paper. Economist, The (2008), “Don’t Mark to Markit,” May 6, 2008; http://www.economist.co.uk/finance/displaystory.cfm?story_id=10809435. Euromoney (2008), “Understanding the Mark-To-Market Meltdown,” March 3, 2008. European Central Bank, (2007), Financial Stability Review. Farris, John, and Christopher Richardson (2004), “The Geography of Subprime Mortgage Prepayment Penalty patterns,” Housing Policy Debate Vol. 15, No. 3, p. 687-714. Fender, Ingo, and Peter Hördahl (200&), “Overview: Credit Retrenchment Triggers Liquidity Squeeze,” BIS Quarterly Review (September), p. 1-16. Fishback, Price, William Horrace, and Shawn Kantor (2001), “The Origins of Modern Housing Finance: The Impact of Federal Housing Programs During the Great Depression,” University of Arizona, working paper. Fitch Ratings (2008), “Fair Value Accounting: Is It Helpful in Illiquid Markets?” Accounting Research Special Report (April 28, 2008). Fitch Ratings (2007), “The Impact of Poor Underwriting Practices and Fraud in Subprime RMBS Performance,” U.S. Residential Mortgage Special Report (November 28, 2007). Frankel, Allen, “Prime or Not so Prime? An Exploration of U.S. Housing Finance in the New Century,” Bank for International Settlements, Quarterly Review, March 2006.

08 Book.indb 255

2/13/09 3:58:38 PM

256

Gary B.Gorton

Gates, Susan, Vanessa Perry, and Peter Zorn (2002), “Automated Underwriting in Mortgage Lending: Good News for the Underserved?” Housing Policy Debate, Vol. 13, No. 2, p. 369-391. Geithner, Timothy (2008), Remarks at The Economic Club of New York, New York City, June 9, 2008. Gieve, John (2008), “The Return of the Credit Cycle: Old Lessons in New Markets,” speech at the Euromoney Bond Investors Congress, February 27, 2008; http://www.bankofengland.co.uk/publications/speeches/2008/speech338.pdf. Gordon, Brian (2008), “Hedges in the Warehouse: The Banks Get Trimmed,” Chicago Fed Letter, Federal Reserve Bank of Chicago, April 2008, Number 249. Gorton Gary (2008), “Information, Liquidity, and the Panic of 2007,” American Economic Review, Papers and Proceedings, forthcoming. Gorton, Gary (1994), “Bank Regulation When ‘Banks’ and ‘Banking’ Are Not the Same,” Oxford Review of Economic Policy, Vol. 10, No. 4, p. 106-119. Gorton, Gary, and George Pennacchi (1989), “Are Loan Sales Really Off-Balance Sheet?” Journal of Accounting, Auditing and Finance 4:2 (Spring), p. 125-45. Gorton, Gary (1988), “Banking Panics and Business Cycles,” Oxford Economic Papers 40 (December 1988), p. 751-81. Gorton, Gary (1985), “Clearinghouses and the Origin of Central Banking in the U.S.,” Journal of Economic History 45:2, p. 277-83. Gorton, Gary (1984), “Private Bank Clearinghouses and the Origins of Central Banking,” Business Review, Federal Reserve Bank of Philadelphia (January-February 1984), p. 3-12. Gorton, Gary, and Lixin Huang (2006), “Banking Panics and Endogenous Coalition Formation,” with Lixin Huang, Journal of Monetary Economics, Vol. 53(7), p. 1613-1629. Gorton, Gary, and Nicholas S. Souleles (2006), “Special Purpose Vehicles and Securitization,” chapter in The Risks of Financial Institutions, edited by Rene Stulz and Mark Carey (University of Chicago Press). Gorton, Gary, and Don Mullineaux (1987), “The Joint Production of Confidence: Endogenous Regulation and Nineteenth Century Commercial Bank Clearinghouses,” Journal of Money, Credit and Banking 19(4), p.458-68. Gorton, Gary, and George Pennacchi (1995), “Banks and Loan Sales: Marketing Non-Marketable Assets,” Journal of Monetary Economics 35(3), p. 389-411. Gorton, Gary, and George Pennacchi (1989), “Are Loan Sales Really Off-Balance Sheet?” with George Pennacchi, Journal of Accounting, Auditing and Finance 4:2 (Spring 1989), p. 125-45.

08 Book.indb 256

2/13/09 3:58:38 PM

The Panic of 2007

257

Gorton, Gary, and Richard Rosen (1995), “Banks and Derivatives,” National Bureau of Economic Research Macroeconomics Annual 1995 (MIT Press). Gorton, Gary, Ping He, and Lixin Huang (2008), “Monitoring and Manipulation: Asset Prices When Agents are Market-to-Market,” working paper. Gorton, Gary, and Andrew Winton (2000), “Liquidity Provision and the Social Cost of Bank Capital,” working paper. Greenspan, Alan, and James Kennedy (2007), “Sources and Uses of Equity Extracted from Homes,” Board of Governors of the Federal Reserve System, Finance and Economics Discussion Series, Working Paper #2007-20. Greenspan, Alan, and James Kennedy (2005), “Estimates of Home Mortgage Originations, Repayments, and Debt on One-to-Four Family Residences,” Board of Governors of the Federal Reserve System, Finance and Economics Discussion Series, Working Paper #2005-41. Gross, David, and Nicholas Souleles (2002), “An Empirical Analysis of Personal Bankruptcy and Delinquency,” Review of Financial Studies, Vol. 15, p. 319 – 347. Grossman, Sanford (1988), “An Analysis of the Implications for Stock and Futures Price Volatility of Program Trading and Dynamic Trading Strategies,” Journal of Business, Vol. 61, p. 275-298. Grossman, Sanford, and Joseph Stiglitz (1980), “On the Impossibility of Informationally Efficient Markets,” American Economic Review 70, p. 393-408. Guerrera, Francesco, and Deborah Brewster (April 30, 2008), “Pimco Scouts for Wall St Cast-offs,” Financial Times: http://www.ft.com/cms/s/0/22861bb0-16f011dd-bbfc-0000779fd2ac.html. Haldeman, Robert (2007), “Fact, Fiction, and Fair Value Accounting at Enron,” The CPA Journal (August 25, 2007). Hester, Elisabeth, and Katherine Bureton (June 19, 2008), “Hedge Funds Hire from Wall Street as Jobs Disappear, Pay Falls,” Bloomberg.com: http://www. bloomberg.com/apps/news?pid=20601087&sid=avItj3PPwSGk&refer=home. International Monetary Fund (April 2008), “Containing Systemic Risks and Restoring Financial Soundness,” Global Financial Stability Report. International Swap Dealers Association (ISDA) (2007), ISDA Margin Survey 2007. International Swap Dealers Association (ISDA) (2005), 2005 ISDA Collateral Guidelines. Jiangli, Wenying and Matt Pritsker (2008), “The Impacts of Securitization on U.S. Bank Holding Companies,” Board of Governors of the Federal Reserve System, working paper.

08 Book.indb 257

2/13/09 3:58:38 PM

258

Gary B.Gorton

Joint Economic Committee of the U.S. Congress, “The Subprime Lending Crisis,” October 2007. Joint Forum, The (2008), “Credit Risk Transfer: Developments from 2005 to 2007,” April 2008. Kau, J.B., and D.C. Keenan (1995), “An Overview of the Option Theoretic Pricing in Mortgages,” Journal of Housing Research Vol. 6(2), p. 217-244. Kelley, Morgan and Cormac Ó Gráda (2000), “Market Contagion: Evidence from the Panics of 1854 and 1857,” American Economic Review, Vol. 90(5), p. 1110-1124. Kemmerer, E.W. (1911), “American Banks in the Times of Crisis under the National Banking System,” Proceedings of the Academy of Political Science in the City of New York, Vol. 1(2), p. 233-253. Kevin Kendra, Fitch, “Tranche ABX and Basis Risk in Subprime RMBS Structured Portfolios,” February 20, 2007. See http://www.fitchratings.com/web_content/sectors/subprime/Basis_in_ABX_TABX_Bespoke_SF_CDOs.ppt. Kiff, John, and Paul Mills (2007), “Lessons from Subprime Turbulence,” IMF Survey Magazine, August 23. Kiff, John, and Paul Mills (2007), “Money for Nothing and Checks for Free: Recent Developments in U.S. Subprime Mortgage Markets,” July, IMP Working paper #07-188. Knight, Malcolm (2008), “Some Reflections on the Future of the Originate-toDistribute Model in the Context of the Current Financial Turmoil,” speech at the Euro 50 Group Roundtable, London, April 21, 2008; http://www.bis.org/ speeches/sp080423.htm. Kohlbeck, Mark, and Terry Warfield (2002), “The Role of Unrecorded Intangible Assets in Residual Income Valuation: The Case of Banks,” see http://papers.ssrn. com/sol3/papers.cfm?abstract_id=296387&download=yes. LaCour-Little, Michael (2000), “The Evolving Role of Technology in Mortgage Finance,” Journal of Housing Research Vol. 11(3), p. 173-205. Leventis, Andrew (2007), “A Note on the Differences between the OFHEO and S&P/Case-Shiller House Prices Indexes,” Office of Federal Housing Enterprise Oversight, working paper. Listokin, David, Elvin Wyly, Larry Keating, Kristopher Rengert, and Barbara Listokin (2000), “Making New Mortgage Markets: Case Studies of Institutions, Home Buyers, and Communities,” Fannie Mae Foundation Research Report. Lonski, John (2008), “Market Breakdowns Prompt Fastest Bank C&I Growth Since 1973,” Moody’s Credit Trends, January 8, 2008.

08 Book.indb 258

2/13/09 3:58:39 PM

The Panic of 2007

259

Mansfield, Cathy Lesser (2000), “The Road to Subprime: ‘HEL’ Was Paved with Good Congressional Intentions: Usury Deregulation and the Subprime Home Equity Market,” South Carolina Law Review 51 (Spring). Marcus, Alan (1984), “Deregulation and Bank Financial Policy,” Journal of Banking and Finance 8 (1984), p. 557-565. Mayer, Chris, and Karen Pence (2008), “Subprime Mortgages: What, Where, and to Whom?” Board of Governors, Federal Reserve System, Working Paper #2008-29. McDermott, Gail, Leslie Albergo, and Natalie Abrams (2001), “NIMs Analysis: Valuing Prepayment Penalty Fee Income,” Standard & Poor’s, New York, January 3, 2001. Michaud, François-Louis, and Christian Upper (2008), “What Drives Interbank Rates? Evidence from the Libor Panel,” BIS Quarterly Review (March), p. 47-58. Milgrom, Paul, and Nancy Stokey (1982), “Information, Trade and Common Knowledge,” Journal of Economic Theory 26, p. 17-27. Miller, Merton (1986), “Financial Innovation: The Last Twenty Years and the Next,” Journal of Financial and Quantitative Analysis, Vol. 21(4), p. 459-471. Mishkin, Frederic (2008), Speech at the U.S. Policy Forum, New York, New York, February 29, 2008, “On Leveraged Losses: Lessons from the Mortgage Meltdown.” Mishkin, Frederic (2008), Speech at the Tuck Global Capital Markets Conference, Tuck School of Business, Dartmouth College, Hanover, New Hampshire, February 15, 2008, “The Federal Reserve’s Tools for Responding to Financial Disruptions.” Missal, Michael (2008), “Final Report of Michael Missal, Bankruptcy Court Examiner,” United States Bankruptcy Court of the District of Delaware, In re: New Century Holdings, Inc. Chapter 11, Case No. 07-10416 (KJC), February 29, 2008. Moen, Jon, and Ellis Tallman (1992), “The Bank Panic of 1907: The Role of Trust Companies,” Journal of Economic History 52, p. 611–30. Moody’s Investors Service (2008), “Understanding the Consequences of ABS CDO Events of Default Triggered by Loss of Overcollateralization,” Structured Finance Special Report, January 7, 2008. Moody’s Investors Service (2007), “Challenging Times for the U.S. Subprime Mortgage Market,” Structured Finance Special Report, March 7, 2007. Moody’s Investors Service (2007), “Impact of Subprime Downgrades on OCLinked Events of Default in CDOs,” Structured Finance Special Report, November 1, 2007. Moody’s Investors Service (2006), “U.S. RMBS: Evaluating Alternative Performance Triggers,” Structured Finance Special Report, September 26, 2006.

08 Book.indb 259

2/13/09 3:58:39 PM

260

Gary B.Gorton

Moody’s Investors Service (2006), “Moody’s Cashflow Assumptions for RMBS Alt-A Transactions,” Structured Finance Special Report, February 9, 2006. Moody’s Investors Service (2005), “Tranching Senior Classes in U.S. Overcollateralization Structures: A Distinct Loss Position May Warrant a Distinct Rating,” Structured Finance Special Report, May 12, 2005. Moody’s Investors Service (2003), “Structural Nuances in RMBS,” Structured Finance Special Report, May 30, 2003. Moody’s Investors Service (2003), “The Fundamentals of Asset-Backed Commercial Paper,” Structured Finance Special Report, February 3, 2003. Moody’s Investors Service (2002), “Protected for Life? Weak Step-Down Triggers May Add Vulnerability in Some Home Equity Securitizations,” Structured Finance Special Report, November 22, 2002. Moody’s Investors Service (2002), “An Introduction to Structured Investment Vehicles,” International Structured Finance Special Report, January 25, 2002. Moody’s Investors Service (1998), “Subprime Home Equity: The Party’s Over,” Global Credit Research, October 1998. Nash, William (1908), “Clearing-House Certificates and the Need for a Central Bank,” Annals of the American Academy of Political and Social Science, Vol. 31, Lessons of the Financial Crisis (March), p. 61-66. Norris, Floyd (2007), “Reading the Tea Leaves of Financial Statements,” New York Times, November 9, 2007. Ó Gráda, Cormac, and Eugene White (2003), “The Panics of 1854 and 1857: A View from the Emigrant Industrial Savings Bank,” Journal of Economic History, Vol. 63(1), p. 213-240. Pagano, Marco (1989), “Endogenous Market Thinness and Stock Price Volatility,” Review of Economic Studies 56, p. 269-288. Plantin, Guillaume, Haresh Sapra, and Hyun Song Shin, 2006, “Marking to Market: Panacea or Pandora’s Box?” Princeton University, working paper. Pollock, Alex (2008), “Conflicted Agents and Platonic Guardians: Interview with Alex Pollock,” American Enterprise Institute Interview, posted May 15, 2008: http://www.aei.org/publications/filter.all,pubID.28012/pub_detail.asp. President’s Working Group on Financial Markets (March 2008), “Policy Statements on Financial Market Developments.” Raiter, Frank, and Francis Parisi (2004), “Mortgage Credit and the Evolution of Risk-Based Pricing,” Joint Center for Housing Studies, Harvard University, BABC 04-23.

08 Book.indb 260

2/13/09 3:58:39 PM

The Panic of 2007

261

Richardson, Gary (2005), “Bank Distress During the Great Contraction, 1929 to 1933, New Evidence from the Archives of the Board of Governors,” University of California-Irvine, Department of Economics, Working Paper. Richardson, Gary, and William Troost (2005), “Monetary Intervention Mitigated Banking Panics During the Great Depression: Quasi-Experimental Evidence from the Federal Reserve District Border in Mississippi, 1929 to 1933,” University of California-Irvine, Department of Economics, Working Paper. Rosengren, Eric (2007), “Subprime Mortgage Problems: Research, Opportunities, and Policy Considerations,” speech by Eric Rosengren at The Massachusetts Institute for a New Commonwealth, December 3, 2007; see http://www.bos.frb. org/news/speeches/rosengren/2007/120307.htm. Securities Industry and Financial Markets Association (2008), Research Quarterly, May. Shiller, Robert (2007), “Understanding Recent Trends in House Prices and Homeownership,” Kansas City Federal Reserve Bank, Jackson Hole Conference Proceedings. Silber, William (2007), When Washington Shut Down Wall Street: The Great Financial Crisis of 1914 and the Origins of America’s Monetary Supremacy (Princeton University Press; Princeton, New Jersey). Sprague, O.M.W. (1915), “The Crisis of 1914 in the United States,” American Economic Review, Vol. 5(3), p. 499-533. Standard and Poor’s (2008), “Is It Time to Write Off Fair Value?,” May 27, 2008. Standard and Poor’s (2007), “Marking to Market When There is No Market,” October 15, 2007. Standard and Poor’s (2007), “Structured Investment Vehicles: Under Stormy Skies, An Updated Look at the Weather,” Ratings Direct, August 30, 2007. Standard and Poor’s (2003), “Structured Investment Vehicle Criteria: New Developments,” September 4, 2003. Straka, John (2000), “A Shift in the Mortgage Landscape: The 1990s Move to Automated Credit Evaluations,” Journal of Housing Research, Vol. 11, N0. 2, 207-232. Taylor, John, and John Williams (2008A), “A Black Swan in the Money Market,” Federal Reserve Bank of San Francisco, working paper. Taylor, John, and John Williams (2008B), “Further Results on a Black Swan in the Money Market,” Federal Reserve Bank of San Francisco, working paper. Temkin, Kenneth, Jennifer Johnson, Diane Levy (2002), “Subprime Markets, the Role of GSEs, and Risk-Based Pricing,” U.S. Department of Housing and Urban Development, March 2002.

08 Book.indb 261

2/13/09 3:58:39 PM

262

Gary B.Gorton

Tran, Kiet (no date), “The Sub-Prime Mortgage Market – A Rough Storm Ahead?” See http://www.markit.com/information/news/commentary/Structured-Finance contentParagraphs/014/document/20071106%20ABX.pdf . UBS (2008), “Shareholder Report on UBS’s Write-Downs,” April 18, 2008. UBS, “Mortgage Strategist,” various issues: June 26, 2007; July 31, 2007; November 27, 2007; October 23, 2007; December 11, 2007. UBS, “Q-Series: Global Banking Crisis,” November 29, 2007. U.S. Treasury (2008), “Blueprint for a Modernized Financial Regulatory Structure,” The Department of Treasury, March. Vanderlip, Frank (1908), “The Panic as a World Phenomenon,” Annals of the American Academy of Political and Social Science, Vol. 31, Lessons of the Financial Crisis, (March), p. 2-7. Wellink, Nout (2007), “Risk Management and Financial Stability—Basel II and Beyond,” Speech at the GARP 2007 8th Annual Risk Management Convention and Exhibition, February 27, 2007; http://www.bis.org/review/r070228a. pdf?noframes=1. Weston, N.A. (1922), “The Studies of the National Monetary Commission,” Annals of the American Academy of Political and Social Science, Vol. 99, p. 17-26. Zelmanovich, Mark, Quincy Tang, Sharon McGarvey, and Bernard Maas (2007), “RMBS NIMS: An Overview,” Journal of Structured Finance (Spring), p. 65- 68. Zigas, Barry, Carol Parry, and Paul Weech (2002), “The Rise of Subprime Lending: Causes, Implications, and Proposals,” Fannie Mae working paper. Zingales, Luigi (2007), “Is the U.S. Capital Market Losing its Competitive Edge?” European Corporate Governance Institute, Finance Working paper No. 192/2007.

08 Book.indb 262

2/13/09 3:58:39 PM

Commentary: The Panic of 2007 Bengt Holmström

Professor Gorton has given us an exceptionally informative picture of the nuclear core of the financial crisis­—the fragile structure of subprime mortgage-backed securitization and the chain of derivatives built on top of it. Those who have the patience to read a long paper filled with lots of important details will be rewarded with a much better understanding and appreciation of the reasons why we fell into the current abyss. It is also an invaluable account for those charged with cleaning up the mess, including regulators, as well as generations of economists who will study the crisis. I want to talk about two questions related to the crisis: (i) In what sense were mortgage-backed securities (MBSs) the cause of the crisis? (ii) Why did we get into subprime lending and securitization in the first place? My main focus will be on the first question, since it is most closely related to the paper and, I think, most relevant for changes in regulation. In what sense were MBSs the cause of the crisis? The paper’s main thesis is that a large part of the problems that emerged in the subprime mortgage markets was due to excessively complex contracting, which obscured the value of the underlying mortgage assets. Unlike traditional mortgages held on the originating bank’s balance sheet, subprime mortgages were increasingly brought 263

08 Book.indb 263

2/13/09 3:58:39 PM

264

Bengt Holmström

to market using an originate-and-distribute model (the fraction of securitized subprime mortgages reached 80% in 2006). Mortgages were pooled into a securitized asset, an MBS, that was sold off in tranches, with each tranche defined by its rights to interest and principal using complex payout rules that depended on how the underlying mortgage portfolio performed. The tranches were created to serve clienteles with different risk appetites. Tranches had ratings ranging from AAA (the highest) to BBB (the lowest). Naturally, the MBS was more valuable at issue if it could push ratings higher. The fact that the credit worthiness of the tranches could be carefully tailored to meet rating requirements changed the role of the rating agencies in a potentially fateful way. Instead of rating tranches ex post, the agencies were often brought in at the design stage to consult on what it would take to pass the bar for a desired rating level. At the time of issue, therefore, an AAA tranche was not an average AAA bond, but a marginal one. The rating game may have misled some market participants, and it certainly softened regulatory requirements on capital adequacy. The subprime MBS was a complex instrument in its own right and challenging to value. But the degree of complexity and opaqueness was vastly increased as the various tranches were sold off and pooled with other MBSs through a chain of structured securities, including those of CDOs (collateralized debt obligations), SIVs (structured investment vehicles), ABCP conduits (asset-backed commercial paper conduits) and other investment vehicles. At each stage, there was “loss of information” about the underlying mortgage assets. At the end of it, minute pieces of the original mortgages had been strewn around hundreds, if not thousands, of structured instruments, making it impossible to trace the value of the fundamental assets, and, in that way, price the structured products. Complexity and opacity are enormous problems in the subprime– related markets today. When housing prices started falling, the subprime–related securities lost value rapidly, and most tranches, including the AAA-rated ones, became sensitive to information about the underlying assets. Without a good way to value the assets, information

08 Book.indb 264

2/13/09 3:58:39 PM

Commentary

265

asymmetries led to serious adverse selection problems and loss of liquidity. Trading ground to a halt in key markets as subprime-related instruments became “toxic” and lost their ability to serve as collateral. There seems to be wide agreement that the proximate cause of the crisis was the sudden loss of collateral quality in liquidity-providing markets. The paper does a great job explaining the rise of complexity and lack of transparency in these markets, but it asserts more than explains the abrupt change in liquidity. Let me try to elaborate on this link in the argument, because it has important consequences for regulatory changes. The first thing to note is that that complexity and opacity need not cause market illiquidity. The initial success of subprime securitization is evidence in point. These markets were liquid for years before problems started to emerge. In fact, they positively thrived, with volumes growing exponentially until the sudden crash. People may have been misinformed. Indeed, the synthetic subprime mortgage index, ABX, which started trading in 2006, signaled a much higher default risk than the markets had priced in. Gorton argues that this revelation triggered the collapse of the structured products markets. If so, more accurate information about the underlying assets­—more transparency—made markets less liquid. To give another example, consider the way de Beers sells wholesale diamonds. They place the diamonds in packets that buyers are forbidden to explore. The packets are sold based on their gross attributes on a take-it-or-leave-it basis. If buyers were allowed to look into the packets first to get a better estimate of their value, packets left behind would become tainted. Inspection would slow down trade and might even prevent trade entirely.1 Placing diamonds in packets eliminates adverse selection problems among buyers. Similar problems between de Beers and the buyers are, in turn, eliminated by de Beers’ concern for its reputation, supported by its rents from the monopoly and from repeat buying. Again, far from being a problem, lack of transparency enhances market liquidity. The shared feature in these examples is symmetric information. Symmetric information about payoffs promotes liquidity. Adverse

08 Book.indb 265

2/13/09 3:58:39 PM

266

Bengt Holmström

selection thrives on asymmetric information, but contrary to what one might think, increased transparency need not reduce information asymmetries and hence may decrease liquidity. Of course, total transparency, where all payoff relevant information is made available to all traders at the same time, keeps information symmetric and markets liquid. But so does complete ignorance. It is partial transparency that is problematic, and adverse selection may increase or decrease with more transparency. I put the word payoffs above in italics, because the sensitivity of payoffs to information is the second critical element in understanding the breakdown of the subprime markets. Financial instruments vary with regard to how sensitive they are to information. Consider the trivial case where a claim on an asset has a constant payoff. Then asymmetric information doesn’t matter, since all information is irrelevant. Debt is an instrument that promises close to a constant payoff as long as the likelihood of default is very low.2 An AAA-rated bond is in that category. It is backed up by sufficiently valuable assets that one does not need to collect much, if any, further information about the underlying assets. The low information sensitivity of AAA-rated bonds makes them more liquid than lower-rated bonds. Lower-rated bonds are more information sensitive because the likelihood of default is higher, and therefore, investors care more about the bond’s underlying assets.3 Interbank markets, repo markets and other near-money markets are not ones in which trading is based on a careful, continuous assessment creditworthiness. They are low-information markets where trading has to be based on trust because there is no time for detailed evaluations. The need for information is kept low by making sure that there are more than enough assets, including reputation, to back up the liabilities. Holding securities that are not information sensitive also helps liquidity. As major liquidity providers, bank assets are often debt instruments or close cousins of debt. Because house prices tend to move slowly compared with other assets, mortgages are especially well-suited for banking business. The connection between asset payoffs and liquidity is developed in detail in Gorton and Pennacchi

08 Book.indb 266

2/13/09 3:58:39 PM

Commentary

267

(1998), and the theory is used to explain the liquidity providing role of banks. Of course, if the value of the assets backing up the debt falls below the face value of debt (or gets close to it), debt becomes much more information sensitive. Creditors become more like equity holders. They now need to assess the value of all the individual assets securing the debt (explicitly or implicitly). This is a much more challenging exercise and one that markets for liquidity provision are ill-equipped to deal with. In a market that is supposed to roll over billions of dollars of debt each day, a sudden need to evaluate counterparty collateral can be devastating. These markets operate on trust, that is, faith that the counter-party is creditworthy, with no time for detailed evaluations. As the saying goes, if a banker has to prove his creditworthiness, he has already lost it. In this respect, markets for liquidity provision are very different than stock markets. In the stock market, uncertainty and adverse selection fears are present all the time, but this does not prevent the markets from functioning. One reason is that stock traders are typically not tied in a chain of debt obligations that make risk hard to bear. Another reason is that the urgency to buy and sell is normally absent. If one is uncertain about the value of a stock, one can simply wait for more clarity. A third reason is that differences in beliefs often alleviate adverse selection. Stock markets thrive on differences in beliefs. Markets for liquidity are killed by it. In light of this discussion, the problem with subprime-related securities was not the lack of transparency as such; in fact, opaqueness is a common characteristic of liquidity-providing markets, as I have tried to illustrate. The real problem was the sensitivity of the MBSs to a fall in the average housing price. The protective layers of lower-rated tranches in the subprime securitization were meant to absorb the losses from a drop in housing prices. But they were not designed properly. One of the most interesting and important details in Gorton’s narrative is the description of how the BBB tranche was meant to grow in size over time. He gives the example of two MBSs, one issued in 2005, the other issued in 2006. Each starts off with a BBB layer that is roughly 1% of the value of the underlying

08 Book.indb 267

2/13/09 3:58:39 PM

268

Bengt Holmström

mortgage portfolio and subordination of between 3 and 4%—a very thin buffer indeed. The idea was that as house prices grew, the increased value of houses would be fed into the BBB buffer, increasing it via refinancing of the mortgages, analagous to the way equity grows if the firm value increases while debt is kept constant. In the case of the 2005 issue, the BBB-layer subordination (i.e., the buffer junior to this layer) grows from 2.95% in 2005 to 9.06% in 2007, providing a lot more protection. In the case of the 2006 issue, however, the BBB-layer subordination only fell from 1.6% to 1.1% because housing prices had started falling. As a result, there was no credit enhancement in the 2006 vintage. The 2005 vintage subprime securities have weathered the initial storm much better than the 2006 vintage, as documented in the paper. Rather than providing a big enough protective layer up front, the MBS structure hoped to build the layer over time, relying on increasing housing prices and favorable refinancing. It was a novel idea that worked wonders initially. Subprime lending boomed, and the returns on the associated securities were attractive due to the high leverage. But the dynamic credit enhancement model only worked as long as house prices were rising, a point that seems obvious in retrospect. Moreover, having a debt instrument move with the aggregate housing price reduces the risk of default, but it makes default much more costly when it happens, since the shock hits everyone at the same time. It is possible that market participants failed to see the correlation of risk induced by securitization, but it is more plausible that they succumbed to a free-rider problem. No individual player had an incentive to care about systemic risk. What lessons can be drawn from this discussion of transparency? First, it calls into question the rush to increase transparency in markets for liquidity provision. Increased transparency will never reach the level of full information, and as we have seen, going half-way may induce parties to acquire information that makes adverse selection worse and markets less liquid. Marking assets to market may not be such a good idea for liquidity-providing institutions.

08 Book.indb 268

2/13/09 3:58:39 PM

Commentary

269

The second lesson is about leverage. I assume that the idea of dynamic credit enhancement was borne out of the need to finance marginal borrowers with little or no collateral and earnings power. Perhaps there is a way to make it work, but the evidence suggests that the problem can be disguised with complex securities only for so long. If the desire is to help marginal borrowers buy houses, it would seem better to be explicit about the subsidies. The third lesson concerns the systemic risks induced by securitization. The problems with highly correlated securities and free-riding need careful scrutiny. Also, the inability to identify the location of toxic assets is a major obstacle for a rescue operation. This has to be factored into the assessment of social risk, and the cost has to be passed in some way onto the originators of loans. For instance, one could have stricter rules on what kinds of structured securities qualify for BIS regulated capital adequacy requirements and how they should be counted. But there are obviously many other possibilities, too. Finally, the general point that markets for liquidity provision are fundamentally different from markets for risk sharing raises important questions. Securitization unwittingly moved some of the banking business into investment banks that were unregulated and could expand fast. The rents of traditional banks were threatened, which led to an increased appetite for risk—a new banking era that supposedly required a different business model. It may still be revived, but for now it seems that a clearer boundary between regulated banks that are in charge of securing and distributing liquidity and the rest of the capital markets, including the former shadow banking system, should be drawn. Had the riskier tranches of the MBSs been held outside the banking sector, the drop in housing values would probably not have caused a big crash, as evidenced by the LTCM crisis and the burst of the tech bubble in 2000. Risk should be shared in equity markets, not in liquidity providing markets. Why did subprime lending and mortgage securitization grow so big? The political desire to expand home ownership, paired with new financial engineering technology, gave birth to the subprime market.

08 Book.indb 269

2/13/09 3:58:39 PM

270

Bengt Holmström

Fannie Mae and Freddie Mac, which were set up to assist poorer households, were encouraged to buy mortgages with capital that was cheaper because it was implicitly secured by government. This may have fueled the growth in subprime. Yet, the fact that broker-dealers were eagerly competing with Freddie and Fannie and in the end held onto a lot of the subprime debt that they originated, because the business was so lucrative (Adrian and Shin, 2008), suggests that government pressure may have been less of an issue than often claimed. For this reason, one must look for another driver. Caballero (2008) has suggested an explanation that sounds more compelling and could account for the strong investor interest in subprime lending as well as home equity loans. In short, they argue that foreign money, especially from emerging economies like China, was looking for a safe haven. The manufacturing boom had generated high incomes, and the savings rates were also high. Because financial markets in developing countries were poorly developed, there was a shortage of domestic savings instruments. The Anglo-Saxon world, especially the US, with their highly developed financial markets, could meet the foreign savings. So, huge amounts of capital began flowing into the US after the turn of the century. The inflow of capital is often interpreted as the result of US consumption greed, but in this story it is the capital account surplus that drives the current account deficit and not the other way around. The evidence that supports this interpretation over the conventional one is the behavior of interest rates. They went down rather than up as the current account deficit widened, suggesting that money was being pushed rather than pulled into the US. The foreigners’ desire to place their money in the US raised asset prices and lowered interest rates, both of which gave financial firms the incentive to create additional means to save. There are three ways to increase the supply of savings instruments: (i) one can build new assets; (ii) one can turn privately held assets into marketable assets; and (iii) one can make better use of the available assets by creating richer state-contingent claims. All three channels were used. And importantly, as they were used, money was necessarily put into more marginal projects. The boom

08 Book.indb 270

2/13/09 3:58:39 PM

Commentary

271

in subprime lending is the prime example. Of course, the foreigners’ desire to invest in the US also happened to match perfectly the government’s wish to expand home ownership, but this may be less important. The flow of money into housing was aided by excess capacity in the US corporate sector; there was little corporate demand for the new funds. In the prevailing low-interest environment, home-equity loans were aggressively peddled, especially to owners who had small or no mortgages. Foreign money was looking for a safe home—literally. It may be more of a stretch to claim that structured securities were created to absorb more foreign investment, except to the extent that such claims made subprime lending feasible. But in theory at least, the contingent use of collateral, which the complex chains of structured securities enabled, was more efficient. Higher efficiency raised the value of collateral, allowing it to absorb more savings. The fact that global imbalances may have played a big role in setting the stage for the crisis does not make the current crisis any easier to deal with. But it puts this painful episode in a very different light when considering regulatory measures to prevent future crises. The desire to blame Wall Street is understandable, and much of it may be warranted. Yet, it is important to realize that the huge profits and bonuses probably came from legitimate efforts to create economic value by satisfying the demand for new financial assets. The crisis is no proof against this hypothesis. Furthermore, and this is really the important point, if the underlying driver was an increase in the foreign demand for US financial assets, we need to think about remedies and counter-measures that go beyond structural changes in financial markets. The US is confronting problems that commonly are associated with emerging markets: how to deal with a burst of foreign investment. The difference is that the money flowing into the US is not likely to be hot money. So far, there are no signs that foreigners are pulling out the money abruptly as typically happens in emerging markets. The other difference, of course, is that the US debt is in dollars, not in foreign currency. That will obviously eliminate many problems. Yet, the fact that the flow of foreign money could so thoroughly throw the US financial system into turmoil

08 Book.indb 271

2/13/09 3:58:39 PM

272

Bengt Holmström

is shocking and requires careful study. Macroeconomic measures that could have prevented the trouble should also be explored. Concluding remark One of the dangers of managing a crisis is that the political and regulatory attention is too much focused on the proximate causes of the crisis. It is evident that the originate-and-distribute model of selling mortgages and the attendant chain of structured securities built around it has been discredited and will take much of the blame. Greater transparency and simplicity in the financial sector is a likely consequence. Gorton’s paper appears consistent with such a response. Yet, it would be important to go beyond the proximate causes and try to develop a deeper understanding of the events before major regulatory changes are made. In my discussion I have tried to illustrate why seemingly obvious recommendations, like increased transparency, may be far off-target once the role of liquidity providing markets is better appreciated. Likewise, knowing what caused the increased demand for subprime securities makes a big difference for regulatory decisions. Of special concern is the tendency to demonize or ban innovations that backfired, not because they were fundamentally wrong, but because the particular implementation was flawed. The originate-anddistribute model and MBSs will certainly have an important place in the future. Gorton’s paper, with its detailed narrative, can help identify the critical errors in contracting that occurred and thereby bring about an improved version of this business model.

Author’s Note: Thanks to Ricardo Caballero, Gary Gorton, Pentti Kouri, Jeremy Stein and Jean Tirole for inspiring discussions on the subject.

08 Book.indb 272

2/13/09 3:58:39 PM

Commentary

273

Endnotes See Milgrom and Roberts (1992) for a more detailed description.

1

Debt payoffs also depend on public information, like interest rates, but being public, this does not usually cause serious information asymmetries. 2

3 Gorton and Pennacchi (1990) have made the same point by showing theoretically that claims that do not move much with the underlying assets are less susceptible to adverse selection. Consequently, liquidity providing intermediaries are focused on creating such claims, deposits being the most obvious example. They also note that securitization (by pooling assets) creates a more liquid claim, because it becomes less sensitive to variations in the values of underlying assets. Money is the ultimate securitized asset. It is highly opaque, because it is a claim on all the productive assets of the economy. It is very information insensitive: The value of money moves slowly, except in states of hyperinflation. Since information about the value of money is so symmetric, money is very liquid.

References Adrian, T. and H. S. Shin (2008), “Financial Intermediaries, Financial Stability and Monetary Policy,” this volume. Caballero R., E. Fahri and P-O. Gourinchas (2008), “An Equilibrium Model of Global Imbalances and Low Interest Rates,” American Economic Review, 98 (1), 358-93. Gorton, G. and G. Pennacchi (1990), “Financial Intermediaries and Liquidity Creation,” Journal of Finance, Vol XLV (1), 49-71. Milgrom P. and J. Roberts (1992), Economics, Organization and Management. Englewood Cliffs, New Jersey: Prentice Hall.

08 Book.indb 273

2/13/09 3:58:39 PM

General Discussion: The Panic of 2007 Chair: Martin Feldstein

Mr. Feldstein: If I could just make one comment on what Bengt said: It is certainly true foreign countries, particularly emerging-market countries and the oil producers, had a lot of funds to invest. Most of that went into Treasuries. They went into mortgages indirectly through the GSEs by buying Fannie and Freddie bonds. But, by lowering interest rates, they created an environment in which investors seeking yield were prepared to take risk and the creation of these wonderful apparently AAA securities, or AAA tranches, were just a tempting vehicle for getting more yield without apparently taking on as much risk as we now know after the fact was actually there. Mr. Moskow: I have a question about one of the details in the paper, not the big picture. That is really on the last page of the paper where you talk about Financial Accounting Standards Board (FASB) 140. This deals with off-balance-sheet entities, or the so-called qualified special-purpose entities. As you mentioned here the FASB is considering changes which could lead to significant modification, maybe elimination, of this provision and then could require consolidation. My understanding is this has been delayed a year, but still it is on the fast track for FASB. You say it is likely to create problems.

275

08 Book.indb 275

2/13/09 3:58:39 PM

276

Chair: Martin Feldstein

On the other side of this is that many believe these never should have been off balance sheet in the first place. They should have been part of the consolidated entity. If they are off balance sheet, the sponsoring entity is not supposed to have any responsibility for that offbalance-sheet entity. As these off-balance-sheet entities have problems, some large banks have brought them back on their balance sheets now, in part because of reputational risk. The result, if you bring them on balance sheet, is there will be less leverage for that institution, and regulators will now include that in their capital ratios. Isn’t this part of the deleveraging process that has to take place in order to get back to conditions of financial stability? Mr. Gorton: I think what happened was, FASB doesn’t really understand what is happening in the crisis and mistakenly thinks there is something wrong with all the securitization. So they somehow have the view that, if you force consolidation for accounting purposes, this is going to solve some problem. It is creating some problems, which, because it has been delayed in the final rule probably will look a lot different. I included that as an example of the kind of policy response that doesn’t really arise from what is actually happening. An implicit contract, which says in certain states in the world, “I reconsolidate my vehicles,” doesn’t bother me. That’s how the world thinks everything functions. I have a paper that shows that is priced. It is a problem for accountants that they don’t know how to deal with implicit contracts, but I don’t think that helps. It doesn’t help a crisis to try to propose something, which really doesn’t have anything to do with the crisis we’re facing. It is misguided. Mr. Fischer: I am also on the last page. You say, “As Merton Miller pointed out more than twenty years ago, financial innovation is largely driven by regulation and taxes. Entrepreneurs will take risk in some form, somewhere. The trends are already clear. Talent is increasingly moving to the least regulated platform: hedge funds.” I’m trying to figure out what to make of that and what it says about the current situation. If hedge funds are not going to have access to the lender-of-last-resort facility—and they shouldn’t—do we need to worry about where the talent is moving?

08 Book.indb 276

2/13/09 3:58:40 PM

General Discussion

277

We have had one example of a hedge fund blowing up, at least one, and getting the Fed into action. That was Long-Term Capital. The Fed got into action as a coordinator of the rescue package and not as a provider of liquidity to the rescue package. I think that was more desirable than the alternative of lending to LTCM. The question to you is, What are we to make of your last paragraph? Mr. Lindsey: I very much liked the detail, but I would like to add another detail to your paper, Gary. This comes as a former chairman of neighborhood reinvestment. I know it’s all blamed on the subprime crisis. The fact is those loans were not demonstrably more risky. In fact, in general they were less risky than the standard mortgage until the start of this crisis. What happened was what Bengt had mentioned, which is a generalized increase in volume for all kinds of loans. We established a basically soft quota system for Community Reinvestment Act (CRA) lending under the CRA provisions. As the aggregate volume went up—and you mentioned these same procedures were done in standard loans and in credit card loans—the volume went up for subprime loans as well. But that overwhelmed the riskmitigation procedures which were in place. That was the reason for the default process. So there was a volume issue that should be inserted as one of the details to be given for the points you made. Mr. Bergsten: I have two questions on Professor Holmström’s bigpicture, long-term view that’s very attractive to me. You suggested the demand for liquid assets—parking space—was the big driver of what’s happened in the housing markets more than a social demand for more housing investment. Why not both? Why not the two moving together? But I really want to focus on the foreign side. You talk about the foreigners looking for parking space for liquid assets. But, the motivations for the buildup in foreign reserves around the world have differed substantially from country to country. In some—including the biggest one, China­—it has been motivated by very traditional mercantilist objectives: Keep the exchange rate undervalued, subsidize trade

08 Book.indb 277

2/13/09 3:58:40 PM

278

Chair: Martin Feldstein

surpluses, and subsidize growth and employment creation through that device. I have interpreted what they have done as essentially an off-market export and job subsidy which, since they don’t mark to market, comes under no domestic surveillance and, at least so far, has come under no effective international response either. I have to believe that in their heart of hearts and mind of minds, they knew those subsidies would cost them something through an eventual loss in the capital value of the assets they were buying to preserve of their huge surpluses. So how is your model affected if some of the investors essentially don’t care if they take a capital loss? Because they don’t mark to market or have any domestic surveillance and discipline over what they do, they are able to invest massive amounts in a way that promotes a social outcome on their side, regardless of the financial outcome. The specifics would be different in Japan or the oil producers or other large accumulators of the assets you are talking about, but in no case do they mark to market or come under any kind of significant domestic discipline. How does that affect your analysis and does it suggest there was really a conspiracy on both sides of the equation to let prices diverge sharply from long-term equilibrium and thus set up the crisis? Mr. Alexander: First of all, I really liked this paper. It has pulled together a lot of things I’ve been thinking about very much from a practical side. The question I would have for both of you relates to the connection between this and information technology and the complexity. Partly what struck me when I read this was this story would have been impossible 10 years ago, just because of the sheer complexity of the contractual arrangements involved. When you think about that and the policy implications, I would link it back to the first part of the chairman’s speech, where he talked about the critical issues of market infrastructure. Partly what strikes me about this is the way information technology has allowed this complexity to evolve and whether or not the institutional infrastructure is really kept up. I would point out that

08 Book.indb 278

2/13/09 3:58:40 PM

General Discussion

279

nobody has repealed Moore’s Law. I suspect we are going to have more of these things going forward. I would be interested in, both the author and discussant, how they see that as contributing to this. Mr. Rajan: One thing I missed in Gary’s excellent paper was why the participants did not realize the loss of information. I am not just talking about the asset managers, but the investment banks who held on to a lot of the stuff. Why did they not realize that complexity would increase multiple-fold, especially when the defaults started multiplying? What did they miss? Do we have to ultimately go to some kind of breakdown of control systems to get an explanation to that? Mr. Trichet: I was very interested in Gary’s paper and Bengt’s remarks. To Gary, I would ask the following question. If it appears that the ABX indices introduction was absolutely decisive to clarify the situation and permit the market to realize the level of risks that was not visible before, could you elaborate on what would have happened had these indices been introduced earlier? Is there a lesson to draw from this situation? Should we be more involved, as regulators or overseers of markets, to deal with ways to permit the market to realize the overall risks itself? I understand that hedging activity in this particular domain has been an important factor triggering large moves in the value of these indices. Mr. Carney: Actually, I would like to pick up where Jean-Claude left off because one of the persuasive aspects of your paper, and also your summary, was the loss of information. Investors did not know what they owned. When we lived through the ABCP debacle in Canada, the investors had the luxury of time, they had all the information because of a standstill, and they still couldn’t figure out what they owned or how to evaluate it. Now they are going to have the luxury of eight more years of owning the stuff to ultimately discover its relative value. One of the big dynamics of the time—and I am getting to my question—was the uncertainty about whether the Canadian non-bank ABCP situation was going to be resolved was one of the things that fed through the ABX indices because of the hedging that was there. That explained part of the dynamic.

08 Book.indb 279

2/13/09 3:58:40 PM

280

Chair: Martin Feldstein

My question is, Given you make a case for the relative uniqueness of subprime and you look at not just the implied default probabilities of the ABX indices—but on the broader derivative indices, default probabilities which are a multiple of any potential, reasonable outcome one would expect—if you think about the CLO market particularly—what is your explanation for that? Is it hedging? Is it a war of attrition between hedge funds that can’t get sufficient financing to get the proper leverage returns and real money which is just waiting to call the bond and come in? To sum up, since one of your core arguments, I thought, was that subprime was relatively distinct from “good” securitization, if you will, why do we still have these price dynamics in the indices for good securitization (“good” is my term obviously)? Mr. Swagel: I wanted to ask Gary about a policy proposal. It is someone else’s proposal. And that is for a database on mortgage information. The idea that people talked about a securitization forum has talked about this. This loan level information would be put out by the servicers—both origination and then ongoing performance. There is some of this available at www.loanperformance.com and www.mcdash.com, but it is very imperfect. This would be more complete, more regular, and more up–to–date. The idea, as people have discussed, is to allow better analysis. So there is complexity, but at least the analysis could pierce through that complexity. We thought about it and said this would also have the benefit of allowing a reputational tail, back to the originator. So the question is, What is your reaction, maybe why didn’t it happen earlier, why didn’t investors demand it, and would it be useful? Mr. McCulley: I have one comment I am going to make about Gary’s paper and then a question to ask. The comment to Gary is—I thought it was an absolutely fantastic paper; I understood every line in it—but importantly, I want to stress I disagree with your viewpoint that the buy side has no talent. I happen to be on the buy side. We have a lot of talent, so therefore I just want to say that we are an exception to your rule. We did understand it and we didn’t buy it.

08 Book.indb 280

2/13/09 3:58:40 PM

General Discussion

281

My question is, I found it very fascinating your describing that the private sector has to act ex ante, whereas the public sector can act ex post. Sometimes the private sector just freezes, which demands the public sector act ex post to try to right the situation. It seems to me that is the framework we all learned back in graduate school about the paradox of thrift, liquidity trap, etc. If the private sector is frozen, then the public sector has to take the other side of the train. Is what you are describing now—with the deleveraging going on—simply the modern-day version of the paradox of thrift? But now it’s the paradox of deleveraging. The profit sector is wrapped up in a downward spiral of deleveraging, which implies that the public sector should go the other way and lever up. Mr. Rosengren: Systematic workouts are very difficult under the servicing agreements and current contractual obligations. I was just wondering what you thought about changing some of the contractual obligations in the governance rules in many of these securitization agreements to make it easier to have a more systematic way of dealing with workouts that don’t seem to have been contemplated in the original agreements? Mr. Gorton: Thank you very much for a lot of very good questions. Let me start by saying I agree with everything Bengt said. I am very sympathetic to the Ricardo Caballero view of the world. Bengt and I talked about what we should each say ahead of time. From my point of view, it worked out very well. Several people read the last page of my paper and had questions on that. It makes me a little bit suspicious about whether they read the other stuff. In response to Stan Fischer’s question: I put that in there because I observe it. There is a lot of talent that goes where the money is. It is a caution to us. Your suspicions are exactly right. The financial landscape is changing so quickly I can’t even keep up with it. There are firms that are mergers of hedge funds and private equity. Citadel has issued rated debt. I watch people, friends of mine from investment banks, move to hedge funds. You read the newspapers. Lloyd Blankfein of Goldman

08 Book.indb 281

2/13/09 3:58:40 PM

282

Chair: Martin Feldstein

made $60 million last year and Ken Griffin of Citadel made $1.2 billion. That tells you how the world is changing. Hedge funds are entities we know almost nothing about. If you try to look at the private data sets, these are strategically manipulated by hedge funds. They report when they feel like it. They can ask if their returns be erased subsequently. There is survivor bias, look-ahead bias, there are all sorts of bias. So we literally know nothing about hedge funds. There is a tendency, and I have this tendency too, to gravitate to the least-regulated platform. That is something we have to be aware of, and Ben alluded to this in the beginning. On Larry Lindsey’s point: Subprime mortgages are demonstrably more risky, if you look at the default rates even in the early 2000s. But I agree with you that in 2006-07, for other reasons I thought that the volumes would decrease. What happened was certain institutions—those that subsequently had large writedowns—did not decrease their activity. Those would be Merrill Lynch, UBS, and Citibank. I was very shocked at that and I agree with you those vintages were very, very troubling. There was a point—I forget who made this point—but it was related to something Bengt said. One of the points I want to emphasize in my paper is that the amount of collateralizable wealth in an economy is very hard to determine ex ante. If you would have asked me two years ago “Will it always be the case that I’ll be able to use agency bonds as collateral under swap CSAs and for repo?” I would have said, “Absolutely!” In August 2007, that was not the case. The problem is that in a crisis, the amount of collateralizable wealth shrinks because of the loss of information. Somehow it can’t be counted on—you can’t count on any certain amount of collateralizable wealth. The question about information technology: That is a really fascinating question. Certainly these structures would not have been possible 10 years ago. Subprime lending, per se, depended upon automated credit scoring and advances in credit scoring which required huge advances in being able to manipulate extremely large datasets. It is also the case, as statistics become a more computer-driven

08 Book.indb 282

2/13/09 3:58:40 PM

General Discussion

283

discipline with bootstrapping everything and simulating things, even these structures are a challenge to simulate. From that point of view, to go to another question, it is not really a matter of the data. It is always better to have more data. The loan performance data, which is available now, doesn’t really help you, if you are further along in the chain. For example, if I want to look at a 2006-vintage, second-lien subprime bond, this is something most people would run screaming from the room if we mentioned buying it (it might be a very good buy), but the structure of that bond is so complicated that you could not link the actual structure to simulate from the actual mortgage data. You cannot do that. That would require one-off coding that would take a long time. That challenge is going to be overcome very shortly, which will mean this will all be possible. The information that it’s the machine power makes a lot of this extremely data-driven. That is likely to continue. Raghu asked why did these banks hold onto this stuff and was there a breakdown in control systems. One of the points that was made in response to Charlie’s talk was really good. Somebody pointed out there is cross-section variation in these firms. Not everybody has the same experience with subprime bonds. There are certain institutions, which have very different experiences. If you read, for example, the UBS Investors Report about them, you’ll see pretty clearly there were problems that didn’t exist in some other institutions. It doesn’t strike me as something that was common to all institutions, but it was there for some institutions. The ABX index: The guys who came up with the ABX index are four investment bankers—one from four different firms. I know these guys. The idea was to create a tradable index, but partly it was largely for hedging purposes. The interesting thing about the ABX index is the first vintage was 2006.01. The way it works is they survey the dealers, they set a coupon, and then it’s based on the price. It trades on the price. The investment banks didn’t ever want to set the coupon too high because you didn’t want people going long to have to pay money. So you always wanted the price to be a bit lower than par.

08 Book.indb 283

2/13/09 3:58:40 PM

284

Chair: Martin Feldstein

The first two vintages were exactly at par. When house price growth rates started to decline, the indices didn’t really do anything. It was only in the 2007.01 vintage where everything fell off the cliff. I am working on another paper on this. What happened, though, is this is an index that is linked to a particular pool of subprime bonds. So there is an arbitrage between those bonds and the index. That arbitrage broke down in 2007 because you couldn’t finance the bonds through repo. Even if you could find the bonds, you couldn’t finance the bonds through repo, so the index delinked from the actual bonds and the market was one-sided. Everybody wanted to buy protection. The thing just went through the floor. The result of that was the prices really had no relationship with the bonds. To go to Mark Carney’s question: The implied default probabilities of the ABX index are completely unrealistic because their arbitrage is broken down. What you’re looking at is not implied default probabilities; you are looking at latent demand for hedging. This was a big problem in marking to market. If I have to mark to market on the ABX index, I’m marking to prices, which are just telling me everybody’s scared out of their minds. Then I have to mark down and everybody is scared out of their minds. That index is dead forever and it has served its purpose for a very brief period before the arbitrage broke down. The breakdown, from an academic point of view, is interesting because it is a glimpse into the magnitude of the liquidity problems in the repo market. Just a couple of other things: I apologize to PIMCO. And then Eric’s question: Is it reasonable to think about changing the servicing agreements to facilitate restructuring the mortgages? The problem is that the claimants in subprime securitizations are divergent. So in a subprime securitization, there is a number of claims that are issued publicly and then there is a whole bunch of residual claims to various cashflows. If I make a prepayment penalty payment on my mortgage, that goes to the securitization and that goes to a certain claimant. Not all of these interests are aligned. The servicers have a lot of skin in the game here. They are doing their best, but if you say to them, “Why don’t you just restructure these things?” the answer is they are

08 Book.indb 284

2/13/09 3:58:40 PM

General Discussion

285

going to be sued. And that’s why there was this discussion that maybe Congress should give them immunity or something. There is no easy way around that without screwing some of the claimants here. That is just the reality of the situation. An easier way out is not even to go down that road. It is to somehow have a mechanism to allow people to refinance without having to change the servicing agreements. Mr. Holmström: Let me just comment on a few things. Stan’s question about hedge funds brings up one important point. One implication of the big–picture view I sketched out in my comment is that other actors that are more proficient in information-intensive financing will step in and play a very important role in the rescue effort. On Fred Bergsten’s question: It is not that important why foreigners came to demand American assets? I just told Caballero’s story because I find it compelling. That said, a lot—more than 50 percent—of foreign investment came from non-government sources, especially from Europeans. I am not exactly sure what you meant by a conspiracy, but I’m rather suspicious of conspiracy theories at this scale. I also want to comment on Raghu’s question, why buyers in the chain didn’t realize that there was a loss of information. My simple answer is that they weren’t even looking for that information. The market for liquidity is set up so that buyers don’t have to worry about information asymmetries. A high-volume market will not have the time to evaluate individual instruments. But that’s what they have ended up doing now, causing great damage to liquidity. Reinventing the wheel is a time honored way of making progress. If this is just the paradox of thrift, that’s of little help. I don’t think we have a good theory of how society should ensure and provide liquidity, even in the most ideal of circumstances.

08 Book.indb 285

2/13/09 3:58:40 PM

Financial Intermediaries, Financial Stability and Monetary Policy Tobias Adrian and Hyun Song Shin

1. Introduction Financial intermediaries have been at the center of the credit market disruptions that began in August 2007. They have borne a large share of the credit losses from securitized subprime mortgages, even though securitization was intended to parcel out and disperse credit risk to investors who were better able to absorb losses.1 The capacity to lend has suffered as intermediaries have attempted to curtail their exposure to a level that can be more comfortably supported by their capital. The credit crisis has dampened real activity in sectors such as housing and has the potential to induce further declines. The events of the last twelve months have posed challenges for monetary policy and have given renewed impetus to think about the interconnection between financial stability and monetary policy. The current credit crisis has the distinction of being the first postsecuritization crisis in which the banking and capital market developments have been closely linked. Historically, banks have always reacted to changes in the external environment, expanding lending when the economic environment is benign. However, the increased importance of intermediaries that mark balance sheets to market both sharpens and synchronizes the responses, giving more impetus to the feedback effects on the real economy. The potential for adverse 287

08 Book.indb 287

2/13/09 3:58:40 PM

288

Tobias Adrian and Hyun Song Shin

real effects are especially strong when banks respond to credit losses or the onset of more turbulent conditions by cutting their exposures, reducing lending and charging higher risk premiums. Prudent risk management dictates such actions, and the script is well rehearsed.2 One notable finding from our empirical analysis is that there are important distinctions between different categories of financial intermediaries. Fluctuations in the balance sheet size of security broker-dealers—the financial sector that includes the major investment banks—appear to signal shifts in future real activity better than the larger commercial banking sector. In fact, the evolution of brokerdealer assets has a time signature that is markedly different from those of commercial banks. We find that the growth in broker-dealer balance sheets helps to explain future real activity, especially for components of GDP that are sensitive to the supply of credit. Our results point to key differences between banking as traditionally conceived and the market-based banking system that has become increasingly influential in charting the course of economic events. Broker-dealers have traditionally played market-making and underwriting roles in securities markets. However, their importance in the supply of credit has increased dramatically in recent years with the growth of securitization and the changing nature of the financial system toward one based on the capital market, rather than one based on the traditional role of the bank as intermediating between depositors and borrowers. Although total assets of the broker-dealer sector is smaller than total asset of the commercial banking sector, our results suggest that broker-dealers provide a better barometer of the funding conditions in the economy, capturing overall capital market conditions. Perhaps the most important development in this regard has been the changing nature of housing finance in the US. As we will see shortly, the stock of home mortgages in the US is now dominated by the holdings in market-based institutions, rather than traditional bank balance sheets. Broker-dealer balance sheets provide a timely window on this world. On a similar theme, we find that market equity (of both the broker-dealer sector and the commercial banking sector) do a better job of signaling future real activity than total assets themselves.

08 Book.indb 288

2/13/09 3:58:40 PM

Financial Intermediaries, Financial Stability and Monetary Policy

289

Having established the importance of financial intermediary balance sheets in signaling future real activity, we go on to examine the determinants of balance sheet growth. We find that the level of the fed funds target is key. The fed funds target determines other relevant short-term interest rates, such as repo rates and interbank lending rates through arbitrage in the money market. As such, we may expect the fed funds rate to be pivotal in setting short-term interest rates more generally. We find that low short-term rates are conducive to expanding balance sheets. In addition, a steeper yield curve, larger credit spreads and lower measures of financial market volatility are conducive to expanding balance sheets. In particular, an inverted yield curve is a harbinger of a slowdown in balance sheet growth, shedding light on the empirical feature that an inverted yield curve forecasts recessions. These findings reflect the economics of financial intermediation, since the business of banking is to borrow short and lend long. For an off-balance sheet vehicle such as a conduit or SIV (structured investment vehicle) that finances holdings of mortgage assets by issuing commercial paper, a difference of a quarter or half percent in the funding cost may make all the difference between a profitable venture and a loss-making one. This is because the conduit or SIV, like most financial intermediaries, is simultaneously both a creditor and a debtor—it borrows in order to lend. Although our results are in line with the economics of financial intermediation, they run counter to some key tenets of current central bank thinking, which has emphasized primarily the importance of managing expectations of future short rates, rather than the current level of the target rate per se. In contrast, our results suggest that the target rate itself matters for the real economy through its role in the supply of credit and funding conditions in the capital market. As such, the target rate may have a role in the transmission of monetary policy in its own right, independent of changes in long rates. When we examine how monetary policy has been conducted in practice in the US, we find that the fed funds target rate tends to be reduced following expansions of balance sheets, and tends to be raised following slowdowns in balance sheet growth. But there is one

08 Book.indb 289

2/13/09 3:58:40 PM

290

Tobias Adrian and Hyun Song Shin

important proviso. In periods of crisis, the fed funds target has been cut sharply to cushion the economy from the fallout from the crisis. Our findings hold implications for the financial stability role of monetary policy. To the extent that the financial system as a whole holds long-term, illiquid assets financed by short-term liabilities, any tensions resulting from a sharp, synchronized contraction of balance sheets will show up somewhere in the system. Even if some institutions can adjust down their balance sheets flexibly in response to the greater stress, not everyone can. This is because the system as a whole has a maturity mismatch. While lender-of-last-resort tools may mitigate the severity of the contraction in balance sheets, they cannot prevent the contraction altogether. Something has to give. There will be pinch points in the system that will be exposed by the de-leveraging. The pinch points will be those institutions that are highly leveraged and who hold long-term illiquid assets financed with short-term debt supplied by lenders who reduce their exposure in response to deteriorating financial conditions. When the shortterm funding runs away, the pinch point financial institutions will face a liquidity crisis. Arguably, this is exactly what happened to Bear Stearns in the US and Northern Rock in the UK, as well as a host of conduits and SIVs that have been left stranded by the ebbing tide of funding in the current credit crisis. In this way, the expansions and contractions of balance sheets have both a monetary policy dimension in terms of regulating aggregate demand, but also a financial stability dimension. Therefore, contrary to the commonly encountered view that monetary policy and policies toward financial stability should be conducted separately, the perspective provided by our study suggests that they are closely related. They are, in fact, two sides of the same coin. The common coin is the markedto-market balance sheet dynamics of financial intermediaries. Although there has been a long-running debate on how far monetary policy should take account of financial stability goals, the debate has primarily focused on either 1) commercial banks, or 2) asset prices. The debate has not focused as much on the institutions that are at the heart of the market-based financial system, such as security broker-dealers. In relation to asset prices, the question has been

08 Book.indb 290

2/13/09 3:58:40 PM

Financial Intermediaries, Financial Stability and Monetary Policy

291

whether central banks should react to asset price bubbles.3 The case against reacting to asset price bubbles is a familiar one, and rests on the following arguments. • Identifying a bubble is difficult. • Even if there were a bubble, monetary policy is not the right policy tool in addressing the problem. An asset price bubble will not respond to small changes in interest rates. Only a drastic increase in interest rates will prick the bubble. • However, such a drastic increase in interest rates will cause more harm than good to the economy in terms of future output and output volatility. The claim that an asset price bubble will not respond to a small change in interest rates has mostly been argued in the context of the stock market, where the proposition is indeed plausible. However, the stock market is not the best context in which to discuss the financial stability role of monetary policy, as stocks are held mostly by unlevered investors. Much more central is the credit market, especially when backed by residential or commercial real estate. As argued already, a difference of a quarter or half percentage in the funding cost may make all the difference between a profitable venture and a lossmaking one for leveraged financial intermediaries. We believe that focusing on the conduct of financial intermediaries is a better way to think about financial stability, since it helps us to ask the right questions. Concretely, consider the following pair of questions. Question 1. Do you know for sure there is a bubble in real estate prices? Question 2. Could the current benign funding conditions reverse abruptly with adverse consequences for the economy? One can answer “yes” to the second question even if one answers “no” to the first. This is because we know more about the script followed by financial intermediaries and how they react to changes in the economic environment than we do about what the “fundamental” value of a house is, and whether the current market price exceeds that value.

08 Book.indb 291

2/13/09 3:58:40 PM

292

Tobias Adrian and Hyun Song Shin

In any case, for a policy maker, it is the second question that is more immediately relevant. Even if a policy maker were convinced that the higher price of housing is fully justified by long-run secular trends in population, household size, rising living standards, and so on, policy intervention would be justified if the policy maker also believed that, if left unchecked, the virtuous circle of benign funding conditions and higher housing prices will go too far, and reverse abruptly with adverse consequences for the economy. The outline of our paper is as follows. We begin with background descriptions of financial intermediation in a market-based banking system. We then present our empirical results on the real impact of broker-dealer balance sheet changes, the determinants of balance sheet changes, and how US monetary policy has reacted to balance sheet changes. We conclude with some implications of our findings for the conduct of monetary policy. 2. Financial Intermediaries in a Market-Based Financial System Behind the development of the market-based banking system is the growth of mortgage backed securities as an asset class. Chart 1 plots the total holding of home mortgages in the US by types of financial institution, drawn from the Flow of Funds accounts for the US. As recently as the early 1980s, traditional banks were the dominant holders of home mortgages, but bank-based holdings have been quickly overtaken by market-based holders of mortgages. In Chart 2, bank-based holdings are defined as the sum of the holdings of the commercial banks, the savings institutions and credit unions. The marketbased holdings are the remainder—the GSE mortgage pools, private label mortgage pools and the GSE holdings themselves. Market-based holdings overtook the bank-based holdings in 1990, and now constitute two-thirds of the $11 trillion total stock of home mortgages. The increased importance of the market-based banking system has been mirrored by the growth of the broker-dealer sector of the economy. Broker-dealers have traditionally played market-making and underwriting roles in securities markets. However, their importance in the supply of credit has increased in step with securitization. Thus,

08 Book.indb 292

2/13/09 3:58:40 PM

Financial Intermediaries, Financial Stability and Monetary Policy

293

Chart 1 Total Holdings of US Home Mortgages by Type of Financial Institution 4.5

4.5 Agency and GSE mortgage pools

4.0

4.0

ABS issuers Savings institutions

3.5

3.5

GSEs Credit unions

3.0

3.0

2008Q1

2006Q1

2004Q1

0.0 2002Q1

0.0

2000Q1

0.5

1998Q1

0.5

1996Q1

1.0

1994Q1

1.0

1992Q1

1.5

1990Q1

1.5

1988Q1

2.0

1986Q1

2.0

1984Q1

2.5

1982Q1

2.5

1980Q1

$ Trillion

Commercial banks

Source: US Flow of Funds, Federal Reserve

Chart 2 Market-Based and Bank-Based Holding of Home Mortgages 7

7 Market-based

2008Q1

2006Q1

0

2004Q1

0

2002Q1

1

2000Q1

1

1998Q1

2

1996Q1

2

1994Q1

3

1992Q1

3

1990Q1

4

1988Q1

4

1986Q1

5

1984Q1

5

1982Q1

6

1980Q1

$ Trillion

Bank-based 6

Source: US Flow of Funds, Federal Reserve

08 Book.indb 293

2/13/09 3:58:41 PM

294

Tobias Adrian and Hyun Song Shin

although the size of total broker-dealer assets is small by comparison to the commercial banking sector (it is around one-third of the commercial bank sector) it has seen rapid growth in recent decades and is arguably a better barometer of overall funding conditions in a market-based financial system In a market-based financial system, broker-dealer assets may be a better signal of the marginal availability of credit as compared to commercial bank assets. At the margin, all financial intermediaries (including commercial banks or GSEs) have to borrow in capital markets through short-term borrowing such as commercial paper or repurchase agreements. But for a commercial bank, its large balance sheet masks the effects operating at the margin. Also, commercial banks provide relationship-based lending through credit lines. Broker-dealers, in contrast, give a much purer signal of marginal funding conditions, as their balance sheet consists almost exclusively of shortterm market borrowing and are not bound as much by relationshipbased lending. The growth of the broker-dealer sector has been striking since the 1980s. Chart 3 charts the increase in the size of the broker-dealer sector in the US relative to the commercial banking sector. Both series are normalized by the size of total household assets. We see that commercial bank total assets have roughly kept pace with total household assets, as evidenced by the flat curve for the series for the ratio of commercial bank assets to household assets. However, the relative size of the broker-dealer sector is more than ten times what it was at the beginning of 1980. Besides growing much more rapidly than commercial bank assets, broker-dealer assets have been more volatile. Chart 4 plots the (annual) growth rates of broker-dealer assets together with the growth rates of commercial bank total assets for the US. We see that brokerdealer assets vary much more sensitively over time, as compared to commercial bank assets. Not only is broker-dealer asset growth more volatile relative to commercial banks, the two series move in quite different ways. Chart 5 is a version of Chart 4 where the commercial bank series has been

08 Book.indb 294

2/13/09 3:58:41 PM

Financial Intermediaries, Financial Stability and Monetary Policy

295

Chart 3 Growth in Broker-Dealer and Commercial Bank Assets Relative to Household Assets (1980Q1 as base) 1100%

1100%

900%

900%

700%

700% Security Broker-Dealer/Household Total Assets: % Growth since 1980

500%

500%

300%

300%

100%

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

-100%

100%

Commercial Bank/Household Total Assets: % Growth since 1980

-100%

Source: US Flow of Funds, Federal Reserve

Chart 4 Growth Rates of Broker-Dealer and Commercial Bank Total Assets 100%

100% Security Broker-Dealer Total Assets: Annual % Growth 80%

80%

60%

60%

40%

40%

20%

20%

0%

0%

2008-

2006

2004

2002

2000

1998

-20%

1996

1994

1992

1988

1986

1984

1982

-40%

1980

-20%

1990

Commercial Bank Total Assets: Annual % Growth

40%

Source: US Flow of Funds, Federal Reserve

08 Book.indb 295

2/13/09 3:58:41 PM

296

Tobias Adrian and Hyun Song Shin

rescaled according to the right hand vertical axis. We see that the peaks and troughs of the two series differ markedly. The chart shows that traditional banking and the new market-based financial system move to a very different beat.4 The balance sheet dynamics of financial intermediaries that mark their balance sheets to market have some distinctive features. Chart 6 is taken from Adrian and Shin (2007) and shows the scatter chart of the weighted average of the quarterly change in assets against the quarterly change in leverage of the (then) five stand-alone US investment banks—Bear Stearns, Goldman Sachs, Lehman Brothers, Merrill Lynch and Morgan Stanley. The first striking feature is that leverage is procyclical in the sense that leverage is high when balance sheets are large, while leverage is low when balance sheets are small. This is exactly the opposite finding compared to households, whose leverage is high when balance sheets are small. For instance, if a household owns a house that is financed by a mortgage, leverage falls when the house price increases, since the equity of the household is increasing at a much faster rate than assets. For investment banks, however, the relationship is reversed. It is as if the householder responded to an increase in house prices by increasing the mortgage loan to value so that leverage increases in spite of the increased value of his house. A procyclical leverage ratio offers a window on the notion of financial system liquidity. When leverage is procyclical, the demand and supply response to asset price changes can amplify shocks. To see this, consider an increase in the price of assets held widely by leveraged market players and intermediaries. The increase in the price of assets strengthens the players’ balance sheets, since the net worth of levered players increases as a proportion of their total assets. When balance sheets become stronger, leverage falls. To the extent that the intermediary wants to avoid holding too much equity (for instance, because return on equity is too low), it will attempt to restore leverage. One way it can do so is by borrowing more, and using the proceeds to buy more of the assets they already hold. Indeed, as we

08 Book.indb 296

2/13/09 3:58:42 PM

Financial Intermediaries, Financial Stability and Monetary Policy

297

Chart 5 Rescaled Growth Rates of Broker-Dealer and Commercial Bank Total Assets 100%

14% Security Broker-Dealer Total Assets: Annual % Growth

80%

12%

60%

10%

40%

8%

20%

6%

0%

4% Commercial Bank Total Assets: Annual % Growth

-20%

2%

0%

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

-40%

Source: US Flow of Funds, Federal Reserve

10

10

2008-1

0

0

2007-3

-10

2007-4

-10

Total Asset Growth (% Quarterly)

20

20

Chart 6 Leverage Growth and Asset Growth of US Investment Banks

-20

-20

1998-4

-20

-10

0

10

20

Leverage Growth (% Quarterly)

Source: SEC; Adrian and Shin (2007)

08 Book.indb 297

2/13/09 3:58:42 PM

298

Tobias Adrian and Hyun Song Shin

see below, the evidence points to broker-dealers adjusting leverage by adjusting the size of their balance sheets, leaving equity intact.5 If greater demand for the asset puts upward pressure on its price, then there is the potential for a feedback effect in which stronger balance sheets feed greater demand for the asset, which in turn raises the asset’s price and leads to stronger balance sheets. Having come full circle, the feedback process goes through another turn. The circular figure on the left illustrates the feedback during a boom. Note the critical role played by procyclical leverage.6 Adjust leverage

Stronger balance sheets

Increase B/S size

Asset price boom

Adjust leverage

Weaker balance sheets

Reduce B/S size

Asset price decline

The mechanism works in reverse in downturns. Consider a fall in the price of an asset held widely by hedge funds and banks. Then, the net worth of such an institution falls faster than the rate at which asset falls in value, eroding its equity cushion. One way that the bank can restore its equity cushion is to sell some of its assets and use the proceeds to pay down its debt. The circular chart above on the right illustrates the feedback during a bust. Note the importance of marking to market. By synchronizing the actions of market participants, the feedback effects are amplified.7 There is a more subtle feature of Chart 6 that tells us much about the financing decisions of financial intermediaries. Recall that the horizontal axis measures the (quarterly) change in leverage, as measured by the change in log assets minus the change in log equity. The vertical axis measures the change in log assets. Hence, the 45-degree line indicates the set of points where equity is unchanged. Above the 45-degree line, equity is increasing, while below the 45-degree line,

08 Book.indb 298

2/13/09 3:58:42 PM

Financial Intermediaries, Financial Stability and Monetary Policy

299

equity is decreasing. Any straight line with slope equal to 1 indicates constant growth of equity, with the intercept giving the growth rate of equity. The feature to note from Chart 6 is that the slope of the scatter chart is close to 1, implying that equity is increasing at a constant rate on average. Thus, equity seems to play the role of the forcing variable, and all the adjustment in leverage takes place through expansions and contractions of the balance sheet. There is a useful perspective on this feature that comes from the risk-management policies of financial intermediaries. Banks aim to keep enough equity capital to meet its overall value at risk (VaR). If we denote by V the value at risk per dollar of assets, and A is total assets, then equity capital E must satisfy E = V x A, implying that leverage L satisfies L = A/E = 1/V If value at risk is low in expansions and high in contractions, leverage is high in expansions and low in contractions—leverage is procyclical.8 One further way we can understand the fluctuations in funding conditions is to look at the implicit maximum leverage that is permitted in collateralized borrowing transactions such as repurchase agreements (repos). The discussion of repurchase agreements is instructive in thinking about leverage and funding more generally, since repos are the primary source of funding for market-based banking institutions. In a repurchase agreement, the borrower sells a security today for a price below the current market price on the understanding that it will buy it back in the future at a pre-agreed price. The difference between the current market price of the security and the price at which it is sold is called the “haircut” in the repo and fluctuates together with funding conditions in the market. The fluctuations in the haircut largely determine the degree of funding available to a leveraged institution. The reason is that the haircut determines the maximum permissible leverage achieved by the borrower. If the haircut is 2%, the borrower can borrow $98 for $100 worth of securities pledged. Then, to hold $100 worth of

08 Book.indb 299

2/13/09 3:58:42 PM

300

Tobias Adrian and Hyun Song Shin

securities, the borrower must come up with $2 of equity. Thus, if the repo haircut is 2%, the maximum permissible leverage (ratio of assets to equity) is 50. Suppose that the borrower leverages up the maximum permitted level. Such an action would be consistent with the objective of maximizing the return on equity, since leverage magnifies return on equity. The borrower thus has a highly leveraged balance sheet with leverage of 50. If at this time, a shock to the financial system raises the market haircut, then the borrower faces a predicament. Suppose that the haircut rises to 4%. Then, the permitted leverage halves to 25, from 50. The borrower then faces a hard choice. Either it must raise new equity so that its equity doubles from its previous level, or it must sell half its assets, or some combination of both. Note that the increase in haircuts will do most harm when starting from very low levels. A percentage point increase from 1% to 2% will mean leverage has to fall from 100 to 50. But a percentage point increase from 20% to 21% will have only a marginal effect on the initial leverage of 5. In this sense, the “chasing of yield” at the peak of the financial cycle is especially precarious, since the unwinding of leverage will be that much more potent. Times of financial stress are associated with sharply higher haircuts, necessitating substantial reductions in leverage through asset disposals or raising of new equity. Raising new equity or cutting assets entails adjustments for the borrower. Raising new equity is notoriously difficult in distressed market conditions. But selling assets in a depressed market is not much better.9 The evidence from the scatter chart above is that borrowers tend to adjust leverage primarily through adjustments in the size of the balance sheet, leaving equity unchanged, rather than through changes in equity directly. For an investment bank, whose assets tend to be short term and liquid (such as short-term collateralized lending), it can adjust its balance sheet size flexibly by reducing lending and not rolling over debt. However, when the financial system as a whole holds longterm, illiquid assets financed by short-term liabilities, any tensions resulting from a sharp, synchronized contraction of balance sheets

08 Book.indb 300

2/13/09 3:58:42 PM

Financial Intermediaries, Financial Stability and Monetary Policy

301

will show up somewhere in the system. Even if some institutions can adjust down their balance sheets flexibly, there will be pinch points in the system that will be exposed by the de-leveraging. We return to this issue below. 3. Macroeconomic Consequences In models of monetary economics that are commonly used at central banks, the role of financial intermediaries is largely incidental; banks and broker-dealers are passive players that the central bank uses to implement monetary policy. In contrast, our argument thus far suggests that they deserve independent study because of their impact on financial conditions and hence on real economic outcomes. In this section, we examine whether financial intermediaries’ impact on financial conditions feeds through to affect real economic outcomes—in particular, on components of GDP. We find that it does, especially on those components of GDP that are sensitive to credit supply, such as housing investment and durable goods consumption. In the language of “frictions,” our empirical findings are consistent with a set of principal-agent frictions that operate at the level of the financial intermediaries themselves. These frictions result in constraints on balance sheet choice that bind harder or more loosely depending on the prevailing market conditions. The fluctuations in haircuts and regulatory capital ratios that are critical in determining the leverage of financial intermediaries can be seen as being driven by the fluctuations in how hard these constraints bind. When balance sheet constraints bind harder, credit supply is reduced. Broker-dealer balance sheets hold potentially more information on underlying financial conditions, as they are a signal of the marginal availability of credit. At the margin, all financial intermediaries (including commercial banks or GSEs) have to borrow in markets (for instance via commercial paper or repos). For a commercial bank, even though only a small fraction of its total balance sheet is market based, at the margin it has to tap the capital markets. But for commercial banks, their large balance sheets mask the effects operating at the margin. Broker-dealers, in contrast, give a purer signal of marginal funding

08 Book.indb 301

2/13/09 3:58:42 PM

302

Tobias Adrian and Hyun Song Shin

conditions, as their liabilities are short term, and their balance sheets are close to being fully marked to market. In addition, broker-dealers originate and make markets for securitized products, whose availability determines the credit supply for consumers and non-financial firms (e.g. for mortgages, car loans, student loans, etc.). So broker-dealers are important variables for two reasons. First, they are the marginal suppliers of credit. Second, their balance sheets reflect the financing constraints of the market-based financial system. To the extent that balance sheet dynamics affect the supply of credit, they have the potential to affect real economic variables. To demonstrate that there are indeed real effects of the balance sheet behavior of intermediaries, we estimate macroeconomic forecasting regressions. In Table 1, we report the results of regressions of the annual growth rate of GDP components on lagged macroeconomic and financial variables. In addition, we add the lagged growth rate of total assets and market equity of security broker-dealers on the right-hand side.10 By adding lags of additional financial variables on the right-hand side (equity market return, equity market volatility, term spread, credit spread), we offset balance sheet movements that are purely due to a price effect. By adding the lagged macroeconomic variables on the right-hand side, we control for balance sheet movements due to past macroeconomic conditions. In Table 1 (and all subsequent tables), * denotes statistical significance at the 10%, ** significance at the 5% level, and *** at the 1% level. All our empirical analysis is using quarterly data from 1986Q1 to 2008Q1. Variable definitions are given in the appendix. The growth rate of security broker-dealer total assets has strongest significance for the growth rate of future housing investment and for durable good consumption (Table 1, columns iv and ii, respectively). Our interpretation of this finding is that the mechanisms that determine the liquidity and leverage of broker-dealers affect the supply of credit, which in turn affects investment and consumption. The finding that dealer total assets significantly forecast durable but not total consumption, and that they forecast housing investment but

08 Book.indb 302

2/13/09 3:58:43 PM

08 Book.indb 303

-0.199 0.066

PCE core inflation (1Q lag)

Fed Funds Target (1Q lag)

0.008 0.018 0.180* -0.023 0.252

S&P500 Return (1Q lag)

S&P500 implied volatility VIX (1Q lag)

10-year/3-month spread (1Q lag)

Baa/10-year spread (1Q lag)

Constant

Financial Market Conditions

0.746***

Lag of left-hand side variable

Macroeconomic conditions

0.003 0.008**

Asset growth (1Q lag)

Equity growth (1Q lag)

Broker-Dealer Variables

1.111

-0.182

1.456**

0.075

-0.002

0.667

-2.225***

0.468***

0.013

0.048*

(ii) Durable Consumption (4Q growth)

(i) Consumption (4Q growth)

1.114

-1.492**

0.460

0.126**

0.039

-0.342***

0.247

0.873***

0.026**

-0.007

Investment (4Q growth)

(iii)

-7.078

0.367

0.972

0.183*

0.041

-0.253

0.344

0.829***

0.055***

0.062**

Housing Investment (4Q growth)

(iv)

0.238

-0.183

0.187**

0.026*

0.009

0.003

-0.112

0.812***

0.006*

0.005

GDP (4Q growth)

(v)

Table 1 Broker-Dealer Assets are Signifigant for Future Macroeconomic Growth

Financial Intermediaries, Financial Stability and Monetary Policy 303

2/13/09 3:58:43 PM

304

Tobias Adrian and Hyun Song Shin

not total investment, lends support to this interpretation, as durable consumption and housing investment could be seen as being particularly sensitive to the supply of credit. The market value of security broker-dealer equity also has predictive power for housing investment, but additionally forecasts total consumption, total investment and GDP. In Table 1, equity is market equity, rather than book equity. To the extent that shifts in market equity are good indicators of the shifts in the marked-to-market value of book equity, we can interpret the empirical finding that equity growth has real impact through the amplification mechanism illustrated in Chart 6, the scatter chart of leverage against assets. When balance sheets become strong, equity increases rapidly, eroding leverage. Financial intermediaries then attempt to expand their balance sheets to restore leverage. Since our data are quarterly, but balance sheets adjust quickly, the one quarter lagged assets may not fully capture this effect. However growth in market equity may be a good signal of growth of spare balance sheet capacity. The forecasting power of dealer assets for housing investment is graphically illustrated in Chart 7. The impulse response function is computed from a first order vector autoregression that includes all variables of Table 1, Column (iv). The plot shows a response of housing investment to broker-dealer assets growth that is positive, large, and persistent. We next ask whether commercial bank balance sheet variables have additional forecasting power for real economic growth. One way to do so is to first orthogonalize commercial bank total asset growth and equity growth with respect to the broker-dealer variables, and then add the commercial banks variables that is unexplained by the broker-dealer variables to the regressions. The results are presented in Table 2. We find that commercial bank equity does have some additional information for housing investment, but total commercial bank assets do not. To further understand differences between the security brokerdealer and commercial bank balance sheet interactions with the macroeconomic aggregates, we run the same regressions as in Table 1,

08 Book.indb 304

2/13/09 3:58:43 PM

Financial Intermediaries, Financial Stability and Monetary Policy

305

Chart 7 Impulse Response Function of Housing Investment Growth to a Broker-Dealer Asset Growth Shock (in units of standard deviations) 2

1

1

Housing Investment Growth

Housing Investment Growth

2

0

0

Impulse Response to Broker Dealer Asset Growth Shock

-1

-1 0

5

Quarters

10

15

but with commercial bank variables instead of security broker-dealer variables (see Table 3). We do find that commercial bank (market) equity is significant in explaining real economic activity, but commercial bank total assets are not. Our interpretation of these findings is that commercial bank balance sheets are less informative than broker-dealer balance sheets as they (largely) did not mark their balance sheets to market over the time span in our regressions. However, market equity is a better gauge of underlying balance sheet constraints, and so better reflects the marginal increases in balance sheet capacity. So, whereas growth in total assets do not signal future changes in activity, growth in market equity does. The finding that commercial bank assets do not predict future real growth is also consistent with Bernanke and Lown (1991), who use a cross-sectional approach to show that credit losses in the late 80s and early 90s do not have a significant impact on real economic growth across states. See Kashyap and Stein (1994) for an overview of the debate on whether there was a “credit crunch” in the recession in the early 1990s.

08 Book.indb 305

2/13/09 3:58:43 PM

08 Book.indb 306

0.009**

Equity growth (1Q lag)

0.092

Fed funds target (1Q lag)

0.006 0.020 0.232 * -0.088 0.339

S&P500 return (1Q lag)

S&P500 volatility VIX (1Q lag)

10-year/3-month (1Q lag)

Baa/10-year (1Q lag)

Constant

Financial Market Conditions

-0.199

PCE core inflation (1Q lag)

Lag of left hand side variable (1Q lag)

0.688 ***

0.004

Macroeconomic conditions

0.060

Asset growth (1Q lag)

Equity growth (1Q lag)

Commercial Bank Variables (Orthogonalized with respect to Broker-Dealer Variables)

0.002

1.426

-0.658

1.636 **

0.081

-0.011

0.716

-2.114 ***

0.418 ***

0.047

0.353

0.015

0.050 *

Durable Consumption

Consumption

Asset growth (1Q lag)

Broker-Dealer Variables

(ii)

(i)

1.315

-1.576 **

0.452

0.125 **

0.037

-0.341 ***

0.258

0.866 ***

0.011

0.038

0.026 **

-0.007

Investment

(iii)

-5.618

0.388

0.542

0.171 *

0.031

-0.375

0.395

0.812 ***

0.088 ***

-0.045

0.057 ***

0.054 **

Housing Investment

(iv)

0.944

-0.516 **

0.167

0.036 *

0.011

-0.038

-0.022

0.770 ***

0.005

0.027

0.007 *

-0.001

GDP

(v)

Table 2 Commercial Bank Assets do not have Additional Explanatory Power for Real Activity (Except for Housing Investment) 306 Tobias Adrian and Hyun Song Shin

2/13/09 3:58:43 PM

08 Book.indb 307

0.009 ***

0.714 ***

Equity growth (1Q lag)

Lag of left-hand side variable (1Q lag)

0.084

Fed funds target (1Q lag)

-0.128 0.080

Baa/10-year (1Q lag)

Constant

0.211 ***

0.017

10-year/3-month (1Q lag)

0.007

S&P500 return (1Q lag)

S&P500 volatility VIX (1Q lag)

Financial Market Conditions

-0.200

PCE core inflation (1Q lag)

Macroeconomic conditions

0.063

Asset growth (1Q lag)

Commercial Bank Variables

-0.779

-0.232

1.397 **

0.067

-0.021

0.642

-1.907 ***

0.412 ***

0.048 **

0.329

(ii) Durable Consumption

(i) Consumption

1.239

-1.741

0.482

0.126 **

0.043

-0.333

0.201

0.882 ***

0.022 **

0.055

Investment

(iii)

0.815

-0.536

0.166

0.035 **

0.012

-0.042

-0.033

0.792 ***

0.007 ***

0.024

Housing Investment

(iv)

Table 3 Commercial Bank Equity has Explanatory Power ... but Commercial Bank Assets do not (v)

0.165

-0.244 **

0.195

0.027 **

0.008

0.004

-0.113

0.785 ***

0.011 **

0.039

GDP

Financial Intermediaries, Financial Stability and Monetary Policy 307

2/13/09 3:58:43 PM

308

Tobias Adrian and Hyun Song Shin

In the same vein, Ashcraft (2006) finds small effects of variations— in commercial bank loans on real activity when using accounting based loan data, but Ashcraft (2005) finds large and persistent effects of commercial bank closures on real output (using FDIC induced failures as instruments). Morgan and Lown (2006) show that the senior loan officer survey provides significant explanatory power for real activity—again a variable that is more likely to reflect underlying credit supply conditions and is not based on accounting data.   The credit supply channel sketched so far differs from the financial amplification mechanisms of Bernanke and Gertler (1989), and Kiyotaki and Moore (1997, 2005). These papers focus on amplification due to financing frictions in the borrowing sector, while we focus on amplification due to financing frictions in the lending sector. Our approach raises the question of whether the failure of the Modigliani-Miller theorem may be more severe in the lending rather than the borrowing sector of the economy. The interaction of financial constraints in the lending and the borrowing sector is likely to give additional kick to financial frictions in the macro context that mutually reinforce each other. These interactions would be fertile ground for new research. We should also reiterate the caveats that underpin the results in Table 2. Inference for macroeconomic aggregates is difficult as all variables are endogenous. In analyzing the data, we started with the prior that balance sheets of financial intermediaries should matter for real economic growth. This prior has guided our empirical strategy. Researchers who look at the data with a different prior will certainly be able to minimize the predictive power of the broker-dealer balance sheet variable. However, analyzing the data with the prior that financial intermediary frictions are unimportant has the potential cost of overlooking the friction. Further searching examinations of the data will help us uncover the extent to which financial variables matter. In addition, we have not analyzed the importance of the balance sheets of other institutions of the market-based financial system, such as the GSEs, hedge funds, etc.

08 Book.indb 308

2/13/09 3:58:43 PM

Financial Intermediaries, Financial Stability and Monetary Policy

309

We now present some additional evidence of the impact of brokerdealer assets from vector autoregressions that summarize the joint dynamics of macroeconomic variables, broker-dealer variables, and monetary policy. Chart 8 is deliberately constructed to resemble the impulse response functions of Bernanke and Gertler (1995). In this exercise, we make no structural assumptions and instead examine the embedded empirical relationships by conducting a VAR exercise in the spirit of Sims (1980). To illustrate the impact of adding financial institutions to the monetary policy transmission mechanism, we plot the impulse response functions of housing investment growth from two vector autoregressions. In the first VAR, only GDP growth, PCE inflation, the federal funds target, and housing investment are included (in that order). The second VAR adds the security broker dealer variables to the macro variables, with the macro variables ordered before the financial institution variables. Each VAR is nonstructural and includes four lags of all variables. By adding the financial institution variables after the baseline macroeconomic variables, we are being conservative, giving the financial institution variables the least possible chance to impact shocks to the fed funds target on housing investment. Each VAR is estimated with four lags, from 1986Q1— 2008Q1. The impulse response functions are plotted in Chart 8. Chart 8 shows that the dynamics of housing investment in response to monetary policy shocks differs in the two VAR specifications. The drop in housing investment in response to a fed funds target increase is both quicker and larger in the VAR with the broker-dealer variables, compared to the baseline model. However, the recovery is also quicker. The two response functions of Chart 8 again illustrate that balance sheet variables of financial institutions have quantitatively important effects on macroeconomic dynamics. We interpret Chart 8 as showing that the market-based financial system has a noticable impact on the monetary policy transmission mechanism. 4. Determination of Broker-Dealer Balance Sheets Having established that broker-dealer balance sheets matter for real activity, we investigate what determines the growth of broker-dealer balance sheets.11 Broker-dealers fund themselves with short-term debt

08 Book.indb 309

2/13/09 3:58:43 PM

310

Tobias Adrian and Hyun Song Shin

Chart 8 Response of Housing Investment to Fed Funds Shock (in Units of Standard Deviations). Comparison of Nonstructural Models with and without Broker-Dealer Variables .5

.5

0

0

-.5

-.5

-1

-1

Baseline VAR

Housing Investment Growth

Housing Investment Growth

VAR with Broker-Dealer Variables

-1.5

-1.5

0

5

10

15

Quarters

(primarily repurchase agreements and other forms of collateralized borrowing). Part of this funding is directly passed on to other leveraged institutions such as hedge funds in the form of reverse repos.12 Another part is invested in longer term, less liquid securities. The cost of borrowing is therefore tightly linked to short term interest rates in general, and the federal funds target rate in particular. Broker-dealers hold longer term assets, so that proxies for expected returns of brokerdealers are spreads—either credit spreads, or term spreads. Leverage is constrained by risk; in more volatile markets, leverage is more risky and credit supply can be expected to be more constrained. Increases in the fed funds target rate are generally associated with a slower growth rate of broker-dealer assets. The first four columns of Table 4 show that this finding holds contemporaneously (column i and ii), it holds with a lag (column iii), and it holds in a forwardlooking sense (column iv). Put differently, expectations of increases in the fed funds target are associated with declines in dealer assets, as are contemporaneous changes. Interestingly, we do not find that commercial bank total asset growth is significantly explained by changes in the federal funds target. This finding is again consistent with two

08 Book.indb 310

2/13/09 3:58:44 PM

08 Book.indb 311

-1.060

PCE Core Inflation (1Q lag)

4.871 **

Baa/10-year (1Q lag)

36.787 ***

-5.697

10-year/3-month (1Q lag)

Constant

-0.013

-1.125 ***

S&P500 Return (1Q lag)

S&P500 Volatility VIX (1Q lag)

Financial Market Variables

0.553

-15.87 ***

Real GDP Growth (1Q lag)

Macroeconomic Conditions

(spread to fed funds)

1-year Eurodollar future

Target (lag, 1Q change)

Target (4Q change)

Target (1Q change)

Fed Funds

37.538 ***

18.131 ***

-3.027

-1.220 ***

0.118

-1.065

0.833

-4.87 ***

(ii) Asset Growth Broker-Dealers

(i)

Asset Growth Broker-Dealers

40.102 ***

14.781

2.503

-0.941 ***

0.159

0.284

-0.740

-11.65 ***

Asset Growth Broker-Dealers

(iii)

36.550 ***

23.398 ***

-2.242

-1.178 ***

0.151

-0.789

0.679

-15.42 ***

Asset Growth Broker-Dealers

(iv)

7.112 ***

0.216

-0.426

0.040

0.013

-0.966 ***

0.516 ***

0.071

Asset Growth Commercial Banks

(v)

Table 4 Increases in the Federal Funds Target (and the Expectation of the Future Target) Tend to Reduce Broker-Dealer Balance Sheets

Financial Intermediaries, Financial Stability and Monetary Policy 311

2/13/09 3:58:44 PM

312

Tobias Adrian and Hyun Song Shin

explanations. Either commercial bank balance sheets do not reflect the current positions of assets and liabilities properly, as their balance sheets have historically not been marked to market, or commercial banks do not manage their balance sheet as actively. In all of the regressions of Table 4, we add GDP growth and PCE core inflation on the right-hand side. Interestingly, neither growth nor inflation are significant determinants of broker-dealer total asset growth. It appears that the level of the fed funds target is sufficient in capturing all of the macroeconomic information that is relevant for broker-dealers. We again find that commercial banks differ sharply in this respect; while their asset growth is not significantly determined by the fed funds target, it is significantly positively correlated with GDP growth and negatively with core PCE inflation. Financial market volatility, as measured by the VIX index of implied volatility, relates negatively to security broker-dealer asset growth, as higher volatility is associated with higher margins and tighter capital constraints, both inducing tighter constraints on dealer leverage. Credit spreads are positively related to dealer asset growth, as they proxy the profitability of holding risky, illiquid, longer maturity assets. Broker-dealers trade actively, so it would be desirable to study their balance sheet behavior at higher frequencies. Fortunately, the Federal Reserve Bank of New York collects financing data of the so-called Primary Dealers at a weekly frequency. Adrian and Shin (2007) document that dealer total asset growth is tightly linked to dealer repo growth, as expansions and contractions of broker-dealer total assets are primarily financed by expansions and contractions in repos. In Table 5, we explain Primary Dealer repo borrowing by the same variables as in Table 4 (except for GDP and inflation, which we had seen were insignificant anyway, and which are not available at a weekly frequency). We use 13-week changes and lags in the regression in order to pick up correlations that occur at the same frequency as the quarterly data. We again find the negative comovement of dealer balance sheets with changes in the fed funds target, and we additionally uncover a positive relation between dealer repos and the term spread.

08 Book.indb 312

2/13/09 3:58:44 PM

Financial Intermediaries, Financial Stability and Monetary Policy

313

Table 5 Primary Dealer Repo Growth Expands when the Term Spread is Large Repo Growth Primary Dealers Fed Funds (13 week change)

-0.037 **

Fed Funds (13 week lag)

0.037 ***

S&P500 Return (13 week)

0.000 *

S&P500 (13 week lag)

0.000 ***

VIX (13 week change)

-0.001

VIX (13 week lag)

-0.007 ***

10-year / 3-month Treasury spread (13 week change)

0.049 **

10-year / 3-month Treasury spread (13 week lag)

0.087 ***

Baa / 10-year credit spread (13 week change)

0.150 ***

Baa / 10-year credit spread (13 week lag)

0.017

Repo Growth (13 week lag)

-0.242 ***

Constant

-0.163

5. Monetary Policy and Balance Sheets We have seen in the previous two sections that dealer asset growth and market equity of broker-dealers and commercial banks explain future real growth of macroeconomic aggregates such as durable consumption and housing investment. In addition, we have seen that changes in the federal funds target rate, and expectations of the future path of policy, are important determinants in broker-dealer total asset growth. Changes to the federal funds target are further primarily determined by real growth and inflation (see Taylor, 1993). So how does monetary policy interact with the waxing and waning of financial intermediary balance sheets? In financial crises, the tight connection between balance sheets and monetary policy certainly becomes apparent. In the fall of 1987 and again in the fall of 1998, the fed funds target was cut in order to insulate the real economy from financial sector distress. While this interaction between monetary policy and financial sector distress is apparent in crises, what is the relationship between the two in normal times?

08 Book.indb 313

2/13/09 3:58:44 PM

314

Tobias Adrian and Hyun Song Shin

We find that higher balance sheet growth of broker assets is associated with a lower fed funds target (see Table 6, column ii), except in crisis, when the sign reverses (see column iii). We also find that increases in dealer balance sheets tend to precede a lower fed funds target in the next quarter. One explanation for these findings may be the slow adjustment of the fed funds rate, and the market’s anticipation of such slow adjustment. Once the interest rate cycle turns, banks expect more moves in the same direction in the future. Anticipating such future moves, banks expand balance sheets following initial cuts in the fed funds rate. Then, the anticipated future cuts materialize. In the time series, the realized subsequent cuts trail the balance sheet expansions. In Chart 9, we plot the impulse response function of the fed funds target rate to a one standard deviation shock to the growth rate of security broker-dealer assets. We include the fed funds target, real GDP growth, PCE core inflation, broker-dealer asset growth, and the interaction of broker-dealer asset growth with the crisis dummy in the VAR.13 The left-hand panel draws the impulse response in crisis periods; the right-hand side draws it in normal times. Note that Chart 9 is drawn as the impulse responses of the fed funds target to a positive broker-dealer asset growth shock. The left-hand panel is familiar from the 1987 crash and the 1998 crisis; the fed funds target was cut aggressively in response to financial sector distress. The right-hand panel of the impulse response function is less familiar: It shows the procyclicality of fed funds policy relative to dealer balance sheet growth.14 As we outlined in section 2, when the relation between financial markets and monetary policy is discussed, the conclusion is often drawn that policy should only incorporate financial market variables insofar as they help to predict future macroeconomic variables such as future output and future inflation. It could be that broker-dealer asset growth in the policy rule is only signifigant because it reflects movements in asset prices that are helpful in predicting future macroeconomic variables. Column (i) of Table 7 shows that the significance of security brokerdealer asset growth is unaffected when additional asset price controls

08 Book.indb 314

2/13/09 3:58:44 PM

Financial Intermediaries, Financial Stability and Monetary Policy

315

Table 6 Monetary Policy is Pro-Cyclical Relative to Broker-Dealer Asset Growth ... Except in Crises (i) Fed Funds Target (change)

(ii) Fed Funds Target (change)

(iii) Fed Funds Target (change)

(iv) Fed Funds Target (change)

-0.070 **

-0.073 ***

Macroeconomic Conditions Fed funds target (1Q lag)

-0.090 ***

-0.070 ***

Real GDP growth

0.240 ***

0.220 ***

0.230 ***

0.229 ***

PCE core inflation

0.130 **

0.110 *

0.140 **

0.123 **

-0.007 **

-0.009 ***

Broker-Dealer Balance Sheets Asset growth Asset growth * crisis dummy

0.038 **

Asset growth (1Q lag)

-0.008 **

Constant

-0.698 ***

-0.567 ***

-0.587 ***

-0.573 ***

Chart 9 Impulse Response of the Fed Funds Target to a (Positive) Shock to Security Broker-Dealer Asset Growth in Crisis and in Normal Times Normal Times

.2

.2

0

0

-.2

-.2

-.4

-.4

Fed Funds Target Response

Fed Funds Target Response

Crisis

-.6

-.6 0

5

10

15

0

5

10

15

Quarters Impulse Response to Broker-Dealer Asset Growth Shock

08 Book.indb 315

2/13/09 3:58:44 PM

316

Tobias Adrian and Hyun Song Shin

are included in the regression. In column (ii) of Table 7, we first regress security broker-dealer asset growth on four lags of PCE core inflation and GDP growth and then add the predicted value from that regression to the right-hand side of the policy rule regression. We see that security broker-dealer assets and the value of security broker-dealer assets that is explained by past macroeconomic variables are both significant, so the security broker-dealer variable is not significant simply because it correlates with past macroeconomic variables. In column (iii), we do the reverse; we first regress GDP growth and PCE core inflation on four lags of security broker-dealer growth and add the predicted value of those regressions to the right-hand side. This captures the degree to which asset growth is forecasting future macroeconomic activity. We again find that the asset growth variable becomes more significant and larger in magnitude, and the predicted values from the first stage regression are not significant. Finally, in the last column of Table 7, we show that commercial bank total asset growth is not significant in the policy rule regression. This again suggests that the transmission mechanism of monetary policy should take the liquidity and leverage of market-based financial intermediaries explicitly into account. We do not interpret the results of Tables 6 and 7 as saying that the balance sheets of broker-dealers are the only relevant measure of financial intermediary liquidity and leverage. There are other leveraged institutions (such as GSEs, hedge funds, to name but a few) whose potential to affect the economy have not been examined here. Our focus on broker-dealers is motivated by the hypothesis that they provide a useful window on the market-based financial system. We do not advocate any particular monetary policy rule that targets balance sheet growth. Considerations of Goodhart’s Law and the Lucas critique will be relevant here as for any simplistic macro policy rule. This point is especially relevant given our observation earlier that the association between balance sheet dynamics and fed funds dynamics may be due to banks’ anticipation of future fed funds changes. One way we can visualize the policy response in setting the fed funds rate to growth in the broker-dealer balance sheet is to compute

08 Book.indb 316

2/13/09 3:58:44 PM

08 Book.indb 317

-0.114* -0.079 0.380

10-year / 3-month (lag)

Baa / 10-year (lag)

Constant

0.009 -0.021

S&P500 return

S&P500 volatility VIX

Financial Markets

CB asset growth (lag)

BD asset growth (explained by four lags of GDP & PCE)

BD asset growth (lag)

BD asset growth

Balance Sheet Variables

-0.012***

-0.096

-0.087

-0.021*

0.01

-0.014

-0.014*

-0.009**

0.187**

0.677

-0.176

-0.113

-0.017

0.011*

-0.014***

-0.239

0.141

0.192***

-0.146**

PCE forecasted by four lags of broker asset growth (lag)

0.200**

PCE core inflation (4Q)

0.189***

-0.116**

0.191

0.215***

(iii) Fed Funds Target (change)

GDP forecasted by four lags of broker asset growth (lag)

-0.135**

Fed funds target (lag)

Real GDP growth (4Q)

Macro Variables

(ii) Fed Funds Target (change)

(i) Fed Funds Target (change)

-0.966***

0.029

0.169**

0.233***

-0.084***

Fed Funds Target (change)

(iv)

Table 7 Pro-Cyclical Monetary Policy is Robust to Asset Price Controls and Controls for Future and Past Macroeconomic Variables

Financial Intermediaries, Financial Stability and Monetary Policy 317

2/13/09 3:58:44 PM

318

Tobias Adrian and Hyun Song Shin

the residual relative to a benchmark Taylor rule, and then plot the residual of the Taylor rule together with the series for the growth in broker-dealer assets. Such plots give an alternative representation of the regression results in Table 7. Chart 10 provides a panel of such plots. The bottom left panel plots the Taylor rule residuals from Table 3, column (i), which best fits the observed fed funds target. The bottom right panel gives the residual from William Poole’s rule, as sketched in Poole (2005). We see the negative correlation between Taylor rule residuals and balance sheet growth clearly in the data.15 6. Implications for Monetary Policy In conventional monetary theory, the primary friction is the price stickiness of goods and services.16 Financial intermediaries do not play a role in these models other than as a passive player that the central bank uses to implement policy. Our findings suggest the need to give these players an independent role. Quantity variables seem to matter—especially components of financial intermediary balance sheets. Using the language of “frictions,” our results suggest a second friction in addition to sticky prices. This second friction originates in the agency relationships embedded in the organization of market– based financial intermediaries, which are manifested in the way that financial intermediaries manage their balance sheets.17 This is a friction in the supply of credit. We are certainly not the first to study frictions in the supply of credit. There has been an extensive discussion of financial frictions within monetary economics (see, for example, the overview by Bernanke and Gertler, 1995). However, it would be fair to say that financial frictions have received less emphasis in the last few years. One reason for the lack of emphasis may be that the earlier literature that focused on commercial bank balance sheets or borrowers’ balance sheets did not produce conclusive empirical results. We conclude from our own study that the time is now ripe to redress the balance and bring financial institutions back into the heart of monetary economics. When we examine an appropriate balance sheet measure that reflects the underlying funding conditions in capital markets, we stand a better chance of capturing the transmission

08 Book.indb 318

2/13/09 3:58:44 PM

Financial Intermediaries, Financial Stability and Monetary Policy

319

Chart 10 Broker-Dealer Asset Growth and Fed Funds Target, Fed Funds Futures, and Two Measures of Taylor Rule Residuals 70 60

60

2

1.5

50 1

e

50

20

-1

10

-2

0

Target (Annual Chang

30

-3 -10

Asset Growth (Annual Percent)

0

-4

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

-30

1

20 0.5

10 0

0

-10 -20

-5

-30

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

-20

30

Security Broker-Dealer Total Asset Growth (Percent Annual)

Fed Funds Target (Annual Change)

Fed Funds Future (One year ahead, in excess of current) 1.5

70

3

60

60

2

1 50

50 1

10

-2

0 -3 -10 -4

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

-5

30

0

20

-0.5

10 0

-1

-10

-1.5 -20 -30

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

-1

Asset Growth (Annual Percent)

20

Taylor Rule Residual

Asset Growth (Annual Percent)

30

Taylor Rule Residual

0

-20

08 Book.indb 319

0.5

40

40

-30

-0.5

Security Broker-Dealer Total Asset Growth (Percent Annual)

70

Future (excess to current)

40

40 Asset Growth (Annual Percent)

2

70

3

-2

Security Broker-Dealer Total Asset Growth (Percent Annual)

Security Broker-Dealer Total Asset Growth (Percent Annual)

Taylor Rule Residual (Specification of William Poole)

Taylor Rule Residual (Specification of Table 3, Column (i))

2/13/09 3:58:45 PM

320

Tobias Adrian and Hyun Song Shin

mechanism through credit supply more fully. In our view, the appropriate balance sheet quantities are those that are marked to market, and hence reflect current market conditions. In this regard, we have seen that broker-dealer assets are more informative than commercial bank assets, and market equity of either commercial banks or brokerdealers do a better job of explaining future activity than (book) asset values. As commercial banks begin to mark more items of their balance sheets to market, commercial bank balance sheet variables are likely to become more important variables for studying the transmission mechanism. Fluctuations in the supply of credit arise from how much slack there is in balance sheets. The cost of leverage of market-based intermediaries is determined by two main variables—risk and short term interest rates. The expected profitability of intermediaries is proxied by carry spreads such as the term spread and various credit spreads. Variations in the policy target determine short-term interest rates and have a direct impact on the profitability of intermediaries. When monetary policy is tightened at the end of an economic expansion, the slope of the yield curve becomes shallower and sometimes inverts. Intermediaries have to reduce the supply of credit when faced with a shallow yield curve.18 Deleveraging is particularly rapid when measured risks also increase. We have already argued that even small increases in repo haircuts can induce drastic reductions in leverage. As the economy slows, financial constraints may bind harder and prices fall more than in the absence of constraints. To the extent that financial intermediaries play a role in monetary policy transmission through credit supply, short-term interest rates matter directly for monetary policy. This perspective on the importance of the short rate as a price variable is in contrast to current monetary thinking, where short-term rates matter only to the extent that they determine long-term interest rates, which are seen as being risk-adjusted expectations of future short rates. Alan Blinder (1998, p. 70) in his book on central banking puts it in the following terms: “Central banks generally control only the overnight interest rate, an interest rate that is relevant to virtually no economically interesting transactions. Monetary policy has important

08 Book.indb 320

2/13/09 3:58:45 PM

Financial Intermediaries, Financial Stability and Monetary Policy

321

macroeconomic effects only to the extent that it moves financial market prices that really matter—like long-term interest rates, stock market values and exchange rates.” Current models in monetary economics emphasize the importance of managing market expectations. By charting a path for future short rates and communicating this path clearly to the market, the central bank can influence long rates and thereby influence mortgage rates, corporate lending rates and other prices that affect consumption and investment. The “expectations channel” has become an important consideration for monetary policy, especially among those that practice inflation targeting.19 Our approach entails quite different policy implications on some key issues. We mention three in particular. One has to do with forward-looking guidance on future policy rates or the publication of the central bank’s own projections of its policy rate. Such communication not only has implications for market participants’ expectations of the future path of short rates, but also for the uncertainty around that path. If central bank communication compresses the uncertainty around the path of future short rates, the risk of taking on long-lived assets financed by short-term debt is compressed. If the compression increases the potential for a disorderly unwinding later in the expansion phase of the cycle, then such compression of volatility may not be desirable for stabilization of real activity. In this sense, there is the possibility that forwardlooking communication can be counterproductive.20 Secondly, there is a case for rehabilitating some role for balance sheet quantities for the conduct of monetary policy. Ironically, our call comes even as monetary aggregates have fallen from favor in the conduct of monetary policy (see Friedman, 1988). The instability of money demand functions that makes the practical use of monetary aggregates challenging is closely related to the emergence of the market-based financial system. As a result of those structural changes, not all balance sheet quantities will be equally useful. The money stock is a measure of the liabilities of deposit-taking banks, and so may have been useful before the advent of the market-based financial system. However, the money stock will be of less use in a financial system

08 Book.indb 321

2/13/09 3:58:45 PM

322

Tobias Adrian and Hyun Song Shin

such as that in the US. More useful may be measures of collateralized borrowing, such as the weekly series on repos of primary dealers. Finally, our results highlight the way that monetary policy and policies toward financial stability are linked. When the financial system as a whole holds long-term, illiquid assets financed by short-term liabilities, any tensions resulting from a sharp pullback in leverage will show up somewhere in the system. Even if some institutions can adjust down their balance sheets flexibly, there will be some who cannot. These pinch points will be those institutions that are highly leveraged, but who hold long-term illiquid assets financed with shortterm debt. When the short-term funding runs away, they will face a liquidity crisis. The traditional lender of last resort tools (such as the discount window), as well as the recent liquidity provision innovations are tools that mitigate the severity of the tightening of balance sheet constraints. However, experience has shown time and again that the most potent tool in relieving aggregate financing constraints is a lower target rate. Past periods of financial stress, such as the 1998 crisis, was met by reduction in the target rate, aimed at insulating the real economy from financial sector shocks. In conducting monetary policy, the potential for financial sector distress should be explicitly taken into account in a forward-looking manner. In analyzing the interaction between financial stability and monetary policy in the time series of the last decades, it is important to keep in mind that the interaction of monetary policy and lender-oflast resort provision was successful in insulating the real economy from financial sector distress. So one can—in our view falsely—conclude that policies toward financial stability are more or less orthogonal to monetary policy analysis. To put it into Bayesian language, when analyzing the data (either via econometric or via structural approaches) with the prior that monetary policy and financial stability are orthogonal, one runs the risk of confirming that prior all too easily. The events of the past twelve months have clearly shown that now is the right time to reset the prior and to rethink the monetary policy transmission mechanism in a market-based financial system. The lesson for financial regulation is that the current risk-based capital requirements are powerless against the pull-back in lending

08 Book.indb 322

2/13/09 3:58:45 PM

Financial Intermediaries, Financial Stability and Monetary Policy

323

that arises from a system-wide deleveraging. When there are spillover effects, actions that enhance the soundness of one institution may end up by undermining another. The prudent curtailing of exposures by the creditors of Bear Stearns will be a run from the point of view of Bear Stearns itself. Secondly, even very safe assets such as reverse repos may be systemically important in that withdrawal of funding creates spillover effects on others.21 As well as the implications for prudential regulation, balance sheet dynamics imply a role for monetary policy in ensuring financial stability. The waxing and waning of balance sheets have both a monetary policy dimension in terms of regulating aggregate demand, but it has the crucial dimension of ensuring the stability of the financial system. Contrary to the common view that monetary policy and policies toward financial stability should be seen separately, they are inseparable. At the very least, there is a strong case for better coordination of monetary policy and policies toward financial stability. 7. Concluding Remarks Financial intermediaries lie at the heart of both monetary policy transmission as well as policies toward financial stability. The key thread to our discussion has been that the interaction of financial intermediaries’ balance sheet management with changes in asset prices and measured risks represents an important component in the transmission mechanism of monetary policy. The current credit crisis has the distinction of being the first postsecuritization crisis, where the market-based banking system has come into its own and has exerted a profound influence in the playing out of events in the financial markets and the wider economy over the last twelve months. We have shown that financial intermediary balance sheet management matters for the real economy, as well as for the soundness of the financial system. There are also important lessons for the conduct of monetary policy—some of them at variance with the current mainstream views on how monetary policy should be conducted. Due to their interaction with the leverage constraints of financial

08 Book.indb 323

2/13/09 3:58:45 PM

324

Tobias Adrian and Hyun Song Shin

intermediaries, short rates are important prices in their own right, and a smaller term premium is associated with contractions in the supply of credit. Our discussion suggests that tracking measures of financial market liquidity derived from the balance sheets of intermediaries has some information value in the conduct of monetary policy. The security broker-dealer assets are a key variable, but certainly not the only balance sheet variable that has the potential to be systemically important. Many other intermediaries use high degrees of leverage and have the potential for disorderly deleveraging. In addition, the economics of commercial banking is becoming more similar to the economics of broker-dealers as their balance sheets are marked to market to a greater degree. An important lesson of our study is that asset prices alone are likely not to be sufficient to summarize the conditions of intermediaries. As a result, balance sheet dynamics are informative both on key components of GDP as well as the resilience of the financial system. Monetary policy and policies toward financial stability are therefore two sides of the same coin.

Authors’ note: We thank the discussant, John Lipsky, and other participants for their comments at the conference. We also thank Adam Ashcraft, Alan Blinder, Markus Brunnermeier, Terrence Checki, William Dudley, Michael Fleming, Mark Gertler, Charles Goodhart, Jan Hatzius, Charles Himmelberg, Bengt Holmström, Anil Kashyap, Nobuhiro Kiyotaki, Meg McConnell, Jamie McAndrews, Don Morgan, Simon Potter, Jeremy Stein, Joseph Tracy, and Michael Woodford for their comments on earlier drafts, and Jiang Wang for earlier discussions. The views expressed in this paper are those of the authors alone and not necessarily those of the Federal Reserve Bank of New York or the Federal Reserve System.

08 Book.indb 324

2/13/09 3:58:45 PM

Financial Intermediaries, Financial Stability and Monetary Policy

325

Appendix: Data Sources and Variable Definitions Chart 1: US Flow of Funds, Board of Governors of the Federal Reserve. 1980Q1-2008Q1. Table L.218, home mortgage assets for various institutions. Chart 2: Source data as in Chart 1. Bank-based institutions: Commercial banking, Savings institutions, Credit unions. Market based institutions: Government-sponsored enterprises, Agency-and GSEbacked mortgage pools, Issuers of asset-backed securities. Charts 3-5: US Flow of Funds, Board of Governors of the Federal Reserve. 1980Q1-2008Q1. Security brokers and dealers, total financial assets, table L.129. Commercial Banks, Total Financial Assets, sum of tables L.110 and L.113. Households, total financial assets, table L.100. Chart 6: 10-K and 10-Q filings of the US Securities and Exchange Commission for Bear Stearns (1997Q1-2008Q1), Goldman Sachs (1998Q4-2008Q1), Lehman Brothers (1994Q4-2008Q1), Merrill Lynch (1992Q1-2008Q1), and Morgan Stanley (1997Q2-2008Q1). Leverage is the ratio of total assets to total stockholders equity. Quarterly growth rates of total assets and leverage are aggregated by weighting by total assets of the previous quarter. Chart 7: Impulse response function computed from a first order vector autoregression of security broker-dealer asset growth, security broker-dealer equity return, housing investment, PCE core inflation, Fed funds target, S&P500 return, S&P500 implied volatility VIX, 10-year/3-month spread, Baa/10-year spread. Variable definitions are given below. Chart 8: Impulse response function computed from a first order vector autoregression of fed funds target, GDP growth, PCE core inflation, security broker-dealer asset growth, security broker-dealer equity return, S&P500 return, S&P500 implied volatility VIX, 10-year/3-month spread, Baa/10-year spread. Variable definitions are given below.

08 Book.indb 325

2/13/09 3:58:45 PM

326

Tobias Adrian and Hyun Song Shin

Chart 9: In all four panels security broker-dealer growth, the federal funds target, and the federal funds future are as defined below. In the lower left hand panel, Taylor rule residuals are the residuals of a regression of Federal fund changes on the lagged federal funds target rate, current GDP growth, and current core PCE growth. These residuals are the residuals of the regression reported in Column (i) of Table 3. The Taylor rule residuals of the lower left hand panel correspond to the Taylor rule described in Poole (2005): Federal Funds = 1.5*(lagged core PCE core inflation - 1.5) + 0.5*Output Gap + 2.3 + 1.5 ). The output gap is computed as the percent deviation of real GDP (Bureau of Economic Analysis) from potential real GDP (Congressional Budget Office). Chart 10: The impulse response functions are computed from a vector autoregression (VAR) using a standard Cholesky decomposition. The baseline specification includes annual GDP growth, annual PCE inflation, the fed funds target, and annual housing investment growth. The specification with financial asset prices adds the market return, market volatility, the term spread, and the credit spread to the VAR (in that order, variable definitions given below). The VAR with security broker-dealers add annual broker-dealer total asset growth and broker-dealer annual equity growth to the macro variables (in that order). Tables 1-4, 6-7: Variable definitions. All variables are quarterly from 1986Q1-2008Q1. GDP growth denotes the annual percentage growth rate of real gross domestic product in chained 2000 US dollars, reported by the Bureau of Economic Analysis. Total consumption, durable consumption, total investment, and housing investment are the respective annual percentage growth rates as reported by the Bureau of Economic Analysis. Core PCE inflation is the annual percentage growth rate of the chained price index of personal consumption expenditures less food and energy, reported by the Bureau of Economic Analysis. The federal funds target is set by the Federal Open Market Committee and calculated as average over the quarter. The term spread is the difference between the 10-year Treasury constant maturity yield and the three month constant maturity yield from the Federal Reserve Board’s H.15 data release. The credit spread is the difference between Moody’s BAA

08 Book.indb 326

2/13/09 3:58:45 PM

Financial Intermediaries, Financial Stability and Monetary Policy

327

yield and the 10-year constant maturity yield, both from the from the Federal Reserve Board’s H.15 data release. The Federal funds future is from the one year Eurodollar rate published by the Financial Times. The S&P 500 return is the quarterly return reported by Standard and Poor’s. The S&P 500 volatility index is the VIX since 1991, and the VXO from 1986-1990, as reported by the Chicago Board Options Exchange. Security Broker and Dealers equity growth is the annual equity return from the Center for Research in Security Prices, according to the Standard Industrial Classification codes. Commercial bank equity is the annual equity return from the Center for Research in Security Prices, according to the Standard Industrial Classification codes. Security broker-dealer total asset growth is the annual growth rate of total financial assets from table L.129 of the Federal Reserve Board’s Flow of Funds. Commercial bank total asset growth is the annual percentage growth rate of the sum of total financial assets from tables L.110 and L.113 of the Federal Reserve Board’s Flow of Funds. The crisis dummy equals 1 in 1987Q4, 1994Q4, and 1999Q1, and 0 otherwise.22 Table 5: The Primary Dealer Repo series is the memorandum item Total Reverse Repurchase Agreements of Table 4 Financing by Primary US Government Securities Dealers from the weekly release of the FR2004 date by the Federal Reserve Bank of New York. The other variables in Table 4 are the same as the ones used in Tables 1-4 and 6-7, at a weekly frequency.

08 Book.indb 327

2/13/09 3:58:45 PM

328

Tobias Adrian and Hyun Song Shin

Endnotes See Greenlaw, et al. (2008). See BIS (2008 chapter 6), IMF (2008a), and Brunnermeier (2008) for an overview of recent events. 1

Kashyap, Rajan and Stein (2008) describe the incentives operating within the institution. See also Rajan (2005). 2

3 Arguments in favor of “leaning against the wind” of financial developments have been given by Blanchard (2000), Bordo and Jeanne (2002), Borio and Lowe (2002), Borio and White (2003), Cecchetti, Genberg, Lipsky and Wadhwani (2000), Cecchetti, Genberg and Wadhwani (2002), Crockett (2003) and Dudley (2006), Goodhart (2000) among others. The argument against is given in Bean (2003), Bernanke and Gertler (1999, 2001), Bernanke (2002), Greenspan (2002), Gertler (2003), Kohn (2005), Mishkin (2008) and Stark (2008).

A forthcoming IMF study (IMF, 2008b) shows how international differences in the incidence of market-based financial intermediaries results in markedly different relationships between asset growth and leverage growth. 4

Commercial and investment banks offset some of the mortgage related losses since the summer of 2007 via equity issuance. In addition, expansionary monetary policy offset some of need to unwind assets. The two red dots corresponding to 2007Q3 and 2007Q4 are below the 45-degree line, showing that not all losses were offset by new equity issuance. The determinants of equity issuance versus asset sales are further discussed in Adrian and Shin (2008b). 5

Gromb and Vayanos (2002) and Brunnermeier and Pedersen (2007) provide models of how balance sheet constraints interact with market developments. See also Kyle and Xiong (2001) and Morris and Shin (2004) for models of liquidity crises. The feedback effects will be larger when market liquidity effects reinforce the balance sheet constraints (Brunnermeier and Pedersen, 2007). 6

See Shin (2005) and Plantin, Sapra and Shin (2008a, 2008b).

7

8 Adrian and Shin (2008b) show how such behavior can be given theoretical rationale in terms of a contracting model between banks and their creditors.

The “margin spiral” described by Brunnermeier and Pedersen (2007) models this type of phenomenon. See also He and Krishnamurthy (2007) for the asset pricing consequences of constrained intermediary capital. 9

We use total asset growth of security broker-dealers as indicator of financial sector balance sheet growth. Deflating the asset growth by the core PCE inflation or household total asset growth does not change the results in this or later tables qualitatively. 10

Adrian and Shin (2008b) present microeconomic foundations for the variables that constrain the size and leverage of market-based financial intermediaries. 11

08 Book.indb 328

2/13/09 3:58:45 PM

Financial Intermediaries, Financial Stability and Monetary Policy

329

Adrian and Fleming (2005) analyze net collateralized financing as an indicator of primary dealer leverage, while Adrian and Shin (2007) focus on gross collateralized financing as an indicator of overall financial system leverage. 12

13 Note that column (iii) of Table 6 does not correspond directly to the VAR, as all variables enter contemporaneously in that Table, while all variables enter as lags in the VAR.

By aggressively cutting the target during times of financial intermediary distress, the Federal Reserve provides liquidity to the economy, which can be rationalized within the context of the Holmström and Tirole (1998) model. 14

Also see also Adrian and Shin (2008a).

15

See, for example, Woodford (2003).

16

These frictions are described in Rajan (2005) and Kashyap, Rajan and Stein (2008).

17

18 Adrian and Estrella (2008) explore further the signal value of the term spread for future macroeconomic outcomes in conjunction with monetary policy tightening cycles.

The expectations channel is explained in Blinder (1998), Bernanke (2004), Svensson (2004) and Woodford (2005). 19

Considerations of international monetary coordination reinforce this point. Hattori and Shin (2008) exhibit evidence that some of the expansions of intermediary balance sheets are financed through the yen carry trade, which rely on predictable discrepancies in short rates across currencies. 20

Morris and Shin (2008) examine issues for financial regulation in a system context. Shin (2008) is a case study of Northern Rock from a system perspective. 21

We use 1999Q1 instead of 1998Q4 as crisis quarter for the LTCM episode as the Flow of Funds data only show a negative growth in security-broker dealer assets in that quarter. In comparison, data from the SEC shows a decline already in 1998Q4 (see Figure 6). This difference is likely a data issue of the Flow of Funds relating to mergers and initial public offerings of some large broker-dealers. 22

08 Book.indb 329

2/13/09 3:58:45 PM

330

Tobias Adrian and Hyun Song Shin

References Adrian, Tobias, and Arturo Estrella (2008) “Monetary Tightening Cycles and the Predictability of Economic Activity,” Economics Letters 99, 260 – 264. Adrian, Tobias, and Michael Fleming (2005) “What Financing Data Reveal about Dealer Leverage,” Federal Reserve Bank of New York Current Issues in Economics and Finance, Volume 11 (3). Adrian, Tobias, and Hyun Song Shin (2007) “Liquidity and Leverage,” working paper, Federal Reserve Bank of New York and Princeton University. Available as Federal Reserve Bank of New York Staff Reports 328. Previous version “Liquidity and Financial Cycles,” presented at the Sixth BIS Annual Conference on Financial System and Macroeconomic Resilience, 18-19 June 2007, BIS Working Paper No. 256. Adrian, Tobias, and Hyun Song Shin (2008a) “Liquidity, Monetary Policy, and Financial Cycles,” Federal Reserve Bank of New York Current Issues in Economics and Finance, Volume 14 (1). Adrian, Tobias, and Hyun Song Shin (2008b) “Financial Intermediary Leverage and Value at Risk,” working paper, Federal Reserve Bank of New York and Princeton University. Federal Reserve Bank of New York Staff Reports, 338. Ashcraft, Adam (2005) “Are Banks Really Special? New Evidence from the FDIC-Induced Failure of Healthy Banks,” American Economic Review 95, pp. 1712-1730. Ashcraft, Adam (2006) “New Evidence on the Lending Channel,” Journal of Money, Credit, and Banking 38, pp. 751-776. Bank for International Settlements (2008) 78th Annual Report, Basel, Switzerland. Bean, Charles (2003) “Asset Prices, Financial Imbalances and Monetary Policy: AreInflation Targets Enough?” in Asset Prices and Monetary Policy, eds. Anthony Richards and Tim Robinson, Reserve Bank of Australia, pp.48-76. Bernanke, Ben (2002) “Asset-Price “Bubbles” and Monetary Policy,” remarks at New York Chapter of the National Association for Business Economics, October 15, 2002. http://www.federalreserve.gov/boarddocs/speeches/2002/20021015/ default.htm. Bernanke, Ben (2004) “The Logic of Monetary Policy,” remarks before the National Economists Club, December 2, 2004. www.federalreserve.gov/boarddocs/ speeches/2004/20041202/default.htm. Bernanke, Ben, and Mark Gertler (1989) “Agency Costs, Net Worth, and Business Fluctuations,” American Economic Review 79, pp. 14 - 31.

08 Book.indb 330

2/13/09 3:58:45 PM

Financial Intermediaries, Financial Stability and Monetary Policy

331

Bernanke, Ben, and Mark Gertler (1995) “Inside the Black Box: The Credit Channel of Monetary Policy Transmission,” Journal of Economic Perspectives 9, pp. 27-48. Bernanke, Ben, and Mark Gertler (1999) “Monetary Policy and Asset Volatility,” Federal Reserve Bank of Kansas City Economic Review 84, pp. 17-52. Bernanke, Ben, and Mark Gertler (2001) “Should Central Banks Respond toMovements in Asset Prices?” American Economic Review 91, pp. 253-257. Bernanke, Ben, and Cara Lown (1991) “The Credit Crunch,” Brookings Papers on Economic Activity 2, pp. 205-247. Blanchard, Olivier (2000) “Bubbles, Liquidity Traps, and Monetary Policy,” in: Japan’s Financial Crisis and its Parallels to the US Experience, edited by Ryoichi Mikitani and Adam Posen, Institute for International Economics Special Report 13. Blinder, Alan (1998) Central Banking in Theory and Practice, MIT Press, Cambridge. Bordo, Michael, and Olivier Jeanne (2002) “Monetary Policy and Asset Prices: Does Benign Neglect Make Sense?” International Finance 5, pp.139-164. Borio, Claudio, and Philip Lowe (2002) “Asset Prices, Financial and Monetary Stability: Exploring the Nexus,” Bank for International Settlements Working Paper 114. Borio, Claudio, and William White (2003) “Whither Monetary and Financial Stability? The Implications of Evolving Policy Regimes,” Proceedings of the Federal Reserve Bank of Kansas City Symposium at Jackson Hole 2003. http://www.kc.frb. org/publicat/sympos/2003/sym03prg.htm. Brunnermeier, Markus (2008) “De-Ciphering the Credit Crisis of 2007,” forthcoming, Journal of Economic Perspectives. Brunnermeier, Markus, and Lasse Pedersen (2007) “Market Liquidity and Funding Liquidity,” forthcoming Review of Financial Studies. Cecchetti, Stephen, Hans Genberg, John Lipsky and Sushil Wadhwani (2000) “Asset Prices and Central Bank Policy,” Geneva Reports on the World Economy 2, International Centre for Monetary and Banking Studies and Centre for Economic Policy Research. Cecchetti, Stephen, Hans Genberg and Sushil Wadhwani (2002) “Asset Prices in a Flexible Inflation Targeting Framework,” in Asset Price Bubbles: The Implications for Monetary, Regulatory and International Policies, eds. William Hunter, George Kaufman and Michael Pomerleano, MIT Press, pp. 427-444. Crockett, Andrew (2003) “International Standard Setting in Financial Supervision,” Institute of Economic Affairs Lecture, Cass Business School, London, 5 February.

08 Book.indb 331

2/13/09 3:58:45 PM

332

Tobias Adrian and Hyun Song Shin

Dudley, William (2006) Panel discussion, NBER conference on Monetary Policy and Asset Prices, May 5-6, 2006. http://www.nber.org/books_in_progress/ assetprices/dudley6-28-06.pdf. Friedman, Benjamin (1988) “Monetary Policy Without Quantity Variables,” American Economic Review 78, 440-45. Gertler, Mark (2003) “Commentary: Whither Monetary and Financial Stability? The Implications of Evolving Policy Regimes,” Proceedings of the Federal Reserve Bank of Kansas City Symposium at Jackson Hole, 2003. http://www.kc.frb.org/ publicat/sympos/2003/sym03prg.htm. Goodhart, Charles (2000) “Asset Prices and the Conduct of Monetary Policy,” working paper, London School of Economics. Greenlaw, David, Jan Hatzius, Anil K. Kashyap, and Hyun Song Shin, (2008), “Leveraged Losses: Lessons from the Mortgage Market Meltdown,” US Monetary Policy Forum Report No. 2 , Rosenberg Institute, Brandeis International Business School and Initiative on Global Markets, University of Chicago Graduate School of Business. Greenspan, Alan (2002) “Economic Volatility,” remarks at Federal Reserve Bank of Kansas City symposium, Jackson Hole, 2002. http://www.federalreserve.gov/ boarddocs/speeches/2002/20020830/default.htm. Gromb, Denis, and Dimitri Vayanos (2002) “Equilibrium and Welfare in Markets with Financially Constrained Arbitrageurs,” Journal of Financial Economics 66, pp. 361-407. Hattori, Masazumi, and Hyun Song Shin (2008) “Yen Carry Trade and the Subprime Crisis,” working paper, Bank of Japan and Princeton University. He, Zhiguo, and Arvind Krishnamurthy (2007) “Intermediary Asset Pricing,” working paper, Northwestern University. Holmström, Bengt, and Jean Tirole (1998) “Private and Public Supply of Liquidity,” Journal of Political Economy 106, pp. 1-40. International Monetary Fund (2008a) Global Financial Stability Report, April, Washington, DC. International Monetary Fund (2008b) World Economic Outlook, forthcoming draft of chapter 4, October 2008 issue, Washington, DC. Kashyap, Anil, and Jeremy Stein (1994) “Monetary Policy and Bank Lending,” in N. Gregory Mankiw (ed.) Monetary Policy, University of Chicago Press. Kashyap, Anil, Raghuram Rajan and Jeremy Stein (2008) “Rethinking Capital Regulation” paper prepared for the Federal Reserve Bank of Kansas City Symposium at Jackson Hole, 2008.

08 Book.indb 332

2/13/09 3:58:45 PM

Financial Intermediaries, Financial Stability and Monetary Policy

333

Kiyotaki, Nobuhiro, and John Moore (1997) “Credit Cycles,” Journal of Political Economy 105, pp. 211-248. Kiyotaki, Nobuhiro, and John Moore (2005) “Liquidity and Asset Prices,” International Economic Review 46, pp. 317-349. Kohn, Donald (2005) “Commentary: Has Financial Development Made the World Riskier?” Proceedings of the Federal Reserve Bank of Kansas City Symposium at Jackson Hole, 2005. http://www.kc.frb.org/publicat/sympos/2005/sym05prg.htm. Kyle, Albert S., and Wei Xiong (2001) “Contagion as a Wealth Effect,” Journal of Finance 56, pp. 1401-1440. Lown, Cara, and Don Morgan (2006) “The Credit Cycle and the Business Cycle: New Findings Using the Loan Officer Opinion Survey,” Journal of Money, Credit, and Banking 38, pp. 1575-1597. Mishkin, Frederic (2008) “How Should We Respond to Asset Price Bubbles?” Speech at the Wharton Financial Institutions Center and Oliver Wyman Institute’s Annual Financial Risk Roundtable, Philadelphia, May 15, 2008. http://www.federalreserve.gov/newsevents/speech/mishkin20080515a.htm. Morris, Stephen, and Hyun Song Shin (2004) “Liquidity Black Holes” Review of Finance, 8, 1-18. Morris, Stephen, and Hyun Song Shin (2008) “Financial Regulation in a System Context,” working paper, Princeton University. Plantin, Guillaume, Haresh Sapra and Hyun Song Shin (2008a) “Marking to Market: Panacea or Pandora’s Box?” Journal of Accounting Research 46, 435-460. Plantin, Guillaume, Haresh Sapra and Hyun Song Shin (2008b) “Fair Value Accounting and Financial Stability,” forthcoming, Financial Stability Review, Banque de France. Poole, William (2005) “The Fed’s Monetary Policy Rule,” remarks at the Cato Institute, October 14, 2005. http://www.stls.frb.org/news/speeches/2005/10_14_05.htm. Rajan, Raghuram (2005) “Has Financial Development Made the World Riskier?” Proceedings of the Federal Reserve Bank of Kansas City Symposium at Jackson Hole, 2005. http://www.kc.frb.org/publicat/sympos/2005/sym05prg.htm. Shin, Hyun Song (2005) “Commentary: Has Financial Development Made the World Riskier?” Proceedings of the Federal Reserve Bank of Kansas City Symposium at Jackson Hole, 2005. http://www.kc.frb.org/publicat/sympos/2005/sym05prg.htm. Shin, Hyun Song (2008) “Reflections on Modern Bank Runs: A Case Study of Northern Rock,” working paper, Princeton University.

08 Book.indb 333

2/13/09 3:58:45 PM

334

Tobias Adrian and Hyun Song Shin

Sims, Christopher (1980) “Macroeconomics and Reality,” Econometrica 48, 1-48 Stark, Jürgen (2008) “Main Challenges for Monetary Policy in a Globalised World” speech at the conference “Monetary Policy in Sub Saharan Africa: Practice and Promise,” Cape Town, 28 March 2008. http://www.bis.org/review/r080331e.pdf. Svensson, Lars (2004) “Challenges for Monetary Policy,” paper for the Bellagio Group Meeting at the National Bank of Belgium, January 2004. www.princeton. edu/~svensson. Taylor, John (1993) “Discretion Versus Policy Rules in Practice,” Carnegie-Rochester Series on Public Policy 39, pp. 195-214. Woodford, Michael (2003) Interest and Prices: Foundations of a Theory of Monetary Policy, Princeton University Press. Woodford, Michael (2005) “Central Bank Communication and Policy Effectiveness,” Proceedings of the Federal Reserve Bank of Kansas City Economic Symposium at Jackson Hole 2005. http://www.kc.frb.org/publicat/sympos/2005/ sym05prg.htm.

08 Book.indb 334

2/13/09 3:58:45 PM

Commentary: Financial Intermediaries, Financial Stability and Monetary Policy John Lipsky

I am very pleased to have the opportunity to discuss this timely and thought-provoking paper. Its implications are particularly relevant, more than a year after the onset of the current financial turmoil. In particular, views remain divided on how economic activity is being affected, and how we can best ensure that lessons learned from this period help to strengthen our crisis prevention and management capabilities. In their paper, Adrian and Shin present a compelling argument why banks and capital markets represent two sides of the same coin in market-based financial systems (Chart 1). They provide a clear explanation for the procyclicality of leverage for market-based financial intermediaries and link the expansion and contraction of balance sheets of these intermediaries to several factors, including, most importantly, the Federal Reserve’s short-term policy interest rate. They then draw links between the behavior of intermediaries’ balance sheets and macroeconomic outcomes. They argue that changes in short-term policy rates have tended to accentuate fluctuations in these balance sheets, with faster balance sheet growth being followed by lower interest rates, and slower balance sheet growth followed by higher interest rates, except during crises. In this sense, monetary policy may have played a procyclical role. Adrian and Shin argue that the potential for financial sector distress should 335

08 Book.indb 335

2/13/09 3:58:46 PM

336

John Lipsky

Chart 1 Main themes of the Adrian and Shin paper •

Balance sheets of market-based financial intermediaries (broker/dealers) are important for monetary policy.

•

Short-term interest rates determine the cost of leverage and the size of their balance sheets, especially because their leverage is highly procyclical.

•

Monetary policy assumes an additional direct financial stability role as it can prevent disorderly unwinding of leverage, which has implications for the real economy.

be taken into account explicitly in a forward-looking manner in conducting monetary policy. This would imply leaning against the wind when leverage is rising rapidly during boom periods. One further conclusion they reach is that forward-looking communications by central banks can be counterproductive by excessively compressing uncertainty around the path of future interest rates. This could lead to excessive leverage building up in the financial system, which might then be reversed in a disorderly manner with harmful consequences to economic activity. While Adrian and Shin focus on the United States, the same can be examined in an international context. After all, this episode of financial stress is truly global in nature, affecting both advanced and emerging market economies (Chart 2). Based on a cross-country “Financial Stress Index” we have developed, the current episode appears large by past U.S. standards—as seen in the upper panel of this chart—but in addition, it has affected a larger share of advanced countries than any other episode since 1980, as shown in the lower panel. Given the global dimension of this crisis, and the importance of understanding the links between stress on financial systems and macroeconomic performance, we are looking at this issue in some detail. In fact, the lead analytical chapter of the October 2008 World Economic Outlook examines the interaction between financial stress and economic performance in a cross-country context. Our upcoming Global Financial Stability Report also examines how disruptions in short-term money markets are affecting monetary transmission mechanisms.

08 Book.indb 336

2/13/09 3:58:46 PM

Commentary

337

Chart 2 The current episode of financial stress ranks as one of the most intense and most widespread United States: Financial Stress Index (shaded areas denote financial stress episodes) 100

100

95

95

90

90

85

85

80

80

75

75

70 1980

70 85

90

95

2000

05

Share of Countries Experiencing Financial Stress 100

100 90

90

Stock market crash

Scandinavian banking crises

80

80

LTCM

70

70

60

60 50

Nikkei/Junk bond collapse

50

ERM crisis

40

40

30

30

20

20

10

10

0 1980

85

90

95

2000

05

0

Adrian and Shin’s analysis complements our own work along several dimensions and is consistent with some of our own findings, although I will also try to highlight some key differences as well. Adrian and Shin underscore the connection between financial intermediaries and markets by focusing on the “mark to market” practices

08 Book.indb 337

2/13/09 3:58:46 PM

338

John Lipsky

of investment houses and the procyclicality of leverage for brokerdealers in the United States. One potential criticism of this approach is that it may overstate the role of investment banks. I believe Adrian and Shin are right to emphasize the general point about the implications of a shift toward more market-based or arm’s length intermediation, while also emphasizing that the distinction between broker-dealers and commercial banks has become blurred. As a result, the financial system as a whole, both in the United States and to a somewhat lesser extent in other advanced economies, has moved toward greater reliance on arm’s length intermediation and fee-generating activities and away from relationship-based lending (Chart 3). Thus, we find that the procyclicality of leverage that arm’s length financial systems generate is a broader phenomenon across financial systems (Chart 4). We see this trend becoming increasingly prominent in a broad range of countries over time. One could argue, as the authors themselves acknowledge, that the apparent economic impact of broker-dealer variables on macroeconomic performance may reflect the priors that underlie the specification, namely that the balance sheets of financial intermediaries they examine should matter for real economic growth. In this regard, controlling for endogeneity of the variables and accounting for missing variables—such as the balance sheets of other institutions—remain areas for further exploration. Moreover, the estimated size of the impact itself appears quite small, as reflected in the coefficients. A broader analysis of the entire financial system’s leverage may provide larger estimates. In fact, our own work supports the conclusion that financial stress has a substantial economic impact, even controlling for other factors that may affect the economy (Chart 5). As can be seen on the right panel, the cumulative output losses in both slowdowns and recessions are substantially higher when they are preceded by financial stress. Moreover, we find that even in financial systems that are more market-based, banks continue to play a very important role. Indeed, our analysis shows that banking-related financial distress on average

08 Book.indb 338

2/13/09 3:58:46 PM

Commentary

339

Chart 3 Financial systems are becoming more arm’s length (September 2006 World Economic Outlook) 0.8

0.8

1995 2004

0.7

0.7

Norway

Denmark

Spain

France

United States

0.0

United Kingdom

0.0

Australia

0.1

Canada

0.1

Netherlands

0.2

Sweden

0.2

Italy

0.3

Japan

0.3

Finland

0.4

Belgium

0.4

Portugal

0.5

Germany

0.5

Austria

0.6

Greece

0.6

Note: The arm’s length index is the average of three sub-indices: relevance of traditional (relationship-based) banking intermediation, development of new types of financial intermediation conducted largely at arm’s length, and the role played by financial markets, where each country’s score is calculated relative to the maximum value across all countries in 2004.

Chart 4 Arm’s length systems may be more procyclical Procyclical Leverage 0.4

Canada

Degree of procyclical leverage (from median regressions)

0.3

Belgium Austria

United Kingdom

0.4

Netherlands

0.3 United

Denmark

States

0.2

0.2

France Australia

Germany

Spain

0.1 Finland

Sweden

0.1

Norway

Italy

0

0

Japan

-0.1

0.2

0.4

0.6

0.8

-0.1

Arm's length index

08 Book.indb 339

2/13/09 3:58:47 PM

340

John Lipsky

Chart 5 Financial stress is associated with larger and deeper economic downturns Cumulative Output Loss During Slowdowns and Recessions

Number of Financial Stress Episodes

(medians; percent of GDP)

120

0 .0 - 0 .5

100

- 1 .0 - 1 .5

2000s 80

- 2 .0

1990s

- 2 .5

1980s 60

- 3 .0 - 3 .5 - 4 .0

40

- 4 .5 20

Preceded by financial stress

0 Total

Followed by slowdown

Followed by recession

Others

Not preceded by financial stress

Slowdowns

Not Preceded preceded by by financial financial stress stress

- 5 .0

Recessions

has higher output costs than other types of financial stress, such as those restricted to securities markets or foreign exchange markets. One implication of Adrian and Shin’s paper is that the procyclicality of leverage in market-based systems amplifies the economic consequences of financial turmoil. Indeed, from a cross-country perspective, we find supporting evidence that arm’s length financial systems appear to be more vulnerable to sharper downturns once a shock hits the system (Chart 6). An important issue raised but not addressed by Adrian and Shin’s paper is to what extent the economic impact of swings in the availability of credit depend on the initial conditions—in particular the balance sheets of households and firms that are the ultimate users of financial intermediation services. Moreover, other initial conditions—such as the build up of asset prices (including for housing) and credit—also could play a role in explaining the impact of a financial system shock on economic activity.

08 Book.indb 340

2/13/09 3:58:47 PM

Commentary

341

Chart 6 Deeper downturns in more arm’s length systems Output Loss under Financial Stress (GDP percent change from a year ago; medians) 5

5 Countries with below median arm's length financial systems

4

4

Countries with above median arm's length financial systems

3

3

Difference between medians statistically significant at time t = 1

2

2

1

1

0

0

-1

-3

-2

-1

0

1

2

3

-

-1

Quarters before and after the start of financial stress episode

Our research indicates that initial conditions do matter. Specifically, we find three key results regarding the importance of initial conditions: First, both house prices and credit-to-GDP ratios tend to rise significantly faster during the upswing of the financial cycle in those stress episodes that eventually are followed by downturns (Chart 7). Second, episodes of financial stress followed by downturns tend to be characterized by non-financial firms being more heavily dependent on external sources of funding in the run-up to the financial stress episode. A higher initial reliance on external funding makes firms more vulnerable to a downswing, setting the stage for a larger impact on the real economy as firms are forced to adjust their spending plans in the aftermath of financial stress. Third, periods of financial stress followed by recessions seem to be characterized by greater household sector reliance on external financing. The policy implications of Adrian and Shin’s paper are particularly interesting, and perhaps somewhat provocative. In particular, I would like to focus on three policy conclusions.

08 Book.indb 341

2/13/09 3:58:47 PM

342

John Lipsky

Chart 7 Initial conditions matter Household Net Lending/Borrowing (% of gross disposable income; deviation from trend one year prior to start of financial stress; averages)

NFC Net Lending/Borrowing (% of GDP; deviation from trend one year prior to start of financial stress; averages)

0.2

0.0

0.0

-0.1

-0.2 -0.4

-0.2

p-value = 0.04 p-value = 0.05

-0.6

-0.3 -0.4

-0.8

-0.5

-1.0

p-value = 0.02

-1.2

-0.6

-1.4

-0.7

-1.6

-0.8 Recessions

Slowdowns

Others

Real House Price (cumulative percent deviation from trend over six quarters prior to start of financial stress; averages)

Recessions

Slowdowns

Others

Credit (% of GDP; cumulative % deviation from trend over six quarters prior to start of financial stress; ave.) 8

30 25

6

20

p-value = 0.05

15 p-value = 0.00

10

4 2

5

0

0 -2

-5 -10

-4

p-value = 0.00

p-value = 0.00

-15

-6 Recessions

08 Book.indb 342

Slowdowns

Others

Recessions

Slowdowns

Others

2/13/09 3:58:48 PM

Commentary

343

First is that the potential for financial distress should be explicitly taken into account in a forward-looking manner when conducting monetary policy. They argue that the most powerful tool in relaxing aggregate financing constraints is lowering the target policy interest rate. Since the financial system as a whole holds long-term, illiquid assets financed by short-term liabilities, lowering the short-term rate can raise the profitability of intermediation, and thus mitigate a sharp pullback in leverage that would have adverse macroeconomic consequences. Adrian and Shin support this claim by citing the timely reduction in target rates during the financial stress episode of 1998. Furthermore, their conceptual framework suggests that even if some financial institutions can adjust their balance sheets flexibly downward, there will be some who cannot—their so-called “pinch points” of the financial system. Therefore, acute financial stress, particularly that resulting from a sharp retraction in leverage, will show up somewhere in the system. In this regard, monetary policy and financial stability policies are linked. The work in our upcoming Global Financial Stability Report corroborates this conclusion. With a more arm’s length financial system, instability in one market may be more easily transmitted to others, making the job of monetary policymakers more difficult. Chart 8 shows that changes in the spreads between the fed funds target rate and other rates become far less predictable during crises. Focusing in on the last two years, the elevated variance in the spread of fed funds rates to rates offered by near-banks—arguably the most leveraged institutions—is particularly dramatic, though so far the rates for endusers—households and corporates—have been less affected. The more bank-based financial system in Europe has also seen less volatility in the relation between the ECB’s Euribor rate and other borrower rates. I agree that monetary policy’s possible role in contributing to booms and busts in leverage is an important issue. Indeed, there may be merit in explicitly leaning against the wind if leverage is rising unusually rapidly. However, the historical evidence suggests that there is a large structural component to rising leverage, as suggested by Chart 3 of their paper. As a practical matter, separating structural and cyclical trends would seem to be quite difficult. Therefore, I consider monetary policy to be a relatively imprecise instrument in this

08 Book.indb 343

2/13/09 3:58:48 PM

344

John Lipsky

Chart 8 Forecast accuracy of spread changes has deteriorated in the subprime crisis 50%

Starting Date of Subprime Crisis under-estimation

40%

Deviation from Actual Value

30%

3-Month LIBOR ABS yield

20%

Mortgage Agency Bond Yield

10% 0% -10% -20% -30%

over-estimation

-40% -50%

Jun-06

Sep-06

Dec-06

Mar-07

Jun-07

Sep-07

Dec-07

Mar-08

Jun-08

respect and see greater merit in the use of macroprudential policies to reduce procyclical tendencies. In this regard, I agree with Adrian and Shin that a closer look at the role of risk-based capital requirements may be in order, and this should be high on policymakers’ agenda. Adrian and Shin stress two other policy implications. First, they make a case for the informational content of investment bank balance sheet variables for the conduct of monetary policy. While the money stock is a measure of the liabilities of deposit-taking banks and may have been useful before the advent of the market-based financial system, this quantity will be of less relevance in financial systems such as that of the United States. Instead, they argue that measures of collateralized borrowing would be more useful, such as the weekly series on repos of primary dealers. My own take would be that central banks need to (and in many cases already do) monitor the broader trends in liquidity and leverage across the whole financial system. Finally, Adrian and Shin argue that if central bank communication compresses the uncertainty around the path of future shortrates, this would reduce the risk associated with taking on longerterm assets financed by short-term debt. If such a compression raises the

08 Book.indb 344

2/13/09 3:58:48 PM

Commentary

345

potential for disorderly unwinding later in the expansion phase of the cycle, then this may not be desirable from the point of view of stabilizing real activity. Therefore, forward-looking guidance may compress volatility excessively and be counterproductive. I would argue that in general, central bank transparency has delivered great benefits across a wide range of countries, and we should be careful in interpreting the findings of this paper as suggesting that central bank communication should be more ambiguous than at present. While excessive certainty about the future path of interest rates may contribute to greater-than-optimal buildup in leverage and the risk of a disorderly unwinding later, I think the real problem is one of incentives in the financial system. As a result, measures to prevent this buildup in vulnerabilities perhaps should be designed more around regulatory policies aimed at aligning the incentives of market participants with sound approaches to risk management. In sum, this paper makes an important contribution to the debate on the link between the financial sector and economic activity, as well as on the appropriate role of monetary and regulatory policies. A thorough debate is warranted on the changing linkages between markets, policies and economic outcomes as financial systems worldwide continue to evolve into more arm’s length–based systems.

08 Book.indb 345

2/13/09 3:58:48 PM

General Discussion: Financial Intermediaries, Financial Stability and Monetary Policy Chair: Martin Feldstein

Mr. Kashyap: This was a great paper. I wish I had written it. I want to check two things that are implications, just to make sure I am following along. Most central bankers in this room have discovered the workhorse models used by their research staffs haven’t been very helpful in thinking about this crisis. The standard new Keynesian model that had become a workhorse and was convenient for talking about inflation targeting has no financial system. It seems to me the first way many central banks then would try to proceed is to follow the Bernanke-Gertler-Gilchrist way of thinking about modeling this. One of the implications of this paper is that it’s not necessarily a good way to go. Thinking about the next imperfection to add should not be on the credit demand side and the contracting problem with entrepreneurs, but it really should be on the credit supply side. That is an imperative task to begin work on. So then I have a question if that is right. Outside of the United States, lots of the central banks aren’t going to have big broker-dealers in their economies. I am wondering if you have thoughts as to how they should go about trying to measure the credit supply conditions to add into their models, at least outside of the couple of very developed markets where broker-dealers operate. 347

08 Book.indb 347

2/13/09 3:58:48 PM

348

Chair: Martin Feldstein

Mr. Stein: I am also enormously sympathetic to the diagnosis and want to ask a question about the prescription that follows from that diagnosis. At the very end of your talk, you talked about monetary policy essentially taking up the slack. That is to say, when the financial sector is getting overleveraged, you might want to raise the funds rate to lean against it. It seems like an alternative takeaway from your evidence and your arguments—and I think I am echoing what John Lipsky said—you are suggesting we need two instruments to solve two problems. It seems like a very natural takeaway would be that what you want to do is have active control of capital requirements. This is how I heard what Chairman Bernanke was saying this morning when he said, “We want to have a more macroprudential approach to capital regulation.” There is a very direct way to do it in your model, which would be raise capital requirements to the point where leverage is no longer procyclical. Wouldn’t that be a more direct way of addressing the problem than trying to get at it through the federal funds rate, which seems like an awkward tool for doing so? Mr. Berner: Like the other questioners, I am pretty supportive of the thesis of your paper that leverage is procyclical in both directions and something that monetary policy needs to think about, both in the traditional sense and in the macroprudential sense. The questions are: Is there in your view an asymmetry in that indirection? In other words, is deleveraging far more procyclical to the downside than is the leveraging up process as euphoric as market participants may think it can be? Maybe that is one of the reasons they complain more about it. Perhaps that is because liquidity dries up more quickly in a bust than is created in a boom. Perhaps it is because lenders are always short the option and those options behave asymmetrically in a bust than the way they do in a boom. What are the policy implications of that? Does it support the idea, as Rick Mishkin for example has said, that you need to act very aggressively in a downturn with monetary policy—more aggressively than you would in leaning against the wind to the upside—or not?

08 Book.indb 348

2/13/09 3:58:48 PM

General Discussion

349

The second question that arises from that is obviously we are going to hear from the folks behind me—Jeremy and Anil—about countercyclical capital regulations. But you also talked about the implied leverage in haircuts and maybe implicitly in trading margins. Should we think in a countercyclical way about those regulations as well from a policy standpoint? Mr. Feldstein: I have a related question to that. If you are thinking about using countercyclical capital requirements, how do you extend that beyond the commercial banks? Interest rate policy affects all financial institutions, but an entirely new set of capital requirements would be required to go beyond the banks. Mr. Redrado: Although the paper has a strong U.S. focus, I think it is very relevant for emerging economies, in particular because it brings back the financial sector to the heart of monetary policies, as it has a very solid foundation in explaining the transmission mechanism of credit supply. You revisit monetary aggregates, which are particularly relevant for countries that have very shallow financial markets of a single-digit ratio of credit to GDP. I would like to focus on bringing the balance sheet dynamics of the financial sectors into monetary policy equations. In countries that have gone through periodic crises, you need to have built-in crisis-prevention mechanisms, or anticyclical policies, in order basically to provide a longer-term horizon. Although you probably haven’t thought about emerging markets in writing your paper, I think it has strong implications for inflationtargeting regimes in developing economies. Mr. Weber: I also think that we should, at this stage, not only look at the monetary policy response to the increased procyclicality of banking, but keep the regulatory side firmly in focus. We very much talked about capital requirements and leverage being procyclical. There is a nice passage in the paper by Gary Gorton beginning on pg. 179 where he highlights these issues for the CDO market. One thing that is striking about capital requirements which are rating-based is that they are intended to be preventive in the sense that if you invest in a noninvestment-grade asset, the capital requirements are so high that you think twice before buying such assets.

08 Book.indb 349

2/13/09 3:58:48 PM

350

Chair: Martin Feldstein

But that is a static view. We are finding out nowadays in the midst of a financial crisis, that the breaking of the overcollateralization triggers in case of events of default can lead to a rating migration of these structured products from AAA to noninvestment grade, just like that. Of course at that point in time, the capital requirements become punitive rather than being preventive. As a consequence, we need to look at better regulation in the sense of a more dynamic approach, which takes into account how regulation interacts with downgrades and capital requirements. Let me come to the issue of deleveraging. The deleveraging process has basically caused the need for many banks to seek new capital, largely due to the rating migrations, less due to an impairment of the underlying cash flows of the assets. This has severe accounting consequences, which we need to adress by allowing more choices for banks. They should be able to move assets from the banking to the trading book and vice versa. And they should be allowed to decide on different avenues on how to deal with distressed assets when they have to repair their balance sheet by taking on balance formerly offbalance-sheet investments. Mr. Trichet: Again, a fantastic paper and very stimulating! Let me only mention, as a matter of record, that as far as we are concerned as you know our concept of monetary policy is based on two pillars: an economic pillar and a monetary pillar. Underlying the latter pillar is a very deep analysis of the monetary situation, which includes both components and counterparts of monetary aggregates. What I derived from the paper that was presented is precisely the extreme importance of what in this framework corresponds to the counterparts of collateralized borrowing. This element seems to contain a lot of relevant information, including for the shorter run. As you know in our own understanding the two pillars are complementary over time: the economic analysis being mostly short to medium term and the monetary analysis being mostly medium to much longer term. The paper is extremely interesting, in that it underlines the importance of collateralized borrowing, which not only conveys information

08 Book.indb 350

2/13/09 3:58:48 PM

General Discussion

351

in the longer term but also in the shorter term. This is a very important new insight that the paper provides. Mr. Sinai: I, too, thought this was an extraordinarily important paper in its focus for monetary policy—at the moment particularly in the practice of U.S. monetary policy. In other papers and research, I’ve seen very little evidence of how relatively small banking-based activity is now in the financial intermediary channel of the United States. If you added the off-balance-sheet activities of the banks into the data, the banking-based activity would be even smaller. The capital market-centric balance sheet is very, very fundamental because in my view the boom-bust cycle we have had, and are having, is a product of balance sheet expansion and impacts on the economy via nonbank financial intermediaries, who structurally now are more important. And your data are showing this increased importance, as is the market-based channels to the economy compared to the more traditional channels that macroeconomists have thought about. I have a question. In your set of financial intermediaries that perform bank-like functions now and work through the capital markets, you didn’t mention private equity venture capital and hedge funds. There may be more. In your analytical framework, in principle, would you include those? The second question I have is, Are there any data on the other intermediaries, so that when we look at the picture of market-based versus bank-based, what would they show? Mr. Shin: The quick answer is a “no.” They are not leveraged mostly. We are only thinking about leveraged institutions. Mr. Summers: Two questions, stimulated by the interesting analysis: First, there seems to be a strong conclusion, if I understood it right, that lower interest rates promoted greater leverage and there was a procyclicality most of the time. This would—it seems to me—have broad implications for monetary policy, whether it wasn’t reasonable to distinguish between normal times when institutions were not capital-constrained and abnormal times when institutions were capital-constrained. If one

08 Book.indb 351

2/13/09 3:58:48 PM

352

Chair: Martin Feldstein

envisioned abnormal times when institutions were capital-constrained, a reasonable thing to assume about those moments would be that monetary policy would affect their profit margins, but would not at all affect the rate at which participants in the economy could borrow from them, which would be the point where the demand curve for loans met the supply that was set by their capital constraints. So, if one thought about the capital-constrained regime, one would think about depressed periods, like potentially the present, it seems to me in a quite different way than was suggested by your analysis. That would have a great deal of implications for the question of the efficacy of reducing interest rates or the negative aspects of raising short-term interest rates in an environment like the present one. It would suggest that, in a capital-constrained environment, the dominant effect was on intermediary profits, whereas in normal times the dominant effect was on economic activity. The second question was, implicitly you commented on most of the range of tools that people talk about. It wasn’t obvious to me what the implications of your analysis were for the implications of the Federal Home Loan Bank System and systems of its kind and for a heavily regulated GSE system. To what extent, as you conflate financial stability and monetary policy, do you think about that government provision of capital or liquidity or lending or however you think about that instrument? It seemed to me to be a question raised by your analysis, and I couldn’t quite see my way through to what the natural answer was from the paper. Mr. Landau: It is a question about how procyclicality really works. What is said in the paper and is very interesting is on page 299, “Equity seems to play the role of the forcing variable, and all the adjustment in leverage takes place through expansions and contractions of the balance sheet.” That doesn’t have to be the case. You could have imagined that those capital gains, which are made from marked to market during expansion, are distributed to shareholders, and the adjustment would take place through distribution. Why did it take place that way? I think that is a very relevant question for now because now

08 Book.indb 352

2/13/09 3:58:48 PM

General Discussion

353

capital is not being created, it is being destroyed inside the financial system, and obviously there are big difficulties in reconstituting it to its normal level. Could we have policies in terms of regulation, provisioning, or accounting which would make sure that this kind of asymmetry between capital helping expansion of balance sheet in ordinary times and accentuating the contraction in bad times could be mitigated? Mr. Meltzer: I like the thrust of this paper very much. I want to put it into a somewhat broader context because while I agree with the main thrust of the paper, the broader context is that there are two basic models that economists have used to discuss financial banking problems. One, which has very currently been popular, is the Woodford view that buries everything in expectations, despite Alan Blinder’s comment in one of his lecture series that this model fails completely because the term structure of interest rate equation just isn’t very good at explaining what happens to long-term rates. The second view, which I have always rejected, was that an essential loanable funds kind of view, which said that monetary policy works through the banking system by making loans available. A third view, which has never gotten much traction, is my own view that said asset prices are very important in monetary policy—all asset prices—because monetary policy works primarily by changing relative prices, that is, by making the price of new or future capital cheaper than the price of existing capital and therefore encouraging investment and similarly encouraging consumption. That is the broader framework in which I think the Shin and Adrian paper fits. And it would be very good if economics would return to the view, or recover the view, that it is really the relative prices of assets that matter most for the transmission of monetary policy. Mr. Blinder: I would like to join the chorus in saying this is a very, very important idea that all of us ought to be thinking about more than we have up to now. I have a question prompted by John’s Chart 4, which was the regression showing basically that in terms, speaking slightly loosely,

08 Book.indb 353

2/13/09 3:58:49 PM

354

Chair: Martin Feldstein

of how much is in broker-dealers versus how much is in banks, the United States is a huge outlier. Whereas, in terms of the procyclicality of those broker-dealers, so to speak, we are not a huge outlier. This is what raises the question. What makes this interesting in terms of relative importance is this Chart 3 of yours that shows the broker-dealers have been growing like mad and the banks have not. We’ve heard in several suggestions in papers in this conference and in Chairman Bernanke’s opening remarks that, in the future, the broker-dealer sector needs to operate first of all with less leverage and maybe with less complexity. If you take down the leverage and you take down the complexity and push these businesses into standardized exchange-traded products that Chairman Bernanke was speaking about earlier, you suck a lot of the profit out of the businesses. So the question is, Do you agree with that and might that mean this Chart 3 starts heading downhill? Mr. Shin: Thank you to John, especially, for the very good discussion. Let me pick on two points that came up. One was about transparency of monetary policy, and the other one was a broader one about whether we could take up slack with better regulation. Transparency is a very important issue and it is intimately tied with central bank accountability. There is a tradeoff here between leaning against the wind and more disclosures. But if we take a step back, we could think of transparency in a broader sense, not simply in terms of publishing your forecast policy rate, but more in terms of first, promoting a better understanding of central banks’ objectives; second, whether the objectives are the right ones; third, promoting a better understanding of the mechanism for the decision making in monetary policy; and lastly, why the means fulfill those ends. I think that will go a long way toward addressing those governance issues without necessarily going down the very narrow route of thinking of transparency just as about having to publish forecasts. If we can break that link between the narrow issue of publishing some stuff or other and the broader one of better governance issue, we may actually have a better debate.

08 Book.indb 354

2/13/09 3:58:49 PM

General Discussion

355

On regulation and how far regulation can go, Marty’s point earlier about why regulatory restrictions on quantities might be more difficult to enforce than doing it through prices is a very important one. As soon as you try to enforce the Basel rules by putting in an anticyclical capital requirement, you are going to see financial innovation being directed toward bypassing those rules. It will be difficult to plug all the gaps. In contrast, prices are very democratic and everyone is ruled by those same prices. That is why the Millennium Bridge example works. Prices affect everyone equally at the same time. There should be an onus on using more price-based methods. Just a brief note on better regulation. I was at a conference in May of last year when a risk manager said, “The value-added of a good risk-management system is that you can take more risks.” That is one of the eternal truths—that the constraint binds all the time. The constraint binds when things are bad, but paradoxically, the constraint binds when things are good as well. You are under constant pressure to meet that 20 percent return-on-equity target. It is not clear that hiring more risk managers and getting more mathematical models in there and making the behavior much more sensitive to outcomes are going to be consistent with the macroprudential approach to regulation. There is a general cautionary tale here about why better risk management from an individual perspective need not promote greater stability over all. Anil Kashyap made the point that there are traditional models out there that look at frictions on the demand for credit. If the borrower’s balance sheet is weakened, then of course there is going to be a reduction in the demand for credit. Most of the models that are add-ons to the DSGE models are models of that kind. But the friction here is very much on the supply of credit and liquidity conditions. If you are going to tinker with the DSGE models, doing it through a demand for credit is not the right way.

08 Book.indb 355

2/13/09 3:58:49 PM

356

Chair: Martin Feldstein

I am short of time and cannot address the other excellent comments and questions. Mr. Lipsky: I thought the comments and questions were very relevant and very good. Happily, they also pointed in directions in which we are working. My IMF colleagues and I consider that this line of research is very promising and important. As you could tell from my presentation, we have been working actively in extending the sort of analysis contained in this paper in an international context. I recommend to all of you and hope you will read the relevant chapter in the upcoming World Economic Outlook and the related work in the Global Financial Stability Report. Regarding ongoing work in this area, we are forming a new unit in the Research Department at the International Monetary Fund focused on macrofinancial linkages. We think this is potentially valuable contribution to our understanding of developments in financial systems and their important interaction with economic outcomes.

08 Book.indb 356

2/13/09 3:58:49 PM

The Current Crisis and Beyond Mario Draghi

More than a year into the most challenging financial crisis of our times, we now face a complex and interlocking combination of rising inflation, declining growth, tightening credit conditions and widespread liquidity tensions pervading the world financial services industry. Authorities are using a range of tools at their disposal to address these challenges. But this crisis has shown, together with the lack of private sector discipline and weaknesses in the regulatory framework, some novel interrelations that call for action on all these fronts concurrently, and this further complicates our tasks. And yet, as we develop responses to these challenges, we also need to step back and consider how we have arrived at where we are. At a very general level it is now becoming apparent that the transformation undergone by the financial services industry in the last few years has not been fully appreciated in its implications for monetary and regulatory policy making. As was well said by Adrian and Shin in the paper presented this morning, this has been the first postsecuritization crisis. Low interest rates and volatility have boosted the size of both regulated and unregulated financial institutions’ balance sheets and their leverage during a long period of time. But of course not all of them were able to withstand a sudden change in financial 357

08 Book.indb 357

2/13/09 3:58:49 PM

358

Mario Draghi

conditions amplified by the accumulated leverage. Market participants failed to soundly manage, measure and disclose risks, with ignorance, greed or hubris playing their customary roles. But that is where, with the benefit of hindsight, regulators should have stepped in, responding to the externalities imposed on the financial system by weak financial institutions, the agency problems that foster excessive risk-taking by financial firms and investors, and the collective action problems in such areas as investment in risk-management capacity and infrastructure, in market infrastructures, in the maintenance of market liquidity and, above all, transparency. Still more broadly, reviewing our experience of the past two or three decades, one is struck by the repeated tendency of the financial system to build up risk and leverage over a series of years, then turn and shed risk rapidly and indiscriminately. While the assets and agents involved and the triggering mechanisms differ from one cycle to the next, the cycles have tended to produce significant deadweight costs and distortions in the real economy, both during the upswing and during the subsequent retrenchment, and this is especially significant the more the financial system is leveraged. While we cannot and would not want to eliminate the bouts of optimism and pessimism that are part of human nature, we must address some of the pro-cyclical implications of our own policy making. The conviction of the Financial Stability Forum (FSF), called a year ago by the G7 to draft the first report in response to the crisis, is that, due in part to a perverse set of incentives, leverage had reached a level that was both excessive for the risk-management capacity of many institutions and misperceived in its real entity. Therefore, improving the incentives in the system so that risks are appropriately managed and risk-control frameworks keep pace with financial innovation; improving the resilience of the system to shocks, whatever their source; and introducing frameworks for dampening the cyclicality of risk-taking, have become the cornerstones of our work plan. Our aim should be to produce a system more immune to the perverse incentives that we have seen, where leverage is lower, and where the sources of leverage and their associated risks are better identified and addressed. Regulatory and supervisory changes will need to go

08 Book.indb 358

2/13/09 3:58:49 PM

Luncheon Address

359

hand-in-hand with enhanced transparency about risk exposures and valuations throughout the system. Let me start with actions to improve incentives. Under the umbrella of the FSF, and in a collaborative effort of national and international regulatory and supervisory authorities, we urged prompt implementation of Basel II, which will align capital requirements more closely to banks’ risks (and I will have more to say about this in a moment). We also recommended strengthening Basel II with respect to capital charges for credit risk in the trading book, for re-securitised assets, and liquidity lines for off-balance-sheet vehicles. We encouraged the Basel Committee to strengthen liquidity-management practices and buffers. We have called on financial institutions to align compensation better with long-term, firm-wide profitability, and recommended a range of measures to discipline the practices of credit rating agencies and to reduce regulatory reliance on ratings so that investors engage in proper due diligence rather than relying exclusively on ratings. We also set out several recommendations to enhance transparency and valuation practices. We outlined “leading practices” for disclosures by financial firms and urged financial institutions to use these as part of their financial reports starting in mid-year 2008. The International Accounting Standards Board, in response to a proposal by the FSF, has initiated a process to develop improved guidance on the valuation of illiquid instruments and related disclosures. We have also called on accounting standards-setters to overhaul standards for asset derecognition and consolidation of off-balance-sheet vehicles. And we recommended that securities market regulators address incentives problems in the securitisation chain and work to expand information for investors on securitisation products and their underlying assets. These measures should improve the ability of investors and others to track the risks taken by financial institutions and thereby increase the effectiveness of market discipline. The progress already achieved testifies to the importance we as financial authorities attach to a stronger framework of regulation. Not only has a consensus been built internationally and cross-sectorally on what is needed, but the process of policy development and implementation

08 Book.indb 359

2/13/09 3:58:49 PM

360

Mario Draghi

is moving forward in coordinated and consistent fashion —and faster than we ever experienced in the past. To illustrate: • Supervisors proposed in July new capital requirements for credit exposures in banks’ and securities firms’ trading books and will set out later this year adjustments to capital requirements for “resecuritisations” and short-term liquidity facilities. • In May, the Basel Committee issued revised guidance on liquidity risk management that materially raises standards for sound liquidity risk management and measurement—including requiring banks to maintain a robust cushion of unencumbered, high quality liquid assets as a safeguard against protracted periods of liquidity stress. • International Organization of Securities Commissions (IOSCO) and the Securities and Exchange Comission (SEC) has set out fundamental changes to its requirements on credit rating agencies to address the quality of ratings, as well as proposals concerning how ratings are used in regulatory guidelines. • Regarding transparency, over this summer, the larger banks have been using our recommended disclosures to provide expanded information on their risk exposures and valuations. • And the International Accounting Standards Board (IASB) is making good progress on new guidance and revised standards for valuations and off-balance-sheet entities, which we expect to see in the next few months. There are many other initiatives underway, alongside complementary efforts in the private sector. We welcome the recommendations, along with accountabilities, that have been set out by the Institute of International Finance (the IIF), the Counterparty Risk Management Policy Group, and the American and European Securitization Forums. The key challenge for us as authorities will be to remain engaged in seeing these and other reforms through, particularly given the shortterm challenges we face. Regulatory changes will need to be phased in over time to avoid adding to the adjustment challenges the system faces now. But there should be no stretching of timetables for

08 Book.indb 360

2/13/09 3:58:49 PM

Luncheon Address

361

enhancing disclosures, including of off-balance-sheet positions, as this is essential to repairing market confidence. Even if we do succeed in rectifying the incentive problems that have tended to generate excesses, our financial system will not be spared risk-management failures, shocks and disruptions in the future. This calls for strengthening the system’s resilience, in terms of both a better financial infrastructure and stronger shock absorbers in financial institutions. At the level of the infrastructure, a resilient system is one that is able to withstand the effects of the failure of a large financial institution. By reducing the centrality of any one institution to the system’s stability, stronger infrastructure should also contribute to reducing moral hazard. A critical priority in this area is to address weaknesses in the operational infrastructure of over-the-counter derivatives markets, and the work undertaken by the New York Fed to this extent should be commended by all the jurisdictions. It is also imperative that we strengthen national and cross-border crisis resolution frameworks so that we can allow weakened financial institutions, including large ones, to fail without putting the remainder of the system at risk. In addition to national reform efforts in a number of countries, work is underway in the FSF and the Basel Committee to strengthen crossborder cooperation and contingency planning among authorities in responding to crisis. At the level of individual institutions, improving resilience means ensuring that capital and liquidity buffers are large enough to enable firms to resist external shocks­—without mandating buffers at a level that impedes efficiency and encourages regulatory arbitrage. The issue is quite complex, because both the actual and the appropriate size of a buffer shift over time depending on the market and systemic environment. With regard to capital buffers, it is not yet clear how required, desired and actual capital levels will evolve over a full cycle under Basel II, but a framework is now in place for tracking and assessing this. Regulatory minima are one of several inputs into firms’ capital decisions. Banks have made significant efforts to raise new capital to meet

08 Book.indb 361

2/13/09 3:58:49 PM

362

Mario Draghi

anticipated needs relative to regulatory minima as well as to respond to the need to reassure markets. Banks clearly see a benefit to maintaining capital significantly above regulatory minima, and markets (including rating agencies) reward them for healthy capital levels. One reason, among others, that banks maintain this margin is underlying uncertainty over risk exposures, valuations and earnings prospects. When as at present this uncertainty increases, the margin required by the market also rises. To a degree, this is unavoidable. But to reduce the tendency that market reaction leads banks to raise capital (or reduce exposures) to possibly inefficient levels in a systemic crisis, we need a more robust ex ante framework of transparency about risk exposures, along with provisions, margins and reserves for valuation uncertainty, than we have at present. The appropriate size for liquidity buffers is even harder to mandate, or to predict, than for capital buffers, given the unpredictable nature of shocks to market and funding liquidity, and the understandable reluctance of monetary authorities to create perfect certainty about when and how they will provide emergency liquidity. But, as with capital, greater ex ante transparency about banks’ liquidity risk management frameworks, cushions and supporting arrangements should help reassure investors and counterparties and reduce the risk of sudden drains on liquidity, as well as the uncertainty that, over the past year, has led to wide and volatile liquidity premia in interbank funding markets. As we consider how to strengthen buffers, more thought needs to be given to how to promote higher buffers above the regulatory minimum in good times, and more flexibility for the system to make use of these buffers when they are comfortably above regulatory minima. Both the banks themselves and their counterparties should be more confident about the banks’ ability to draw on such buffers in bad times. Banks will only be willing to do this if they do not fear they will be punished by the market—which brings us back to the importance of transparency improvements. If banks can credibly assure markets and their regulators that risks to their asset values and earnings prospects are being soundly managed and contained, then they

08 Book.indb 362

2/13/09 3:58:49 PM

Luncheon Address

363

may be able to survive a temporary decline in capital levels, while still above their regulatory minimum, during cyclical downturns. Authorities will be closely monitoring banking conditions to inform decision on the best timing for introducing tighter capital requirements. We will not unduly burden financial institutions during the adjustment phase, but also not miss the opportunity of the next calm phase in financial markets to strengthen the structure of the financial system. Pacing regulatory changes, but firmly committing to them, will contribute to reducing the potential for cyclicality in the system to be deepened by regulatory capital requirements. There is no shortage of ideas around for how the capital regime might be modified to dampen potential pro-cyclicality. Some have suggested that a regime that adjusts regulatory minima with the cycle could help reinforce the message that buffers need to be built up in good times and are there to be used when a rainy day arrives. To a degree, this is already possible under the second pillar of Basel II. However, too much divergence in national implementation of pillar 2 would raise issues of transparency and consistency in international regulatory arrangements and should therefore be constrained. To be credible, a discretionary system would need to be limited in its scope, fully transparent, and broadly consistent across countries; ad hoc, uncoordinated reductions in required minima could be viewed as forbearance and could give the wrong signal about authorities’ judgement as to the overall strength of the system. It may also be possible to give more attention to the mix of capital instruments supporting banks’ risk exposures. There may be scope to enhance the quality of capital during strong economic conditions. A complementary policy aimed at strengthening buffers could look at mechanisms to automatically increase bank equity levels in bad times, such as reverse convertible debentures proposed by Flannery or the insurance scheme proposed by Kashyap, Rajan and Stein. Although many issues of security design would need to be addressed for these instruments to effectively discipline banks, generate adequate investor demand, and to be reliably collected on when bad states occur, these mechanisms could potentially reduce the bad-signal problems associated both with discretionary changes in minima and with capital-raising

08 Book.indb 363

2/13/09 3:58:49 PM

364

Mario Draghi

efforts by the banks themselves. A necessary complement would be a disclosure framework for risk exposures and valuations that prevents the market taking conversion events or the resort to insurance as a trigger for a broader downgrade of the system’s future prospects. Strengthened capital and liquidity buffers and improved incentives to assess and manage risk should all help reduce the pro-cyclicality of the financial system, and, in the new configuration of the financial services industry more than ever, help monetary policy to attain its traditional objectives. But should monetary policy itself embody in its objective function the health of the financial system? As Chairman Bernanke and others have noted, we cannot afford to ignore the health of financial markets when setting monetary policy. It would be difficult to disagree with this statement, and we should try to make it operational considering, however, that its application may have different degrees of intensity. We should certainly use the monetary and credit aggregates that give the best projections of financial conditions for a given structure of the financial services industry. But how to go beyond that is a matter of considerable complexity. Though I share the importance of the policy goal, it should be clear that how to define the objective function is not only analytically difficult, but it would also pose serious institutional challenges. For a central bank whose primary mandate is to preserve price stability alone, introducing financial stability as an additional objective could introduce a trade-off where none exists today. Indeed, at times of extraordinary volatility and dramatic risk re-pricing, maintaining price stability could be the best contribution that monetary policy could give to the return to financial stability. The same remains true during peaceful times, when prolonged periods of double-digit growth in several credit aggregates should call for action to protect future price stability even if in absence of an immediate danger. Second, I can only point at the difficulty of having a framework that is as clear and measurable as the one we have today for our monetary policies. But where I see the greatest difficulty is in the amount of information that would be needed in order to run a monetary policy having financial stability among its objectives. In other words, we would have to know about both the regulated and the unregulated parts of

08 Book.indb 364

2/13/09 3:58:49 PM

Luncheon Address

365

the financial services industry as much as we know about inflation and output. Of course this would also be the area where, were we to move in this direction, the benefits from enhanced transparency and interactions with regulatory policy would be greatest. In conclusion, the next few years are likely to be ones of low risktaking and progressively low leverage, as financial institutions and households repair their balance sheets and as internal and external macroeconomic imbalances are resolved. These adjustments will not be painless, and ensuring that they take place in an orderly manner will pose substantial challenges for policymakers: Preserve price stability while supporting growth, and continue injecting liquidity as needed by an industry that is still far from having resolved its problems, at a time of strong inflationary pressures and tightening credit conditions. In such a situation, monetary policy cannot be the only, or even the main tool for reflating the economy and the financial system. Fiscal policy will have a key role to play, in its various configurations. But equally important would be to recreate an environment where banks during this current transitional period were both able and willing to use their buffers to lend, and capital markets were able and willing to fulfill their functions. For this to happen, the uncertainties about risk exposures and asset valuations; about developments in the real economy; about the strength of financial, corporate and household balance sheets need to be adequately addressed. This implies a central role for improved transparency as well as action by the private sector to repair balance sheets and strengthen the functioning of markets.

08 Book.indb 365

2/13/09 3:58:49 PM

General Discussion: Luncheon Address

Mr. Bergsten: I would like to leap for a moment beyond all the daunting immediate problems and ask a longer-run question about the work you and the FSF are doing. If I heard Chairman Bernanke correctly this morning, his discussion of macroprudential regulation included the notion of thinking over time about a common or consistent regulatory regime across different classes of financial institutions and across countries. So, I have four closely related questions for you. Would that be a good idea over the longer run? Is it a feasible idea? If so, roughly over what time period could you imagine that happening? Does your FSF provide the nucleus of the institutional framework that might be needed to move in that direction? Mr. Draghi: Let me rephrase it in the following way. Would the FSF answer Chairman Bernanke’s point about the need to have a macroprudential framework that is both coordinated across institutions and across countries? And how long is it going to take? You’ve seen from what I have said today, that the line I’m proposing the FSF should take in the future is a line very much geared towards this macroprudential oversight concept. The FSF is a group of people that reflects different constituencies. It is basically a group where people who are interested in individual institution 367

08 Book.indb 367

2/13/09 3:58:49 PM

368

Luncheon Speaker, Mario Draghi

oversight—the supervisors—meet. Then you have the Treasuries, or taxpayers’ money. And then you have the central bankers. This architecture—which was designed during the post-Asian crises in the late 1990s—has proven to be very useful in addressing the present set of problems. It has proven to be quite agile and has evolved from being a place where, at the end of the 1990s, the greatest interest was in trying to understand how the Asian crisis had developed and what sort of mistakes should be avoided in the future. Gradually, most of the discussions became more concerned with regulatory issues. And now that we start seeing the effects of spillovers and the externalities of this crisis, it’s naturally turning into what Chairman Bernanke defined today as macroprudential oversight. The FSF is a group which has a well-defined constituency where the major financial centers are represented. Looking forward, we are actually planning two things. One is to see whether the membership should evolve in receiving the contributions of countries that are today still emerging countries but already big financial centers. The second thing is what sort of outreach efforts we should undertake, and there is a series of planned actions in this respect. Your point about the need to have coordinated action is very much at the center of the FSF identity in all sorts of initiatives. Imagine a different situation. A crisis like the one we are living through naturally elicits national responses, and to the extent that these responses affect the level playing field of our globalized financial industry, this would be highly disruptive. What’s happened instead is that all the regulatory discussions that have taken place since the crisis started have been fully shared and coordinated. There is another part of the authorities’ response to the crisis which we decided at the very beginning ought to remain national because there was little value added in having an international response. This is the lender-of-last-resort function, which involves the state stepping in to overcome a crisis. The experience has proved us right because we have had crises in various countries, and in each one, it took a different form from the other.

08 Book.indb 368

2/13/09 3:58:49 PM

General Discussion

369

And in each one, taxpayers’ money was involved so it would have been very difficult to have an internationally coordinated response in that specific field. Mr. Fischer: Before transparency became a virtue, banks, especially in Europe, used to have hidden reserves and do things which regulators now don’t like because they allow the banks to smooth earnings. I’m referring to the SEC regulators; the bank examiners are more inclined to see the value of large provisions. Presumably there is a social value to the banks adding to reserves and more generally to their financial resources in good times. This ends up as a tax issue, since banks are perfectly free to retain after-tax profits. Do you see any social value to giving some tax incentive in this area to reduce the procyclicality of bank operations? Mr. Draghi: The question is: Should we go back to a system where banks could have higher provisions than what is now currently determined by the current accounting rule IAS 39? The answer is yes. There is a widespread conviction that the previous system—the system before this accounting provision was put into place—had higher levels of capitalization, more flexible ways of handling credit cycles. As a matter of fact, we have a real example of a better provision, which is given by the Spanish system. Spanish banks nowadays have a capitalization level which is way higher than European average, and this is because before IAS 39 they had what we call “dynamic provisioning,” which basically amounts to what you said. The sense is that we have to do something about this, which of course is not an easy thing because of the reasons you said—because investors’ interest ought to be protected. So you don’t want to have situations, for instance, where the level of provisions is such that you are not actually taking care of the investors’ interests. It is not an easy question to answer. I think our current system is not as good as it was in the past. Mr. Feldstein: You’ve come back to the theme of transparency a number of times. And, it’s hard to object to transparency. On the other hand, I wonder how feasible it really is. I think about the series of announcements from major financial institutions that a loss has been discovered. Do they know it in advance? Could they really

08 Book.indb 369

2/13/09 3:58:49 PM

370

Luncheon Speaker, Mario Draghi

have disclosed it? Or, is it news even to the insiders themselves? Simply trying to describe the positions that they hold on their balance sheets, I think we’ve learned today the complexities of all of these instruments are a warning that you can’t simply describe what you hold and expect anybody from the outside or even anybody on the inside to understand. Mr. Draghi: I will rephrase this question as saying, Is this quest for higher transparency a realistic undertaking? Well, I can answer that: Yes, it is. It is the most important thing we can do now and the first. But the fact that firms keep on giving different figures about their losses may be due to what you suggested: opaqueness, ignorance, complexity. Or it may also be due to the fact the current economic environment is changing, therefore they simply upgrade their estimates or downgrade their estimates. But if we discuss the first possibility—as a matter of fact, especially the largest financial institutions have made a lot of progress in this regard. Much of the ignorance about exposures came from poor risk management, because the prerequisite for managing risk across the firm, in the different lines of business, was the knowledge of what the actual exposure was. So the evidence is that risk management had been very poor, therefore their knowledge was very poor. They are updating their risk management systems now, and I would say that much progress is being made by the major firms at least. And, in so doing, they are also upgrading their knowledge. I think this is not an effort that is going to finish anytime soon, but there is a certain convergence of interest between the regulated and the regulators in producing more transparency. However, one of the main reasons for not being transparent about certain products, especially complex ones, was that these products have a lot of property value embedded, and once you are too transparent about the complexity of these products you can really lose money. I do not know exactly what is going to happen with respect to this dimension of transparency. For the time being, it is not the dominant one because much of the value of such complex products has gone anyway. So there is very little left to protect in this specific area. But

08 Book.indb 370

2/13/09 3:58:49 PM

General Discussion

371

as I said, it is a battle that ought to be fought because it is a prerequisite for everything we have said today—for introducing these new ideas about how to improve flexibility in the use of capital, even for having monetary policy run with an eye on financial stability. Mr. Crockett: Mario, we’ve talked a lot this morning about procyclicality. Of course, you covered it in your talk. My question is: How far does that procyclicality arise from regulatory or accounting conventions, which tends to be an assumption that people make? And, how far is it inherent essentially in the way in which the financial system works? If I think of my own institution in reaction to the events of the past year, we have raised our tier-one capital ratios substantially from what it was a year ago. That is partially because internally we feel that the uncertainties are such we need the security of a stronger balance sheet, and it’s partly because the market rewards us for that. That is not because of accounting conventions, at least not directly to accounting conventions, or regulatory intervention. And, I suspect many other institutions of similar size and function as our own would feel the same. So, in addition to the question of how far is it due to the inherent nature of the financial system, to the extent it is, how can regulation or supervisory oversight deal with that? Mr. Draghi: It is clear that some of our regulations do induce procyclicality, and I have said something about this before, whether it is because of the accounting system or marking to market. Here, I think we have to take something as a fait accompli. We are not going to change the mark-to-market system of valuation for the thousands of reasons this was found to be good to begin with. And if we start fighting this battle, the battle will be over even before we start. However, there are specific narrowly defined instances where there is a case for looking at the mark-to-market accounting. I am referring now explicitly to something that might have been hinted at in the discussion this morning: when trading is disrupted and you have to price by proxies, namely an index. An index during these periods when trades are disrupted no longer reflects the underlying value of the assets, but it does reflect things concerning risk or liquidity or other factors. Certainly you have a case for thinking seriously about that, and the IASB, having followed our recommendation, has an

08 Book.indb 371

2/13/09 3:58:49 PM

372

Luncheon Speaker, Mario Draghi

expert valuation group where all the interests of the industry, regulators, and central bankers are represented. They’re going to produce guidance on this specific issue. So, in my view, procyclicality is not necessarily induced by accounting. Accounting’—marking to market —tells you how much assets are worth at that precise point in time if markets keep on functioning. If markets don’t function, then we are in that precise circumstance. By the way, accountants tell me—I don’t know whether they say this now, but they said it in the past —that people were not compelled to price using indexes. Not even before. They should price using their best judgment. Now I don’t know whether this guidance comes a year too late or if it actually was there before. Mr. Summers: This is somewhere between a question and a comment. I have to say when I hear the sentence there is a convergence of interests between the regulators and the regulated, I react the same way when I used to react when somebody said, “I understand the general laws of economics, but in my country it works different.” There is a prospect it could be true. But you can attach a negative presumption to almost anything that follows. There is certain confusion about when you aspire to transparency what it is that we are aspiring to. Are we aspiring to transparency about the current opinion of management who have bought a set of financial assets—who, it is very clear from what has happened over the previous year, have not understood those assets very well. Is our objective to achieve transparency in respect to what they think? Or is our objective to achieve transparency with respect to some underlying reality that seems to be set or in place? I understand the sort of hedge fund world point of view, which is everything should be marked to market as close as possible where there is a good market approximation for something that is a loan book that should be used and that should be done in response to transparency. I understand that view. I understand the view of the people who think we need to interfere substantially with the normal operation of mark to market accounting because it substantially induces procyclicality. I recognize that means you are going to have less transparency, but transparency about an illusory reality doesn’t have any meaning and simply operates to cause panics. I understand that point completely. The point of view that frankly

08 Book.indb 372

2/13/09 3:58:49 PM

General Discussion

373

tends to come out of the official sector when it goes into conclave with the private sector—which is that we should have more transparency and at the same time be smoothing the excesses of the abuses of mark to market—strikes me as understandably in political terms that gets everyone out of the room in agreement with each other. What I actually have a lot of trouble understanding is the coherent doctrine. Can you help me? Mr. Draghi: I can help you, but I have to say I was deliberately very, very circumspect about the coverage of this sentence. Let me tell you, we are in a sense at a stage which is far from the sort of dilemmas you’ve painted. Let me give you an example of how greater transparency can be useful and shouldn’t face and has not, in fact, faced great objections by the bankers. One of the recurring questions here is comparing European and U.S. banks. What you get commonly in discussions here on this side of the Atlantic is: “We are sure that European banks haven’t made enough provisions.” And certainly as somebody said to me, there isn’t in the United States one stone that has not been turned by now. While in Europe I’m sure nobody understands that claim. The fact that these 25 major financial institutions have accepted to respond to our quest for greater transparency, and are using the template we have given them—which is quite complete, or at least it is quite extensive—will allow us by September to know exactly the comparative state of provisioning between the two sets of banks. This is something that some banks complained about, saying that it was very expensive, that they didn’t have the management capacity to fill out this form, but basically in the end, they did it. In fact, one British bank, which has already produced its response, did a very good job. That’s the sort of transparency that doesn’t necessarily elicit hostility or suspicion between the regulated and the regulator. Other than that, I would agree with you. We have to start with the general presumption that the regulated entities follow a different set of interests from our own interests. There I would agree with you. I would not agree with you when you said that the two statements—agreeing with mark to market and pricing based upon indices when you have market disruption—are inconsistent. I don’t think they are. I think

08 Book.indb 373

2/13/09 3:58:49 PM

374

Luncheon Speaker, Mario Draghi

that when markets don’t function, mark to market becomes something different. I think the IASB’s interest in pursuing this line of thinking is actually quite legitimate. Having said that, don’t expect a big revolution from the IASB or big changes. We’ll have a millimetric response, a millimetric change there. Mr. Summers: Do you accept what would be the hard-core economists and finance people’s view that no financial institution should be allowed to use a pricing model which has the implication that there are obvious and easy profit opportunities by trading in freely liquid securities—that is, when, as is common in these discussions, you say, “Well, the index moves don’t have anything to do with the value of our stuff because it’s a crisis and it’s obviously distorted and doesn’t have anything to do with what stuff is worth.” A corollary of that is there is a very easy way to make money by buying the index. Most of the markets, people tend to resist arguments that have as their premise that there are a lot of thousand-dollar bills lying around on the street, and so do you accept that there is a kind of general doctrine for thinking about transparency, that yes, in some circumstances, you may need to use models and not price right along with markets? But, that there should be a systematic effort to purge the system of accounting conventions that have implicit in them the idea that there are easy profit opportunities like trading in freely liquid securities? Mr. Draghi: If I could rephrase your question, Larry, I would be against any use of price modeling which would be behaving in an asymmetric fashion. In other words, if I have to change—if I have to move from mark to market—this move ought to be symmetric. Both on the downside and on the upside. And I would be against any change that would not be symmetric. Mr. Summers: I understand that and I agree with you that symmetric assumptions that you can make profits on instruments that you misvalue in previous years is better than asymmetric assumptions of that kind. But, in your symmetric world you are assuming that, well, we know it’s overvalued, so it can be easier to make money shorting it sometimes, and other times we know it’s undervalued, so it’s easier to make money being long in it. Do we want to allow accountants and managers to sort of make assumptions that imply

08 Book.indb 374

2/13/09 3:58:50 PM

General Discussion

375

that there are easy profit opportunities? And isn’t that the logic of a bunch of what you and others are advocating? Mr. Draghi: I never looked at it that that way, quite frankly. If it is symmetric, then it means I have lost money on the way up and made money on the way down. What I am saying is that if I have a model, I have to have a consistent application of this model throughout the cycle. Mr. Calomiris: I wanted to talk little bit about whether transparency is meaningful, especially during the good times; whether transparency is meaningful if banks aren’t on a margin that rewards something. If you go back to the 1920s—they called it the Roaring ’20s, right?—loan growth by U.S. banks was enormous, and equity ratios were hugely procyclical both because of retained earnings, stock offerings and price increases. What was different? As I think Andrew pointed out too, the regulatory environment wasn’t providing so much protection for banks that they didn’t have to worry about procyclical leverage. And that’s the reason that the shadow financial regulatory committees of the United States, Europe, Japan, Latin America have all said that the sine qua non of reform of capital standards is to require banks to be on a debt market margin—a risky debt market margin. It is not easy to do, but it can be done. That’s what real market discipline means: not just disclosing things to the market when there’s no consequence. Equity markets are not enough of a discipline or reward. And you don’t have to legislate transparency once you create an incentive for transparency. Banks historically found ways to communicate with markets without regulators telling them how to because they were so scared about the consequences of not communicating with the market. The big problem is banks aren’t scared enough. They weren’t scared enough in the boom period. Now everyone is adjusting. Now everyone is trying to get the market reward that comes from stabilizing. The problem is that happened only in the bad times. We need it to happen in good times. You’re not going to get that under the current regulatory structure because the banks have lobbied to avoid discipline, as you pointed out with respect to the ratings problem. The banks lobbied to prevent this. And this gets us back to Larry’s point—that we really have a political economy problem, and discipline doesn’t just mean transparency.

08 Book.indb 375

2/13/09 3:58:50 PM

376

Luncheon Speaker, Mario Draghi

Mr. Draghi: No it certainly doesn’t. Much of what I said is meant to introduce discipline at good times, and so whether it will make it or not is going to be a matter of politics. Right now you have a uniform view by all the regulators across the world that once you introduce flexibility in your capital buffers that should be disciplined in such a way that banks are, I would say, compelled to pile up buffers in good times so that they can depend on them in rainy days. I think one country has succeeded in doing it. And I think it could be done. That is my perception, and of course it is a matter of perception here. It is not certainty. So far, it is a very weak statement because so far everybody is so scared, as you said, they would say yes to anything. Mr. Weber: Well, last year I was quite concerned with the sort of industry leveraging process and whether we would coordinate quickly enough or provide liquidity quickly enough. I think what you said is right, if we cooperate very well with our treasuries and supervisors. This year, I am more concerned. Do we have an exit strategy? Can we make sure that the current redefinition of the regulatory framework doesn’t bind us as central banks in pursuing the objectives that we should pursue? I am quite concerned that in this getting used to coordinating with all parties involved leads to a state of poor perception that this should continue into the indefinite future, and for me it is also very important what you said about the financial stability not being a concern for our primary objectives, however they are defined. Really my question is, Do you guys also discuss the exit strategy? How are we going to get out of this when the market is somewhat calmer? I think that this always needs to be kept in mind or we are lost as central banks. Mr. Draghi: Again, let me rephrase your point. You are worried about too much coordination on one front, and with the function of lender of last resort. There it is clear for all central banks concerned that coordination has proved to be extremely useful. It has really played a fundamental role in the time of crisis. But none of us, well, I am speaking for people who are here, would in a sense deflect from their own mandatory, statutory objectives. So coordination is useful in the sense that it is helpful to pursue objectives that are either present in your statutory mandate or have to do with specific crisis

08 Book.indb 376

2/13/09 3:58:50 PM

General Discussion

377

situations at a certain point in time. A totally different area of coordination where I don’t think we should have an exit strategy is the one that pertains to both individual institution oversight, or to what Chairman Bernanke called macroprudential oversight. There, I think there is clearly value added in having coordination and in remaining coordinated. As I said before, the main value added is in preserving the level playing field of our financial services industry. Mr. Trichet: From the very beginning of this interaction between the various countries concerned was the fact that after we worked out some kind of consensus amongst the “specialists,” we would be able to get it done in the various countries. The Basel committee started after the crisis of sovereign debt, and then we worked it out on the basis that we would be able to get approval through normal mechanisms in the various countries concerned. We discovered it might be more complicated than that. My main question is the following. Do you believe that at the present juncture, with the pressure of the credit crisis ongoing, we can consider that this working assumption remains reasonable, and that whatever consensus we get we can move through the process and directives in Europe and the Congressional procedures in the U.S.? Are you reasonably confident that again these macroprudential recommendations that Ben mentioned will crystallize into the legal framework that is needed in most cases in our various countries? Mr. Draghi: Well, another question about the realism of our recommendations, of our prescriptions, of our dreams. Let me distinguish between two issues. One is that many of the recommendations of the FSF in the report that we produced in April can actually be implemented without legislative steps. Some not, some do need that. I do not know about this country. But for the countries I can recall now, I don’t see big problems in Europe with the EU commission. Actually, we have been in close touch with them since they are the ultimate legislative power in the union on financial matters. So they had better be kept on board of the work we do. So far they have been collaborative. In a sense, they should only have their enthusiasm restrained from time to time. A very different issue is the one you raised about having a sort of an internationally agreed framework for

08 Book.indb 377

2/13/09 3:58:50 PM

378

Luncheon Speaker, Mario Draghi

macroprudential oversight. My view on this point is that I wouldn’t know. I wouldn’t know the answer. The concept is already complex by itself. I think you said so this morning, it should be defined without thinking about international coordination. But it certainly is worth fighting to have its principles—that are, I wouldn’t say coordinated, but basically, shared, agreed. Whether this is going to be possible really depends on what the final framework is. Ms. Malmgren: If I heard what you said correctly, and I think a message from today is as a market person it sounds like effectively through new measures regulators basically want to take the keys of the Maserati away, and get the markets with less leverage and more capital requirements to behave somewhat more sensibly and into a more sensible kind of vehicle. But, the thing is, markets still will want the Maserati, and we know from the past re-regulation experience, Sarbanes-Oxley, where we effectively deemed off-balance-sheet activities to not be material as one example. We now know how that played out. What are the chances regulators will begin to look again at what will be effectively deemed not material because that is where the markets are going to try to scheme to get the keys to the Maserati back? Mr. Draghi: Again, another question which is in a sense in the same vein. What will happen when the situation becomes less strained than it is today or when the fear disappears? One answer that is more a strategic answer than a substantial one would be if we all have this worry—and we all have it—that it could be a pushback by the relevant sectors of the industry that then very little is achieved, then we shouldn’t miss the opportunity to cast our intentions or policies in iron now, using the current fear as a good reason to avoid future pushbacks. I don’t know whether this is possible everywhere, but certainly it is a strategy that the EU Commission is pursuing of speeding up deliberations so that some of what is being recommended in this report gets into place, into legislation soon. To say that this is possible over all constituencies—well, I would not be able to say anything like that.

08 Book.indb 378

2/13/09 3:58:50 PM

The Role of Liquidity in Financial Crises Franklin Allen and Elena Carletti

I.

Introduction

The crisis that started in the summer of 2007 came as a surprise to many people. However, for others it was not a surprise. John Paulson, the hedge fund manager, correctly predicted the subprime debacle and earned $3.7 billion in 2007 as a result.1 The vulnerabilities that the global financial system has displayed were hinted at beforehand in the Bank of England and other Financial Stability Reports.2 The Economist magazine had been predicting for some time that property prices in the U.S. and a number of other countries were a bubble and were set to fall.3 Although the fall in U.S. property prices that is the fundamental cause of the crisis was widely predicted, the effects that this had on financial institutions and markets were not. In particular, what has perhaps been most surprising is the role that liquidity has played in the current crisis. The purpose of this paper is to use insights from the academic literature on liquidity and crises to try to understand the role of liquidity during the last year. We focus on four possible effects of liquidity: on pricing, on interbank and collateralized markets, on fear of contagion, and on the real economy.

379

08 Book.indb 379

2/13/09 3:58:50 PM

380

Franklin Allen and Elena Carletti

One of the most puzzling features of the crisis has been the pricing of AAA tranches of a wide range of securitized products. It appears that the market prices of many of these instruments are significantly below what plausible fundamentals would suggest they should be. This pricing risk has come as a great surprise to many. We argue that the sharp change in pricing regimes that started in August 2007 is consistent with what is known in the academic literature as “cash-in-the-market” pricing. Holding liquidity is costly because less liquid assets usually have higher returns. In order for providers of liquidity to markets to be compensated for this opportunity cost, they must on occasion be able to make a profit by buying up assets at prices below fundamentals. Once the link between prices and fundamentals is broken, then arbitrage becomes risky and the usual forces that drive prices and fundamentals together no longer work. This limit to arbitrage means that prices can deviate from fundamentals for protracted periods. The second surprise has been the way in which the money markets have operated. The interbank markets for terms longer than a few days have experienced considerable pressures. In addition, the way that the collateralized markets operate has changed significantly. Haircuts have increased and low-quality collateral has become more difficult to borrow against. The Federal Reserve and other central banks have introduced a wide range of measures to try to improve the smooth functioning of the money markets. The extent to which these events affect the functioning of the financial system and justifies central bank intervention depends on the possible explanations as to why the markets stopped operating smoothly. One of the main roles of interbank markets is to reallocate liquidity among banks that are subject to idiosyncratic shocks. If banks hoard liquidity and as a result they are able to cover idiosyncratic shocks from their own liquidity holdings, then their unwillingness to lend to other banks is not a problem. If, on the contrary, the liquidity hoarding prevents the reshuffling of liquidity to deficient, but solvent banks, then the badly functioning interbank market is a problem warranting central bank liquidity provision. Allowing banks to exchange mortgage-backed securities for Treasuries is desirable if it improves collateralized lending in the repo market, but is not if it simply leads to more window dressing by financial institutions.

08 Book.indb 380

2/13/09 3:58:50 PM

The Role of Liquidity in Financial Crises

381

In this case, the actions of the Federal Reserve are simply removing market discipline. The third aspect of the crisis that we consider relates to contagion risk. The controversial use of public funds in the arranged merger of Bear Stearns with J. P. Morgan was justified by the possibility of contagion. If Bear Stearns had been allowed to fail, its extensive involvement as counterparty in many derivatives markets may have caused a string of defaults. There is a large literature on the likelihood of contagion between banks based on simulations. The conclusion of this literature is that contagion in banking is unlikely. However, some have argued that these simulations do not capture important elements of the process. Whatever one’s view of the likelihood of contagion in banking, it is important to conduct similar studies in the context of counterparty risk in derivatives and other markets. Much of the academic literature on the role of liquidity in financial crises has focused on the effects on the real economy, mainly through the provision of liquidity to non-financial firms. We argue this has not been a significant factor to date in the current crisis. However, this may change going forward. There is a growing literature on understanding the current crisis. Brunnermeier (2008) provides an excellent account of the sequence of events in the crisis focusing on a wide range of factors. Adrian and Shin (2008) argue that the dynamics of the crisis are driven by deleveraging. What sets our study apart from these papers is its primary focus on liquidity. We start in Section II with a brief overview of the crisis focusing on the factors that are important for our subsequent discussion. Section III considers what liquidity in our context actually is and how liquidity created by banks, which is the focus of our study, can be measured. In Section IV we explain a theoretical framework for understanding liquidity provision. Section V applies this framework to gain insights into the current crisis. Finally, Section VI contains concluding remarks.

08 Book.indb 381

2/13/09 3:58:50 PM

382

II.

Franklin Allen and Elena Carletti

Liquidity and the Crisis

The crisis that started in the summer of 2007 is one of the most dramatic and important crises of recent decades. Its causes and unfolding have highlighted a number of new concerns and issues for policy makers, practitioners as well as academics interested in financial and monetary issues. In the following we briefly outline the sequence of events. This provides a starting point for our discussion in subsequent sections. This description is mostly drawn from Federal Reserve Bank (2008a, 2008b) (see also Bank of England, 2008; Bernanke, 2008; European Central Bank, 2008; International Monetary Fund, 2007 and 2008; and Kohn, 2008). The crisis started in the first half of 2007 when the credit quality of subprime residential mortgages, in particular adjustable-rate ones, started to deteriorate. Mortgage companies specializing in subprime products experienced funding pressures and many failed. Although problems were initially confined to the subprime mortgage markets, further deterioration of credit quality and increases in the delinquency rates led to a spread of the crisis to other markets and products. By mid-2007 investors started to retreat from structured credit products and risky assets more generally, as rating agencies started downgrading many mortgage-backed securities. The securitization market for subprime mortgages simply broke down. Chart 1 shows that in July 2007 there was a tremendous jump in the co-movement of AAA-rated tranches of subprime mortgage-backed securities, commercial mortgage-backed securities, and securities linked to corporate credit quality. A general loss of confidence started to become pervasive. Signs of strain appeared in the leveraged syndicated loan market and in other leveraged lending markets in late June 2007, in the asset-backed commercial paper (ABCP) and in the term bank funding markets in August 2007. Spreads of collateralized loan obligations (CLOs) increased while the issuance of such debt reduced significantly, thus also reducing leveraged lending. Spreads on U.S. ABCP widened significantly in mid-August, while the volume of ABCP outstanding dropped significantly. This put substantial pressure on the structured

08 Book.indb 382

2/13/09 3:58:50 PM

The Role of Liquidity in Financial Crises

383

Chart 1 Co-movement between AAA-rated U.S. Structured Financial Instruments (in percent) 75.00

75.00

70.00

70.00

65.00

65.00

60.00

60.00

55.00

55.00

50.00

50.00

45.00

45.00

40.00

40.00

35.00

35.00

30.00

Jan/07 Feb/07 Mar/07 Apr/07 May/07 Jun/07 Jul/07 Aug/07 Sep/07 Oct/07 Nov/07 Dec/07 Jan/08 Feb/08Mar/08Apr/08

30.00

Source: Bank of England calculations using data from JP Morgan Chase and Co. – Bank of England Financial Stability Report (2008, p. 24, Chart 1.19). Graph plots the proportion of the variation in exponentially weighted daily changes in credit default swap premia for the most senior tranche of the ABX HE 2006 H1. CMBX NA Series 1 and CDX NA explained by the first principal component over a three-month rolling window.

investment vehicles (SIVs) that had heavily invested in structured financial products. Many had to activate the contingent liquidity support from their sponsor banks. At the same time, problems arose in the term interbank funding markets in the U.S., Europe and the U.K. Banks suddenly became much more unwilling to provide liquidity to other banks, especially for maturities longer than a few days. Reflecting that, Libor spreads rose significantly (Chart 2). The apparent reason for this liquidity hoarding was twofold. On the one hand, banks wanted to protect themselves against potential larger-than-anticipated liquidity needs deriving from the disruptions in the mortgage, syndicated loans and commercial paper markets. On the other hand, uncertainty about the counterparty risk increased as banks could not precisely assess their counterparties’ exposure to the subprime related securities and also to the other disrupted markets.

08 Book.indb 383

2/13/09 3:58:50 PM

384

120

Franklin Allen and Elena Carletti

Basis Points

Chart 2 Three-month Interbank Rates Relative to Expected Policy Rates 120 Sterling

100

US dollar

100

Euro

80

80

60

60

40

40

20

20

0

0 January

July

April

October

April

January

2007

2008

Sources: Bloomberg and Bank calculations – Bank of England (2008), Chart 10, p. 11. Spread of three-month Libor to three-month overnight indexed swap (OIS) rates.

After a relief of the tensions in September and October following a 50 basis point reduction in the federal funds rate, tensions mounted again in November and December when end-of-the-year considerations became an additional element fueling the uncertainty deriving from the subprime market crisis. Spreads widened significantly again in all affected markets, and a flight to quality led to a strong demand for safe assets and a sharp drop in Treasury bill yields. Problems mounted again in March 2008 when the release of news of further losses and write-downs due to the use of mark-to-market accounting increased concerns about the creditworthiness and the capital position of several institutions. Financial markets continued to be under great stress, particularly the markets for short-term uncollateralized and collateralized funding. Tensions culminated in midMarch 2008 when a sudden wholesale run on Bear Stearns impeded the investment bank obtaining funding on both unsecured and collateralized short-term financing markets. Indicators of counterparty risk started being more significantly affected. For example, the cost of

08 Book.indb 384

2/13/09 3:58:51 PM

The Role of Liquidity in Financial Crises

385

insurance against the default of large complex financial institutions (LCFIs), as measured by the credit default swap spreads, rose steadily in 2008 and reached an unprecedented peak around the time of the collapse of Bear Stearns (Chart 3). Central banks around the world accompanied the unfolding of the crisis with numerous interventions. Some of these interventions concerned reductions in policy rates (but the Fed also reduced the discount window rate in September 2007) as well as liquidity injections into the system. Other interventions concerned changes in the standard operational frameworks or the creation of more unusual, innovative forms of special liquidity schemes. Changes involved extensions in the maturity of central bank lending (in the U.S. both with respect to the discount window loans in September 2007 and the open market operations in March 2008) and widening of the collateral accepted. Special liquidity schemes introduced during the crisis include the Term Auction Facility in December 2007, through which credit is auctioned to depository institutions against Discount Window collateral; the Term Securities Lending Facility in March 2008, which allows primary dealers to swap less-liquid mortgage and other asset-backed securities for Treasury securities; and, after the collapse of Bear Stearns, the Primary Dealer Credit Facility, through which the discount window was extended to primary dealers. Similarly, a special liquidity scheme was introduced in the U.K. in April 2008, according to which institutions eligible for the standing facilities can swap collateral with Treasury bills. Furthermore, both the Bank of England and the Federal Reserve were directly involved in managing and orchestrating the rescue, respectively, of Northern Rock and Bear Stearns, and the Federal Reserve recently established a temporary arrangement to provide emergency liquidity to Fannie Mae and Freddie Mac, should it become necessary. More recently, the U.S. Treasury has been given the power, though on a temporary basis, to extend unlimited credit to (and invest in the equity of ) the two government-sponsored enterprises. Although the real effects of the crisis have so far been contained to some extent, initial signs of propagation seem to be emerging. Credit standards and terms on both commercial and industrial (C&I) loans

08 Book.indb 385

2/13/09 3:58:51 PM

386

Franklin Allen and Elena Carletti

Chart 3 Major Large Complex Financial Institutions Credit Default Swap Premia 330

Basis Points

330 U.S. Securities Houses

300

300

U.S. Commercial banks Major U.K. banks

270

270

European LCFIs 240

240

210

210

180

180

150

150

120

120

90

90

60

60

30

30 0

0 Jan.

Mar.

May

July

September 2007

November

March

January 2008

Sources: Markit Group Limited, Thomson Datastream, published accounts and Bank calculations – Bank of England (2008), p. 11, Chart 9.

and commercial real estate loans tightened, and the yields on corporate bonds increased significantly over the first half of 2008 (see Federal Reserve Bank 2008, p. 12), indicating increasing pressures and risks for the nonfinancial corporate sector. Credit has remained available to the business sector so far, but household borrowing has slowed. Similar changes are occurring in the U.K. and Europe. The exchange rate of the dollar fluctuated during the crisis with a general trend towards depreciation against most currencies. Private payroll employment started falling substantially in February 2008, and inflation started also to be a source of concern. Economic growth remained slow in the first half of 2008, and the persistent weaknesses in the housing markets, together with the tightened conditions for credit to businesses and households, also weakened the projections for the second half of the year. III.

Liquidity Provision by Banks

The term liquidity is used in many different ways. For our purposes, liquidity is the ability to buy financial assets and real goods and services

08 Book.indb 386

2/13/09 3:58:51 PM

The Role of Liquidity in Financial Crises

387

immediately. The most liquid asset is cash. Current and deposit accounts and assets such as Treasury bills are also very liquid. They can be sold to raise cash at short notice with very little fall in price. How should the liquidity of the financial system be measured? The focus of our study is on financial institutions and in particular on banks. Berger and Bouwman (2008a) have suggested a method for measuring liquidity created by the banking system and have applied it to the U.S. They start by classifying all bank assets and liabilities together with off-balance sheet items as liquid, semi-liquid and illiquid. They then assign weights to these three categories and calculate the amount of liquidity created by the banking system. They consider several possible measures. Their preferred measure includes off-balance sheet activities. According to this measure, in 2003 the U.S. banking system created $2.843 trillion of liquidity. This represented 39% of gross total assets and 4.56 times the overall level of bank capital. The amount of liquidity created by the banking system increased every year between 1993 and 2003 and during this period almost doubled. In a subsequent paper, Berger and Bouwman (2008b) use their measure of liquidity to investigate the relationship between liquidity and crises. Their sample period from 1984-2008Q1 includes two banking crises, the credit crunch of the early 1990s and the current crisis. They focus on “abnormal” liquidity creation. This is defined to be the deviation from the time trend of liquidity creation adjusted for seasonal factors. They find that both banking crises in their sample have the feature that they were preceded by abnormal positive liquidity creation. This was particularly true for the current crisis. This reflects a buildup of capital and a loosening of lending standards. During the credit crunch of the early 1990s, liquidity fell. For the current crisis there is an indication of a fall after the start of the crisis, but unfortunately, their data set only goes up to the end of 2007. In order to understand the role of liquidity in the current financial crisis, it is necessary to develop a theoretical framework for understanding liquidity creation by the banking system and how this relates to crises.

08 Book.indb 387

2/13/09 3:58:51 PM

388

IV.

Franklin Allen and Elena Carletti

A Theoretical Framework of Liquidity Provision

Liquidity has clearly played a very important role in the current crisis. Therefore it is important to have a theoretical framework for thinking about liquidity provision by the banking system and its contribution to the occurrence of crises. What follows is not meant to be a literature review, but rather a very brief description of the relevant concepts related to the crisis using a few papers. Private Provision of Liquidity by the Financial System Asset pricing theory in financial economics that provides the tools for asset valuation and risk management relies on the assumptions of fully rational agents and perfect and complete markets. In these models, agents understand the risks involved in the investments they undertake and price them correctly. In a similar spirit, much of the theory that underlies central bank inflation-target policy in recent years relies on similar assumptions. In this frictionless world, financial institutions have no role to play, and financial crises should never occur. However, they do occur, and as the current crisis shows, badly functioning money markets, financial institutions and their role as liquidity creators can be at center stage. Understanding recent events in terms of models without financial intermediaries is difficult, to say the least. The first step in analyzing the role of liquidity in financial crises is to develop a model of liquidity provision in the context of financial institutions and markets. We need to understand how a financial system can provide liquidity efficiently and what can go wrong. We also need to consider the potential role of central banks in improving the allocation of resources and maintaining financial stability when there is a problem. The standard model of banking that allows consideration of the role of banks as liquidity providers was introduced by Bryant (1980) and Diamond and Dybvig (1983). There is a short asset that provides liquidity in the next period and a long asset that provides a higher return but at a later date. Consumers are initially unsure when they will require liquidity, and they cannot directly insure this risk. In this

08 Book.indb 388

2/13/09 3:58:51 PM

The Role of Liquidity in Financial Crises

389

view of the world, the role of banks is to provide liquidity insurance to depositors. The original banking models do not include financial markets. To understand the current crisis, it is essential to have a framework with both financial intermediaries and markets. Allen and Gale (2004a, 2007), among others, develop such an approach. They argue that in modern financial systems financial markets are essential for financial institutions. Consumers invest in financial intermediaries such as banks and mutual funds, and these institutions then invest in financial markets. Information and transaction costs make it too costly for individual investors to trade directly in the full range of financial markets. Both financial intermediaries and markets play an important role in this environment. Financial intermediaries provide liquidity insurance to consumers against their individual liquidity shocks. Markets allow financial intermediaries (and hence their depositors) to share aggregate risks. This general equilibrium framework allows a normative analysis of liquidity provision by the financial system. It is like the Arrow-Debreu model of resource allocation but includes financial institutions. It provides a benchmark for the efficient provision of liquidity by intermediaries and markets and an ideal allocation for a central bank to aim at implementing. Banks allow consumers to deposit funds that they can withdraw when they have liquidity needs. This liquidity provision allows banks to accumulate funds that they can use to lend to firms to fund longterm investments. Banks must manage their liquidity so that they can meet the liquidity needs of their depositors. There are two types of uncertainty concerning liquidity needs. The first is that each individual bank is faced with idiosyncratic liquidity risk. At any given date its customers may have more or less liquidity needs. The second type of uncertainty that banks face is aggregate liquidity risk. In some periods aggregate liquidity demand is high, while in other periods it is low. Thus, aggregate risk exposes all banks to the same shock, by increasing or decreasing the demand for liquidity that all banks face at the same time. The ability of banks to hedge themselves against these liquidity risks crucially depends on the functioning, or, more precisely, the completeness of financial markets.

08 Book.indb 389

2/13/09 3:58:51 PM

390

Franklin Allen and Elena Carletti

If financial markets are complete, the financial system provides liquidity efficiently in that it ensures that banks’ liquidity shocks are hedged. One way to implement complete markets that allow every bank to hedge itself against idiosyncratic liquidity risk is as follows. Each bank issues a small amount of a security contingent on the idiosyncratic liquidity shock experienced by each other bank. With the funds generated by these securities, each bank buys all of the securities issued by the other banks that are contingent on its own idiosyncratic shock. Thus when a bank is hit by a high liquidity shock, it obtains the funds it needs to cover its liquidity requirements.4 The equilibrium prices of all these bank-specific securities, together with securities that allow aggregate risk to be hedged, lead to the efficient provision of liquidity by the financial system. The invisible hand of the market ensures that the pricing of the complete set of securities provides the correct incentives for the provision of liquidity by the banking system in every state of the world. The key point here is that the implementation of complete markets requires a large number of bank-specific securities, but in practice we do not see anything that resembles this kind of situation or provides an equivalent allocation. One possible reason is that the infrastructure needed to support all the securities required for markets to be complete can be very costly in practice and thus not convenient. Although the current U.S. financial system has many securities and many are specifically contingent on the particular experiences of specific firms such as credit default swaps, it is still a far cry from enabling the type of hedging transactions that correspond to the theoretical benchmark of complete markets. If markets are incomplete, banks can trade only a limited number of assets and their ability to hedge liquidity risk changes dramatically. The incompleteness of markets leads to inefficient provision of liquidity by the financial system. This can generate cash-in-the-market pricing, where even the prices of safe assets can fall below their fundamental value, and lead to financial fragility, where even small shocks have large effects on asset prices. In addition, there can be contagion where shocks spread from one institution to another, leading to a

08 Book.indb 390

2/13/09 3:58:51 PM

The Role of Liquidity in Financial Crises

391

chain of bankruptcies. These effects provide an explanation of what can go wrong in imperfect financial markets. Financial Fragility and Cash-in-the-Market Pricing The problem with incomplete markets is that liquidity provision by the financial system is inefficient. The nature of risk management to ensure that the bank or intermediary has the correct amount of liquidity changes significantly from the case of complete markets. When markets are complete, it is possible, as explained above, to use securities to ensure liquidity is received in the situations when it is needed. The price system ensures adequate liquidity is provided in every state and is priced properly state by state. In this case, banks and other intermediaries buy liquidity in states where it is scarce by selling liquidity in states where it is plentiful for them, and the financial system allows risk sharing and insurance. In contrast, when markets are incomplete, liquidity provision is achieved by selling assets when liquidity is required. When liquidity is scarce, asset prices are determined by the available liquidity, or in other words, by the cash in the market. It is necessary that a proportion of financial institutions hold extra liquidity that allows them to buy up assets when liquidity is scarce. These suppliers of liquidity are no longer compensated for the cost of providing liquidity state by state. Instead, the cost must be made up on average across all states, and this is where the problem lies. The providers of liquidity have the alternative of investing in a productive long asset. There is an opportunity cost to holding liquidity since this has a lower return than the productive long asset. In order for agents to be willing to supply liquidity they must be able to make a profit in some states. If nobody held liquidity, the price of the long asset would collapse to zero. This would provide an incentive for some agents to hold liquidity since they can acquire assets cheaply. But if the price increased too much, then nobody would hold liquidity as this would not make any profit. Thus, in equilibrium prices will be bid to the level where the profit in the states where banks face high liquidity demand is sufficient to compensate the providers of liquidity for all the other states where they do not make any profit

08 Book.indb 391

2/13/09 3:58:51 PM

392

Franklin Allen and Elena Carletti

and simply bear the opportunity cost of holding liquidity. In other words, prices are low in the states where banks need more liquidity. But this is exactly the wrong time from an efficiency point of view for there to be a transfer from the banks who need liquidity to the providers of liquidity. There is in effect negative insurance and suboptimal risk sharing. Asset price volatility is costly because depositors are risk averse and their consumption varies across banks with high and low idiosyncratic liquidity risk.5 This leaves scope for central bank intervention. By engaging in open market operations to fix the price of the long asset (or equivalently fix the short-term interest rate), central banks can remove the inefficiency deriving from the asset price volatility and achieve the same allocation as with complete markets (Allen, Carletti and Gale, 2008). To summarize, when markets are incomplete, asset prices must be volatile to provide incentives for liquidity provision. This asset-price volatility can lead to costly and inefficient crises. There is a market failure that provides the justification for central bank and other kinds of intervention to improve the allocation of resources. Liquidity provision in the complete markets allocation provides a benchmark for judging the effectiveness of such intervention. Contagion A second important concept when markets are incomplete is contagion. The linkages between banks that interbank markets provide imply that problems in one bank can spread to other banks and can potentially disrupt the whole financial system. Allen and Gale (2000) analyze a variant of the basic model of liquidity provision described above to consider how this process works and the inefficiencies involved.6 As with financial fragility, the problem is concerned with liquidity provision but in a somewhat different way. The possibility of contagion arises from the overlapping claims that different banks have on one another rather than from asset price volatility. When one bank suffers a shock and defaults as a consequence, the other banks suffer a loss because their claims on the troubled bank fall in value. If this spillover effect is strong enough, it can cause a crisis throughout the system. In extreme cases, the crisis passes from bank to bank,

08 Book.indb 392

2/13/09 3:58:51 PM

The Role of Liquidity in Financial Crises

393

eventually having an impact on a much larger set of banks than the one in which the original shock occurred. If there is a large degree of interconnectedness between banks in the sense that many hold the assets of others, there are many links through which a crisis can spread. On the other hand, the importance of each link will be smaller. This means that a shock can be more easily absorbed by the capital buffer of each institution. If there are a few links but each involves a larger amount of funds, crises are more likely to spread because each bank’s capital buffer will be overwhelmed if another bank fails. Thus the case of some interconnectedness but not too much represents the most likely situation for contagion to occur. Contagion is an extremely worrying phenomenon for policy makers. The costs of bankruptcy of financial institutions can be large. A whole string of bankruptcies among banks can cause tremendous damage to the financial system, and this in turn has the potential to have large spillovers to the real economy. If firms no longer have access to funding from banks or other financial institutions then they may have to cut investment and their level of output significantly. Many factors affect the probability and the extent of contagion. One that seems to have played a role in the current crisis relates to the use of mark-to-market accounting. This accounting method has the benefit of reflecting the market value of the balance sheets of financial institutions and therefore of allowing regulators, investors and other users of accounting information to better assess the risk profile of financial institutions. This is true provided financial markets operate perfectly and prices correctly reflect the future earning power of assets. However, when markets do not work perfectly and prices do not always reflect the value of fundamentals as in the case where there is cash-in-the-market pricing, mark-to-market accounting exposes the value of the balance sheets of financial institutions to short-term and excessive fluctuations, and it can ultimately generate contagion. If there is cash-in-the-market pricing in one sector of the financial system, then other sectors can be affected by the change in the prices and may be forced to write down the value of their assets.7

08 Book.indb 393

2/13/09 3:58:51 PM

394

Franklin Allen and Elena Carletti

Asymmetric Information In our discussion of liquidity provision so far, asymmetric information has played a relatively small role. In particular, the assets that are traded are not characterized by asymmetric information. In the current crisis, many people believe that asymmetric information has played an important role (see, for example, Gorton, 2008). Bolton, Santos and Scheinkman (2008) have provided an interesting theory of liquidity provision with asymmetric information. In their model there are three sets of agents. These are investors with a short horizon, intermediaries and investors with a long horizon. The basic source of inefficiency is asymmetric information about asset values between long-horizon investors and financial intermediaries. Long-horizon investors cannot distinguish between an asset sale that is due to a liquidity need and an asset sale to offload low-quality securities. This asymmetric information leads to an adverse selection problem and consequently to a price discount. Bolton, Santos and Scheinkman assume that as time passes, the intermediaries learn more about the assets that they hold. This ensures that over time the adverse selection problem gets worse, and the price discount if an intermediary sells becomes greater. The basic problem an intermediary faces if it is hit by a liquidity shock is whether to sell its assets now at a discount or to try and ride out the crisis. The danger of doing this is that the intermediary runs the risk of having to sell at a greater discount if the crisis lasts longer than expected. It is shown that two types of rational expectations equilibrium exist. In what they call the immediate trading equilibrium, intermediaries sell assets immediately to ensure they have enough liquidity. In the delayed trading equilibrium, intermediaries try to ride out the crisis and only sell if they are forced to. For some parameter values, only the immediate trading equilibrium exists, while for others both do. Surprisingly, the authors are able to show that the delayed trading equilibrium is Pareto superior when both exist. The reason is that short-horizon investors undervalue long

08 Book.indb 394

2/13/09 3:58:51 PM

The Role of Liquidity in Financial Crises

395

assets while long-horizon investors undervalue cash. There is a gain from inducing short-horizon investors to hold more long assets and long-horizon investors more cash. This is what the delayed trading equilibrium does. The worse is the asymmetric information problem, the less is the gain as it impedes the operation of the market for the long assets. Spillovers to the Real Economy Much of the literature on liquidity provision has been concerned with the provision of liquidity to firms and resulting spillovers to the real economy. One of the important issues in crises is why problems in the financial system spill over into the real economy. The seminal contribution here is Holmström and Tirole (1998). In their model, entrepreneurs operate firms. These entrepreneurs need to provide costly effort for the firm to be successful. In order to ensure they are willing to do this, they need to be provided with part of the equity of the firm. This limits the ability of the firm to raise funds by issuing securities to outside investors. If a firm is hit by a liquidity shock and needs more funds to continue, it may be unable to raise them in the market. If it cannot continue because of this, then it may go bankrupt, and this can cause a significant loss in welfare. The occurrence of this event is more likely when credit markets are disrupted. In order to overcome this problem, the firm may need to hold liquid securities that it can sell in the event of a liquidity shock. If the private supply of such securities is insufficient, the government may be able to improve welfare by issuing government debt that can be held by firms. Now when firms are hit by a shock, they will have sufficient liquidity to continue. Another important contribution is Kiyotaki and Moore (1997). They show that small shocks can lead to large effects because of the role of collateral. A shock that lowers asset prices lowers the value of collateral. This means that less borrowing is possible, asset prices are further lowered and so on in a downward spiral. Disruptions in liquidity provision can be the shock that initially lowers asset prices and starts the problem.

08 Book.indb 395

2/13/09 3:58:51 PM

396

V.

Franklin Allen and Elena Carletti

Insights into the Current Crisis

In this section, we focus on four of the crucial features of the crisis that we argue are related to liquidity provision. The first is the fall of the prices of AAA-rated tranches of securitized products below fundamental values. The second is the effect of the crisis on the interbank markets for term funding and on collateralized money markets. The third is fear of contagion should a major institution fail. Finally, we consider the effects on the real economy. 1.

Effects of Liquidity on Pricing

One of the most surprising aspects of the crisis has been the collapse in prices of even the AAA-rated tranches of mortgage-backed securities and other structured credit products. Some banks have had to write down the AAA-rated super senior tranches of mortgage-linked collateralized debt obligations by as much as 30% (Tett, 2008) due to a fall in their market prices. According to the Bank of England (2008, pp. 18-21), if this change in price was due to deterioration in fundamentals, then it would be necessary to believe that the ultimate percentage loss rate of securitized subprime mortgages would be 38%. This would be justified, if, for example, 76% of households with subprime securitized mortgages would default and the loss given default rate was 50%. This seems, however, implausible given that none of the AAA-rated tranches have yet defaulted and, as the Bank of England also estimated, there should not be any future default in AAA-rated subprime mortgage-backed securities, even with a continued decline in U.S. house prices. It is not only AAA-rated tranches of subprime mortgage-backed instruments that have suffered but also commercial mortgage-backed securities and securitizations linked to corporate credit quality. As Chart 1 illustrates, at the start of the crisis the co-movement of these instruments rose dramatically. The high co-movement among different types of AAA-rated securities with different fundamentals suggests that it is probably not fundamentals driving the falls in prices. The framework developed in the previous section provides some insight into what could be determining prices. The movements

08 Book.indb 396

2/13/09 3:58:51 PM

The Role of Liquidity in Financial Crises

397

observed are consistent with the cash-in-the-market pricing of securities explained above. In this framework, it can be shown that aggregate shortages of liquidity can cause even risk-free securities to trade at a significant discount to their fundamental. Usually the theory is developed in terms of a single asset. However, the analysis can be applied to the case of multiple assets. With segmented markets, the theory can also explain why different but related types of securitites would also be affected so their prices would tend to fall as well. Participating in a market involves the initial fixed cost of finding out information about the security being traded. This fixed cost limits the number of participants. The structure of investment banks and other participants in markets is usually such that a desk will trade a number of related products to try to economize on this fixed cost. Risk management in these firms is such that in the short run there is a fixed limit on the total amount of cash available to purchase these securities. Our view is that as news about the subprime default problems came out, many investors changed their estimate of the risk of these securities and readjusted their portfolios. This led to a wave of selling and overwhelmed the capacity of the market to absorb sales. As a result, prices of even the AAA-tranches fell. The reason that the prices of other securities such as AAA-rated tranches of commercial mortgage-backed securities also fell is that they are traded by the same desks as securitized subprime products, and so sales of these also led to a drop in prices. One important feature of this pricing of AAA-rated tranches at such large discounts is their persistence. One might expect cash-inthe-market prices to persist for a few days. But once the limits on each desk’s ability to trade have had time to be adjusted, it would be natural to expect the desks to bid up the prices of the securities since there would appear to be a significant arbitrage opportunity. By going short in similar maturity Treasuries and investing in these AAArated tranches, a significant premium could apparently be earned. What prevents this? The answer is limits to arbitrage (Shleifer and Vishny, 1997). In particular, once the link between prices and fundamentals is broken, the difference between them may widen in the wrong direction during the period of holding the position.

08 Book.indb 397

2/13/09 3:58:51 PM

398

Franklin Allen and Elena Carletti

It is well known that such limits to arbitrage can prevent even virtually identical securities from trading at the same price. The classic example is the shares of the Dutch company Royal Dutch Petroleum and the British company Shell Transport and Trading. Before July of 2005 when the two entities were formally merged into a single company, the shares of Royal Dutch Petroleum and Shell Transport and Trading were Siamese twins that shared in the profits of the oil major. Royal Dutch received 60% of the dividends and earnings of the joint company, and Shell Transport and Trading received the remaining 40%. Standard asset pricing theory suggests they should have traded at a ratio of 60/40 = 1.5. In fact they traded at very different price ratios than this (see, for example, Brealey, Myers and Allen, 2008, p. 367). It is interesting to note that although the prices of AAA-rated tranches of non-subprime mortgage-backed securities such as commercial mortgage-backed and securitizations linked to corporate credit quality were significantly affected, the prices of conforming prime mortgagebacked securities issued by Fannie Mae and Freddie Mac were much less affected. This is not surprising given that here the arbitrage is virtually risk free given the implicit government guarantee provided to the securities of these government-sponsored enterprises. Once the value of AAA-rated tranches of securitized products fell significantly, it no longer became possible to fund the Structured Investment Vehicles (SIVs) and similar entities holding them using short-term finance. Thus the market for asset-backed commercial paper to finance such SIVs dried up since it was now clear the collateral was lower in value and also risky, whereas before it was thought to be safe. To avoid loss of reputation, the banks that had set up these SIVs were forced to bring the underlying assets back on to their balance sheets. Their need for liquidity was thus dramatically increased. In our view, one of the important features of the current crisis is therefore that cash-in-the-market pricing combined with limits to arbitrage has significantly affected the pricing of large volumes of fixed-income securities for significant periods of time. Effectively this means that the creation by banks of uninsured off-balance sheet vehicles that borrow short and invest long has significantly increased risk in the financial system. Moreover, until significant experience has

08 Book.indb 398

2/13/09 3:58:52 PM

The Role of Liquidity in Financial Crises

399

been gained concerning this type of risk of the cash-in-the-market pricing of such assets, the ability of financial institutions to manage risk exposures will be considerably impaired. Another possible explanation of the pricing anomalies in the AAArated tranches of securitized securities is that they are due to asymmetric information as, for example, in Bolton, Santos and Scheinkman (2008). Strong adverse selection and moral hazard problems provide a potential explanation for the large discounts in prices for risky securities like those backed by subprime mortgages. However, the fall of other AAA-rated securities as well as the co-movements of prices of these products as shown in Chart 1 are more difficult to explain. The deterioration in the fundamentals of the underlying instruments in commercial mortgagebacked securitizations and securitizations linked to corporate credit quality was much less. Some other factor must be at work for the asymmetric information to be consistent with what happened. 2.

The Effects on Interbank Markets and Collateralized Markets

The second feature of the current crisis that has caused some surprise is the effect on the money markets. In particular, volumes in the interbank markets for maturities beyond a few days were significantly reduced. Less surprisingly, in the collateralized money markets, the haircuts on collateral increased significantly, particularly for mortgage-backed securities as shown in Table 1. We consider each of these in turn. One of the important issues with the interbank markets is the cause of the increase in spreads shown in Chart 2. These strains were particularly severe in December of 2007 and led the Fed to introduce special measures to provide liquidity, including the introduction of the Term Auction Facility to lend against discount window collateral. Subsequently in March 2008, they lengthened the term they were willing to lend for in open market operations, introduced the Term Securities Lending Facilities to lend Treasuries against a broad range of collateral, and announced the Primary Dealer Credit Facility to lend bilaterally to primary dealers.

08 Book.indb 399

2/13/09 3:58:52 PM

400

Franklin Allen and Elena Carletti

Table 1 Typical “Haircut” or Initial Margin (in percent) January-May 2007

April 2008

U.S. treasuries

0.25

3

Investment-grade bonds

0-3

8-12

High-yield bonds

10-15

25-40

Equities

15

20

Investment grade CDS

1

5

Synthetic super senior

1

2

Senior leveraged loans

10-12

15-20

2nd lien leveraged loans

15-20

25-35

Mezzanine level loans

18-25

35+

ABS CDOs: AAA

2-4

15

AA

4-7

20

A

8-15

30-50

BBB

10-20

40-70

Equity

50

100

AAA CLO

4

10-20

AAA RMBS

2-4

10-20

Alt-a MBS

3-5

20-50

Sources: Citigroup; and IMF staff estimates – from International Monetary Fund (2008), Table 1.2, p. 23. Note: ABS = Asset-backed security; CDO = collateralized debt obligation; CDS = credit default swap: CLO = collateralized loan obligation; RMBS = residential mortgage-backed security.

An important question is why these strains occurred and whether the actions of the Federal Reserve were warranted. As mentioned in Section II, two explanations are typically given as to why the interbank markets came under such strain. The first is that banks were hoarding liquidity in anticipation that they would have significant liquidity needs going forward. For example, they faced the possibility of having to bring many assets in SIVs and other off-balance sheet entities back on balance sheet as asset-backed commercial paper markets dried up. Also, banks faced the prospect as the economy slowed down of corporations drawing down their lines of credit. All in all, liquidity had become scarce, and the prospect of uncertainty in aggregate demand for liquidity going forward meant banks wanted to hold onto as much as possible.8

08 Book.indb 400

2/13/09 3:58:52 PM

The Role of Liquidity in Financial Crises

401

The second explanation for the drying up of interbank markets is that increased uncertainty about the solvency of banks meant that they became unwilling to lend to each other. It is argued that uncertainty over which banks held subprime mortgages and the value of these, together with the uncertainty concerning other securitized assets, made it very difficult for banks to judge which banks they should lend to. If this is the explanation of the drying up of markets, then one would expect to see distrust of banks’ prospects going forward to be reflected in the pricing of credit default swaps on banks. It can be seen from Chart 3 that the spread on credit default swaps on banks were elevated in December 2007 but by a relatively small amount. This was much less than the spreads that occurred in March 2008 at the time that Bear Stearns collapsed. The relatively low spreads in December 2007 suggest that banks’ reluctance to lend to each other probably plays a relatively small part in explaining why markets dried up. Liquidity hoarding is probably a more important factor. If liquidity hoarding is the explanation, then the drying up of interbank markets may in fact not be a problem. It can be argued that the main role of interbank markets is to reallocate liquidity between banks to allow them to meet idiosyncratic liquidity shocks. If there is increased aggregate uncertainty about liquidity demand, banks will hold more liquidity and can then cover idiosyncratic demands without resorting to the interbank market. In this case, the drying up of liquidity does not pose a threat to financial stability. In contrast, if the unwillingness of banks to provide liquidity prevents the efficient reallocation of liquidity to banks in need of liquidity, then financial stability can be affected and central bank intervention is warranted. We next turn to the collateralized money markets. Much of the lending that occurs between financial institutions takes the form of short-term collateralized repurchase agreements. In normal times, a wide range of assets from Treasuries to mortgage-backed securities are used as collateral, and they are regarded as close substitutes. Haircuts vary but by relatively small amounts. Table 1 shows that this changed as the current crisis progressed. This is partly because of the valuation issues discussed in the previous section that makes the securities more risky as collateral. In addition, there is the issue that if there is a

08 Book.indb 401

2/13/09 3:58:52 PM

402

Franklin Allen and Elena Carletti

default, particularly of a major financial institution, there is likely to be a flight to quality. This should increase the value of Treasuries but reduce the value of lower quality collateral such as mortgage-backed securities. In extreme circumstances, the flight to quality may cause the value of the lower quality collateral to fall below the haircut the lender took. Thus Treasuries become a preferred form of collateral in times of crisis. In this view, the actions of the Federal Reserve and other central banks in making Treasuries more available by swapping them for lower quality collateral significantly helps the functioning of the repo markets. One of the interesting characteristics of the strains in the interbank markets is that they were most severe in December of 2007 and around quarter’s end in September 2007 and March 2008. This suggests that other considerations such as the desire of financial institutions to window dress may have also contributed to the strains. Musto (1997, 1999) presents persuasive evidence that financial institutions’ desire to look good at year’s end and the end of quarters leads to significant pricing effects in the money markets. Such desire may have been even more accentuated during the recent crisis. In this case, the actions of the Federal Reserve in exchanging Treasuries for mortgage-backed securities and lower quality collateral may actually hurt rather than help. Financial institutions can hold low-quality securities for the period where no reporting is required. They then briefly buy Treasuries so that the balance sheet they report to shareholders or regulators is high quality. Temporarily increasing the supply of Treasuries makes this kind of deception easier. It helps remove market and regulator discipline. An important issue is the extent to which the strains in the market and the increased appetite for Treasuries occurred because of a need for improved collateral or because of a desire to window dress. More research is needed to settle this issue and evaluate the desirability of the actions undertaken by the Federal Reserve and other central banks. One piece of information that would shed some light on the importance of these two factors is the extent to which low-quality collateral was swapped for Treasuries and the extent to which these transactions were reversed afterwards.

08 Book.indb 402

2/13/09 3:58:52 PM

The Role of Liquidity in Financial Crises

3.

403

Fear of Contagion

The justification that the Federal Reserve gave for arranging the takeover of Bear Stearns by J. P. Morgan was the fear of contagion (Minutes of the Federal Reserve, March 14, 2008). Bear Stearns was the counterparty in a large number of derivative transactions. The fear was that if they had gone bankrupt there would have been contagion through the network of derivative contracts that they were part of, and a large number of other financial institutions may have been adversely affected. Contagion was discussed above in Section IV. Theories of contagion have mostly been developed in the context of banks and interbank markets. They show how a shock to one bank that causes bankruptcy can cascade through the financial system and cause a string of bankruptcies. If bankruptcy costs are high, then this string of failures can be very costly. The effect on asset prices may be large if failed institutions are forced to liquidate assets and there is cashin-the-market pricing. Moreover, there may be significant spillovers into the real economy if a significant number of financial institutions fail. Contagion potentially provides a strong justification for central banks to intervene and save institutions such as Bear Stearns. The key issue is how likely this kind of damaging contagion is in practice. This depends on the number and size of counterparties active in the market as well as on the size of the interrelations among them. The more numerous are the counterparties and the smaller the interrelations, the less likely it is that a default of one counterparty leads to contagion. The reason is that the buffers of capital of the surviving intermediaries are more likely to be large enough to absorb the default, especially if each of them has only small claims with the troubled intermediary. Given the characteristics of the markets where Bear Stearns operated, it is quite possible that this would have been the case and no contagion would have occurred. Upper (2007) provides a survey of simulation exercises that look for evidence of contagious failures of financial institutions resulting from the mutual claims they have on one another. Most of these papers use balance sheet information to estimate bilateral credit relationships for different banking systems. The stability of the interbank

08 Book.indb 403

2/13/09 3:58:52 PM

404

Franklin Allen and Elena Carletti

market is tested by simulating the breakdown of a single bank. This methodology has been applied to the Belgian, German, Swiss, U.K. and U.S. banking systems, among others. These papers find that the banking systems demonstrate a high resilience, even to large shocks. Simulations of the worst-case scenarios show that banks representing less than 5% of total balance sheet assets would be affected by contagion on the Belgian interbank market, while for the German system the failure of a single bank could lead to the breakdown of up to 15% of the banking sector in terms of assets. These results heavily depend on how the linkages between banks, represented by credit exposures in the interbank market, are estimated. For most countries, data is extracted from banks’ balance sheets, which can provide information on the aggregate exposure of the reporting institution vis-à-vis all other banks. To estimate bank-to-bank exposures, it is generally assumed that banks spread their lending as evenly as possible. In effect, this assumption requires that banks are connected in a complete network. Hence the assumption might bias the results, in the light of the theoretical findings that better connected networks are more resilient to the propagation of shocks. The main finding of this literature is that contagion is unlikely. However, there are a number of reasons for caution in accepting this result and concluding that policy makers need not worry about contagion between banks. The first is that they do not model price effects of bankruptcy. Cifuentes, Ferrucci and Shin (2004) have argued that these price effects are the main transmission mechanism for contagion. As Upper (2007) points out, they also rely on the initial shock being confined to a single bank. If there is an initial shock that affects several banks simultaneously, then this can also lead to contagion being more likely. In the case of Bear Stearns, it is not clear from publicly available information how much contagion there would have been had it been allowed to fail. Press reports stress the large number of derivative contracts that Bear Stearns was a counterpart in. However, as argued above, this could mean that contagion was less likely because there would be more institutions with capital buffers to absorb the defaults. In any case, more simulations like those undertaken for banks

08 Book.indb 404

2/13/09 3:58:52 PM

The Role of Liquidity in Financial Crises

405

are needed in the context of derivatives to assess the likelihood of contagion with this kind of default. As a final point, one also has to keep in mind that even when there is a realistic risk of contagion that justifies central bank or government intervention, this also involves costs that should be traded off against the costs deriving from contagion. These costs of intervention include the future moral hazard associated with increased risk taking by financial institutions going forward. 4.

Effects on the Real Economy

As discussed in Section IV, much of the academic literature on liquidity has been concerned with firm’s access to funds. If firms are limited in the amount they can raise because of factors such as moral hazard and adverse selection, they may be limited in the amount they can invest or may even fail if they suffer a liquidity shock. By holding liquid assets they can avoid this problem. So far the indications outlined in Section II indicate that firms’ financing has not been affected too much, and in particular, firms have not had to greatly restrict their investment plans because of a lack of finance. However, credit standards and terms on corporate and real estate loans have tightened. In the first half of 2008, yields on corporate bonds also increased significantly. If the crisis continues to worsen, the effects on corporate finance discussed in the literature may begin to bite more seriously. VI.

Concluding Remarks

The fundamental cause of the current crisis has been the dramatic fall in property prices. Although this fall in property prices was widely anticipated, many aspects of the crisis that resulted were not, and these have considerably exacerbated the effects of the crisis. We have focused on three of the most important. These are the following. • The significant fall in prices of many AAA-rated tranches of securitized products, including many unrelated to subprime mortgages.

08 Book.indb 405

2/13/09 3:58:52 PM

406

Franklin Allen and Elena Carletti

• The drying up of interbank markets for maturities beyond a few days and the change in haircuts on collateralized lending. • The fear of contagion. We have argued that these phenomena are all intimately connected with the role of liquidity in financial crises. They have greatly exacerbated the effects of the crisis. We suggest that the significant discounts on AAA-rated tranches of securitized products that are too large to be explained by the underlying fundamentals are the result of cash-in-the-market pricing. These price movements were unanticipated and have produced a whole set of problems for risk management going forward. The drying up of liquidity in interbank markets is usually attributed to a mixture of liquidity hoarding by banks to counter the increased uncertainty over aggregate liquidity demand and fear of lending to other banks. At the end of 2007, the evidence seems to be that banks were to a large extent hoarding liquidity rather than refusing to lend to counterparts because credit default swaps on banks were only elevated somewhat. This is less of a problem than fear of lending as banks are not being refused credit. In normal times, high-quality asset-backed securities and Treasuries are close substitutes for collateral in the money markets. However, in crisis times they are not because the possibility of default will cause a flight to quality. This leads to a demand for Treasuries rather than asset-backed securities. It is desirable for central banks to meet this demand to improve the efficiency of the money markets. However, in times of stress there is also a heightened demand for Treasuries for window dressing purposes at quarter and year end. Meeting this increased demand for Treasuries is not desirable as it removes an important market discipline. It is important that current facilities that allow asset-backed securities to be swapped for Treasuries be evaluated in this light. Theoretical analysis suggests that the process of contagion where default cascades through the financial system represents a significant danger. Contagion was the justification for preventing the bankruptcy

08 Book.indb 406

2/13/09 3:58:52 PM

The Role of Liquidity in Financial Crises

407

of Bear Stearns as they were heavily involved as counterparties in the derivatives markets. However, little empirical work on the plausibility of contagion in the context of derivatives markets has been done. This is urgently needed. In the remainder of this section, we consider some open issues related to the role of liquidity in financial crises that deserve attention. The first concerns mark-to-market accounting. One of the points we have emphasized is that cash-in-the-market pricing leads to prices that do not reflect fundamentals. If that occurs, mark-to-market accounting for financial institutions has the disadvantage that it can understate the value of banks and other intermediaries and makes them appear insolvent when in fact they are not. Historic cost accounting has the advantage that it does not do this. On the other hand, it leads to bankrupt institutions that deserve to be closed being able to continue and possibly gamble for resurrection. In Allen and Carletti (2008b), we suggest that in financial crisis situations where liquidity is scarce and prices are low as a result, market prices should be supplemented with both model-based and historic cost valuations. The rest of the time, and in particular when asset prices are low because expectations of future cash flows have fallen, markto-market accounting should instead be used. The second issue is the “too big to save problem” of large banks in small countries. The Federal Reserve could easily prevent the threat of contagion posed by Bear Stearns. Even the threat of contagion posed by the failure of the largest banks in the U.S. such as Citigroup and Bank of America could be avoided by central bank and government intervention even though this may require the outlay of very large amounts of government funds. However, some banks are so large relative to the countries in which they are based that this is not the case. One example is Fortis in Belgium. This has assets that are greater in size than the GDP of Belgium. If it were to fail, it would be quite likely that a Belgian government (if one existed at the time) would be unwilling to intervene and assume fiscal responsibility because of the large size of the burden. In this case, the key issue would be how the burden would be shared between countries of the European Union. Ecofin (2008, p. 5) specifies that, “If public resources are involved,

08 Book.indb 407

2/13/09 3:58:52 PM

408

Franklin Allen and Elena Carletti

direct budgetary net costs are shared among affected Member States on the basis of equitable and balanced criteria.” Unfortunately, this lack of specificity is likely to lead to substantial delays in dealing with the situation as each country vies to improve its fiscal position. During this time, the prospect of contagion could effectively freeze many European and some global capital markets with enormous effects on the real economy. It is an urgent matter for the European Union to agree on specific ex ante burden sharing criteria for the costs of preventing large banking crises. The work along these lines that is currently under way needs to proceed rapidly. Even more worrying is the fact that there exist banks that may fail in small countries that are not part of a larger grouping. The classic example here is UBS and Credit Suisse in Switzerland. These two banks both have assets significantly in excess of Swiss GDP. It may literally be infeasible for the Swiss government to raise the funds to prevent their failure. In such cases, the potential damage caused by the prospect of contagion if one of them were to fail is very large. It is again an urgent task to devise a system to prevent this kind of problem from occurring. The International Monetary Fund or the Bank for International Settlements are obvious institutions to be assigned to deal with such problems. The alternative is to wait for the catastrophe to occur. In that case, consumers will subsequently be unwilling to invest in large banks in small countries. In the meantime, however, very large costs will have been imposed on the global economy.

Author’s note: We are grateful to Alessio De Vincenzo of the Bank of Italy for numerous helpful discussions and to our discussant, Peter Fisher. Radomir Todorov and Zhenrui Tang provided excellent research assistance.

08 Book.indb 408

2/13/09 3:58:52 PM

The Role of Liquidity in Financial Crises

409

Endnotes Financial Times, January 15, 2008, and June 18, 2008.

1

See, for example, Bank of England (2006) and (2007).

2

See, for example, The Economist (2005) and (2006).

3

See Allen, Carletti and Gale (2008) for a full description of how complete markets can be implemented. 4

Allen and Carletti (2006, 2008a) analyze in detail how this pricing mechanism works.

5

For a survey of the literature on contagion, see Allen and Babus (2008).

6

See Allen and Carletti (2008a) for an analysis of mark-to-market accounting when there is cash-in-the-market pricing. 7

8 See Allen, Carletti and Gale (2008) for an analysis of the relationship between aggregate liquidity risk and liquidity hoarding.

08 Book.indb 409

2/13/09 3:58:52 PM

410

Franklin Allen and Elena Carletti

References Adrian, T., and H. Shin (2008). “Liquidity and Leverage,” Federal Reserve Bank of New York Working Paper 328, May. Allen, F., and A. Babus (2008). “Networks in Finance,” Working Paper 08-07, Wharton Financial Institutions Center, University of Pennsylvania. Allen, F., and E. Carletti (2006). “Credit Risk Transfer and Contagion,” Journal of Monetary Economics 53, 89-111. Allen, F., and E. Carletti (2008a). “Mark-to-Market Accounting and Liquidity Pricing,” Journal of Accounting and Economics 45, 358-378. Allen, F., and E. Carletti (2008b). “Should Financial Institutions Mark to Market?” Bank of France Financial Stability Review, October, forthcoming. Allen, F., E. Carletti and D. Gale (2008). “Interbank Market Liquidity and Central Bank Intervention,” working paper, University of Pennsylvania. Allen, F., and D. Gale (2000). “Financial Contagion,” Journal of Political Economy 108, 1-33. Allen, F., and D. Gale (2004a). “Financial Intermediaries and Markets,” Econometrica 72, 1023-1061. Allen, F., and D. Gale (2004b). “Financial Fragility, Liquidity, and Asset Prices,” Journal of the European Economic Association 2, 1015-1048. Allen, F., and D. Gale (2007). Understanding Financial Crises, Clarendon Lecture Series in Finance, Oxford: Oxford University Press. Bank of England (2006), Financial Stability Report, October. Bank of England (2007), Financial Stability Report, April. Bank of England (2008), Financial Stability Report, April. Berger, A. and C. Bouwman (2008a). “Bank Liquidity Creation,” Review of Financial Studies, forthcoming. Berger, A., and C. Bouwman (2008b). “Financial Crises and Bank Liquidity Creation,” working paper, University of South Carolina. Bernanke, B., (2008). “Liquidity Provision by the Federal Reserve,” Speech, May 13, Board of Governors of the Federal Reserve System, www.federalreserve.gov/ newsevents/speech/bernanke20080513.htm. Bolton, P., T. Santos and J. Scheinkman (2008). “Inside and Outside Liquidity,” working paper, Columbia University.

08 Book.indb 410

2/13/09 3:58:52 PM

The Role of Liquidity in Financial Crises

411

Brealey, R., S. Myers and F. Allen (2008). Principles of Corporate Finance, Ninth Edition, New York: McGraw Hill. Brunnermeier, M. (2008), “Deciphering the 2007-08 Liquidity and Credit Crunch,” Journal of Economic Perspectives, forthcoming. Bryant, J. (1980). “A Model of Reserves, Bank Runs, and Deposit Insurance,” Journal of Banking and Finance 4, 335-344. Cifuentes, R., G. Ferrucci and H. Shin (2005). “Liquidity Risk and Contagion,” Journal of European Economic Association,” 3(2-3), 556-566. Diamond, D., and P. Dybvig (1983). “Bank Runs, Deposit Insurance, and Liquidity,” Journal of Political Economy 91, 401-419. Ecofin (2008). Memorandum of Understanding on Cooperation between the Financial Supervisory Authorities, Central Banks and Finance Ministries of the European Union on Cross-Border Financial Stability, 1 June 2008. Economist (2005). “Hear that Hissing Sound,” The Economist, December 8, 2005. Economist (2006). “What’s that Hissing Sound?” The Economist, August 24, 2006. European Central Bank (2008). Financial Stability Review, June. Federal Reserve Bank (2008a). Monetary Policy Report to the Congress, February. Federal Reserve Bank (2008b). Monetary Policy Report to the Congress, July. Gorton, G. (2008). “The Panic of 2007,” 2008 Jackson Hole Symposium. Holmström, B., and J. Tirole (1998). “Private and Public Supply of Liquidity,” Journal of Political Economy 106, 1-40. International Monetary Fund (2007). Global Financial Stability Report, October. International Monetary Fund (2008). Global Financial Stability Report, April. Kohn, D. (2008). “Money Markets and Financial Stability,” Speech May 29, Board of Governors of the Federal Reserve System, www.federalreserve.gov/newsevents/speech/kohn20080529a.htm. Kyotaki, N., and J. Moore (1997). “Credit Chains,” Journal of Political Economy 99, 220-264. Musto, D. (1997). “Portfolio Disclosures and Year-End Price Shifts,” Journal of Finance 52, 1563-1588. Musto, D. (1999). “Investment Decisions Depend on Portfolio Disclosures,” Journal of Finance 54, 935-952.

08 Book.indb 411

2/13/09 3:58:52 PM

412

Franklin Allen and Elena Carletti

Shleifer, A., and R. Vishny (1997). “The Limits to Arbitrage,” Journal of Finance 52, 35-55. Tett, G. (2008). “Why Triple A Prices are Out of Sync with Fundamentals,” Financial Times, May 2, 2008. Upper, C. (2007). “Using Counterfactual Simulations to Assess the Danger of Contagion in Interbank Markets,” Bank for International Settlements Working Paper 234, Basel.

08 Book.indb 412

2/13/09 3:58:52 PM

Commentary: The Role of Liquidity in Financial Crises Peter R. Fisher

Allen and Carletti provide an insightful review of the literature on liquidity and financial crises and a useful framework for considering the role of liquidity in the events of the past year. I find myself in fundamental agreement with what I take to be their two key points: first, on liquidity hoarding as the more significant explanation of the breakdown in interbank markets and, second, on the impact of cashin-the-market pricing on asset values. As a consequence of this agreement, my comments will necessarily digress into quibbling about how one reaches these conclusions, how they should be characterized and into my own thoughts on the key puzzle of the past year, the Federal Reserve’s new facilities and suggested areas for further work. Liquidity hoarding as “balance sheet defensiveness” In their analysis of the drying up of interbank lending markets, the authors conclude that “liquidity hoarding” by banks has probably been the more-important factor than has uncertainty about the condition of borrowers (Allen and Carletti, beginning on pg. 399). I certainly agree. (See Fisher, 2008.) In public, bankers would always prefer to blame uncertainty about their borrowers’ balance sheets than anxiety about their own balance sheets. However, in my own conversations with bank CFOs, treasurers, and trading desks 413

08 Book.indb 413

2/13/09 3:58:52 PM

414

Peter R. Fisher

from August of 2007 through March of 2008, there was a frank acknowledgement of a defensive concern with their ability to finance their own positions and those of their key customers. The simultaneous and generalized widening of unsecured, interbank lending rates across U.S. dollar, sterling and euro markets last August and the persistence of these wider spreads for the past year also support the idea of a lenders’ strike as the more useful explanation. I see “liquidity hoarding” as a form of “balance sheet defensiveness” by bankers unwilling to rent space on their balance sheets to their competitors at traditional spreads. A broad definition of liquidity as the growth of balance sheets, as expressed in the other recent work of Adrian and Shin (2008), should not be seen as a different subject but rather as the flip side of the same coin. This broad definition of liquidity as the growth rate of financial intermediaries’ aggregate balance sheets helps explain both the abundance of liquidity earlier in this decade and the subsequent scarcity of liquidity that began last summer. More importantly, it locates the concept of liquidity in a behavior (the willingness and ability to expand one’s balance sheet) that creates a flow rather than simply viewing liquidity as a stock to be allocated. Allen and Carletti’s discussion of aggregate as contrasted with idiosyncratic liquidity shocks (pg. 401) would benefit from further thinking about behaviors and flows rather than stocks. Having concluded that liquidity hoarding was the better explanation of interbank behavior, the authors surprisingly focus on “uncertainty in aggregate demand for liquidity” without corresponding attention to “aggregate supply.” Let me make a plea to the regulators and central bankers, however, to consider carefully the distinction between aggregate and idiosyncratic liquidity shocks before designing new liquidity rules or ratios or further altering central bank operations. It is critical that any new rules recognize the behavioral dimension of liquidity as something that a banking system creates (or destroys) and not as a stock to be rationed among banks. Thus, I would be skeptical as to whether different liquidity rules or ratios had they been adhered to, by themselves, would

08 Book.indb 414

2/13/09 3:58:52 PM

Commentary

415

have made things any better over the past year and I can easily see how they could have made things procyclically worse. I would also suggest further work on the appropriate central bank response to aggregate as opposed to idiosyncratic liquidity shocks as the issue seems much less clear cut to me. I can see the case for central bank intervention in both cases, depending on circumstances. For example, an aggregate liquidity shock caused by a central bank firming of monetary policy would not be a likely candidate for an aggressive central bank reaction. An idiosyncratic shock to a single firm of an extraordinary scale (such as a computer malfunction of a major clearing bank) or one that raised solvency concerns in the interbank market which the central bank knew to be unfounded would both be candidates for central bank lending. Cash-in-the-market pricing is an accurate description Allen and Carletti’s description of the impact of scarce liquidity on asset prices, in conditions of incomplete markets and as constrained by the limits to arbitrage (Allen and Carletti, 391-392, 397 citing Shleifer and Vishny 1997), is hauntingly familiar to the investment management practitioner, particularly one that thought high-quality, mortgage-related securities looked cheap in December, and in March, and again in June. Unfortunately, “cash-in-the-market pricing” by itself describes but does not explain the divorce of asset pricing from fundamentals— meaning the credit fundamentals of the underlying cash flows, not macro-economic fundamentals. Allen and Carletti observe: “When liquidity is scarce asset prices are determined by available liquidity or in other words by cash in the market.” But when liquidity is abundant asset prices are also determined by cash in the market, as was the case from 2004 through early 2007. But it is also the case that balance sheet expansion and contraction, and the broader definition of liquidity, do not explain the divorce between asset pricing and credit fundamentals.

08 Book.indb 415

2/13/09 3:58:52 PM

416

Peter R. Fisher

The puzzle that should haunt us With the benefit of hindsight, we cannot claim to be puzzled by the fact of falling house prices nor by the fact of a financial crisis. If we are candid, however, we should admit that we are still perplexed by the severity and longevity of the crisis, by the loss of financial firms’ ability to absorb losses and to provide liquidity and, thus, by the jeopardy this crisis poses to the real economy. The key questions that should haunt us are: (1) How can a system that was thought to be so well capitalized just 18 months ago have proved itself to be much more highly-leveraged (so much more poorly capitalized) than we thought? And (2) How did this leverage so abruptly and persistently translate itself into both a lack of liquidity and falling credit asset values? My own attempts to answer these disturbing questions focus on the prevalence of asset-based or “repo financing” and on the transformation—or degradation—this has wrought to our credit system. Let me acknowledge that in our highly-evolved financial system there is a daisy chain of agency problems—of misaligned incentives— both in the creation of credit (from asset originators to asset distributors to asset managers) and in the investment process (from beneficial owners of assets, to boards of directors, to staffs, to consultants and again to asset managers). But these agency problems in finance have been with us for some time and could just have easily been described in 1978, 1988 and 1998 as today. I see the daisy-chain of secured financing arrangements that have run through our financial system, and the asset-based rather than income- or cash-flow-based credit process which they reflect, as providing the more compelling insight into both the surge in liquidity and credit prices early in this decade and their subsequent collapse over the past year. The theory of a lower capital charge for secured financing rests on the assumption that the addition of pledged collateral lowers the risk to the lender. In the presence of both belts and suspenders it is assumed that the lender need hold less of a cushion (in the form of

08 Book.indb 416

2/13/09 3:58:52 PM

Commentary

417

loss bearing capital) against the risk of loss, where the belt is presumably the borrower’s ability to repay the debt out of cash flow and the suspenders are the borrower’s pledge of collateral. The degradation of our credit process comes about not by the fact of secured financing but when lenders cease to pay attention to the borrowers’ ability to repay out of cash flow and make their lending decisions solely on the basis of the expected value of the collateral and whatever haircut (or down payment) the lender can secure whether the borrowers be households or hedge funds. In our current system of transaction-based leverage the haircut becomes the loss absorber of first recourse. But the haircut is only a slice of the asset itself and, thus, the “capital” available to absorb losses on the asset is perfectly correlated with the asset. As the asset goes up in value this correlation appears to create an additional cushion and to justify the wisdom of the loan; but when the asset falls in value, the cushion decays at the same rate as the asset. As lenders seek to protect themselves by increasing their implicit capital cushion through increasing haircuts (as many intermediaries attempted to do earlier this year) their actions both confess their failure to look to the borrowers’ cash flow as the first recourse and demonstrate the procyclical nature of asset-based financing as the impact of rising haircuts on asset values becomes self-defeating. This is exactly parallel to the procyclical nature of secured financing described in a more general context by Kiyotaki and Moore (1997) as referenced by Allen and Carletti (pg. 395). With all the discussion about underwriting standards for home mortgages, it strikes me as more than a little odd that we have been observing and discussing a crisis in the financial system for more than a year and yet nobody has spoken about underwriting standards for lending to hedge funds, or SIVs, or REITS, or CDOs or broker dealers or banks. I believe this is a reflection of how deeply we are immersed in a culture of asset-based finance. But perhaps after a quarter century of a bull market in credit asset values—brought on by the persistent decline in nominal interest rates caused, in sequence, by disinflation, productivity gains, and an extended period

08 Book.indb 417

2/13/09 3:58:52 PM

418

Peter R. Fisher

of abnormally low real rates—we should not be surprised that our financial system has been re-engineered into an asset-based process that presumes rather than inquires into the cash flows of borrowers. While there are significant differences between the events of 2008 and of 1998, I am struck by the parallel in the procyclical mechanics that repo-based financing played both in story of LongTerm Capital Management and in the systemwide dynamics that began to unfold last summer. I would also suggest that the prevalence of repo-based financing helps explain the abruptness and persistence with which the de-levering has been translated into illiquidity and sharp asset price declines. For some time, the marginal buyer (or seller) of assets has been a levered buyer (or seller). Not in the sense of balance sheet leverage but, rather, levered in the transactional sense of only being in a position to buy those assets which can be funded in the repo market. This is true not only of the firms that are thought of as highly levered, like hedge funds, but also of a great deal of “long only” activity where the high volume and velocity of transactions creates reliance on repo financing to support the timely purchase of assets and a subsequent sorting out of positions and cash flows. As a consequence, “funding liquidity” has come to mean the ability to fund the purchase of an asset on leverage and illiquidity means the inability to fund (or extend the funding) of an asset on leverage. The procyclical nature of raising haircuts as a form of lender self-defense triggered both a shift in demand from secured to unsecured markets, overwhelming the traditional interbank markets, and a fall in asset prices that could not be sustained at higher haircuts. While economists and commentators can distinguish between funding liquidity and asset market liquidity (or depth), in market practice the two terms are commonly conflated because they are so closely linked. While different types of assets are recognized as having different liquidity characteristics, outside of money market eligible instruments, this liquidity itself is thought of as an asset’s ability to be financed. Thus, liquidity is not so much an alternative to investment (as in “being liquid” or “being invested”) but, in a world of

08 Book.indb 418

2/13/09 3:58:53 PM

Commentary

419

transactional leverage, “liquidity” is the means of becoming invested and illiquidity is the corresponding explanation for downward pressure on asset prices. In sum, the “cash in the market” that has driven asset prices both up and down is the cash that comes from lenders, not investors. The Federal Reserve’s New Facilities In discussing the Federal Reserve’s new facilities, Allen and Carletti focus principally on the swapping of Treasury securities for lower quality collateral and suggest contrasting perspectives on how this might be evaluated (pg. 401). On one hand, they point out that the collateral swap “helps the functioning of the repo markets in times of crisis” by expanding the supply of the preferred collateral. But on the other hand, to the extent that the swapping of Treasuries for lower quality collateral helps financial institutions window dress, they suggest that this may have contributed to the strains and “actually hurt more than help” by making it easier for the Fed’s counterparties to engage in the deception of hiding the quality of their balance sheets on reporting dates. I have several reactions. First, these are essentially the same thing: you cannot help the repo market without affecting the balance sheets of repo market participants. Second, of course it is about window dressing—trying to make balance sheets look less leveraged—but it is always about window dressing. Ten years ago a broker-dealer CFO described to me the process of managing his balance sheet through quarter-end statement dates as like flying a jumbo jet under the Gateway Arch in St. Louis. Banks and broker-dealers are always trying to manage down their leverage on quarter-end dates and over the past year this has been particularly intense. Twenty years ago, central bank orthodoxy, which came from the Bundesbank, held that no self-respecting central bank would want to use its balance sheet to monetize the profligacy of its own finance ministry. The irony was that accumulating foreign exchange reserves forced the Bundesbank to finance the profligacy of the U.S. Treasury—foreshadowing our current imbalanced relationship across the

08 Book.indb 419

2/13/09 3:58:53 PM

420

Peter R. Fisher

Pacific. Today, a new orthodoxy suggests that a central bank should only hold sovereign credit on its balance sheet as a way of avoiding the messy business of credit judgments. But in today’s monetary world, a central bank that lends only against sovereign credit is like a gold-regime central bank that lends only against gold: in a crisis it will end up sucking all of the preferred assets out of the market—by hogging the base asset for the central bank’s own balance sheet. To be relevant in a financial crisis, central banks have to lend against the assets the banks have not the assets they wish the banks had. The time to be fussy about the asset quality of the financial system’s balance sheet is when the assets are being created, not when they need lender-of-last-resort financing. The swapping out the Fed’s balance sheet holdings of Treasuries, and the expansion of the Discount Window both to an auction format and to primary dealers, are useful and necessary steps that indirectly help give the banking system time to de-lever—to shrink balance sheets down to their sustainable capital and income base. But none of the Fed’s facilities directly help the banks and broker-dealers to de-lever, because you cannot de-lever by borrowing money. In creating the auction mechanism for the Discount Window the Federal Reserve has sought to re-activate the banking system’s use of the lending facility that accepts a broader pool of collateral. As a former Manager of the System Open Market Account, a guilty conscience obliges me to confess that the non-use of the Discount Window by banks has been, to some extent, a self-inflicted wound. By providing an entirely elastic supply of reserves at a constant, targeted price and aiming to minimize the volatility in the fed funds rate, the Open Market Desk habituated the banking system to the non-use of the Discount Window. While the stigma of weakness associated with use of the Discount Window in the late 1980s and early 1990s certainly played a role in banks’ reluctance to seek borrowed reserves, by never forcing the banking system to take out borrowed reserves, the Federal Reserve habituated the banking system to a regime in which all needed reserves were provided through open market operations. Neither the Desk nor the Committee was willing

08 Book.indb 420

2/13/09 3:58:53 PM

Commentary

421

to tolerate the volatility in the funds rates that would, over time, have trained bank treasurers to use the Discount Window. Thus, I fear we have had too little rather than too much volatility in the fed funds rate. If the Federal Reserve’s actions have contributed to the practice of window dressing it is not through the advent of the recent swapping of Treasury securities for lower quality collateral but, rather, by the Fed’s routine willingness to provide a super abundance of reserves on quarter-end dates. Finally, Allen and Carletti may want to reflect upon the seemingly perverse consequences of the Fed’s efforts to limit the volatility of the fed funds rate as a contributor to higher intra-period leverage with reference to their conclusion that central bank interventions “can remove the inefficiency deriving from asset price volatility and achieve the same allocation as with complete markets” (pg. 392). We must be careful to distinguish removing volatility from merely shifting it. Contagion Allen and Carletti also discuss the fear of contagion as a rationale for central bank intervention, concluding that the main finding from the literature is that contagion is unlikely but that there are reasons for being cautious in accepting this result and that further work in this area should be undertaken (pg. 403-405). I certainly concur on the need for further work, particularly to get beyond consideration of direct exposures between financial firms and to delve further into indirect exposures. Counterparties should have a quite accurate picture of their direct exposures to a firm at risk of being closed. However, indirect exposures caused by parallel and correlated asset positions, as well as proxy hedging strategies, are harder to ascertain, harder assuage and, thus, more likely to stimulate herding behavior that could give rise to contagion. A final thought We need to be careful with the words we use. We have a problem of both too little capital and of too much capital. There is too little loss bearing capacity inside many financial intermediaries in the form of equity; but there is too much capital in the business of financial

08 Book.indb 421

2/13/09 3:58:53 PM

422

Peter R. Fisher

intermediation. The easy part of de-levering is the selling of financial assets to shrink balance sheets and the raising of new equity for those firms presumed to be survivors. The harder part will be contraction of the financial services industry. In the 1990s Japan made two mistakes of consequence. First, in the early 1990s the Bank of Japan ran a too restrictive monetary policy. In the latter part of the decade, the Japanese authorities were too slow in managing the process of consolidating their weakened banks. I hope we have learned both lessons from the Japanese experience.

References Adrian, T. & H. Shin (2008). “Liquidity, Monetary Policy, and Financial Cycles,” Federal Reserve Bank of New York, Current Issues in Economics and Finance, Vol. 14, No.1, Jan/Feb 2008. Fisher, P. (2008). “What happened to risk dispersion?” Bank of France, Financial Stability Review, Feb. 2008. Kiyotaki, N. & J. Moore (1997). “Credit Cycles,” Journal of Political Economy, Vol. 105, No. 2.

08 Book.indb 422

2/13/09 3:58:53 PM

General Discussion: The Role of Liquidity in Financial Crises Chair: Stanley Fischer

Mr. Makin: I would like to ask the authors and Peter if the liquidity problems they are discussing—and, Peter, your experience in the marketplace over the past 12 months—suggests to you the Fed ought to consider enlarging its balance sheet? More specifically, the Fed is the place where you can go for Treasuries to swap against securities that may be more difficult to turn into liquid assets. In the wake of problems, such as the failure of IndyMac and the incipient failure of other institutions, we see a situation developing where there is a run out of large deposits and into cash and/or Treasuries. (Personal anecdote: In March of this year when I was very nervous about my deposits in large institutions, I approached a mutual fund and asked them if I could put a substantial amount of money into their Treasury-only fund. And they said, “No, we already have too much of that going on.”) So, the notion there is going to be in a crisis entailing an excess demand for Treasuries suggests that the Fed ought to start buying more Treasuries in order to be able to supply them to panicky market participants who are running out of bank money. Does that notion follow from your discussion? 423

08 Book.indb 423

2/13/09 3:58:53 PM

424

Chair: Stanley Fischer

Mr. Lacker: The last few decades I’ve noticed an empirical regularity about financial crises that hasn’t gotten as much attention at this conference, but this conference is a good illustration—and it’s that financial crises give rise to a significant increase in references to asymmetric information and market frictions and appeals to them as rationales for government intervention of various sorts. This emerged as a promising line of research in the very early 1980s and was pursued with diligence and industry by many economists— some of them in this room. It has been a very helpful and very useful line of research. It has illuminated very many important phenomena. But it has been disappointing as well because what we found from those research endeavors is that it’s fairly difficult—not impossible, but fairly difficult—to build an efficiency-related rationale for government intervention. Obviously, what it requires is some comparative advantage with a government actor, such as superior information, superior technology, or the ability to tax. But the ability to tax implies the intervention is a redistribution rather than an efficiency enhancement. A fair reading of the literature on financial arrangements under limited information suggests deep humility about the economics of central bank credit market intervention. It occurred to me yesterday in the discussion about how prudential regulators ought to respond to credit cycles. You don’t have to stand on your head to build a model in which financial intermediaries varied their credit standards over the cycle in response to varying economic conditions in which that is optimal. In other words, the cyclical variation in credit standards is an effect and not a cause. It will be difficult to implement an optimal calibration intervening in those credit standard judgments. This humility suggests we entertain when we consider interventions or consider how we understand financial market crises a range of potential explanations for observable phenomena and check how well they line up against observations. The authors of this paper propose a cash-in-the-market friction as an explanation of last year’s phenomenon. I have a hard time buying this because we’ve heard all these

08 Book.indb 424

2/13/09 3:58:53 PM

General Discussion

425

reports of vast sums sitting on the sidelines waiting for more attractive prices. In fact, the discussant seems to be an instance of that. So, I would be interested in how they reconciled that observation with their friction. Besides, even if you grant the friction, it would explain the need for unsterilized intervention. Yet, what we have done is sterilized intervention because sterilized intervention doesn’t increase the amount of cash in the market. Here, Bordo’s distinction is important. As the authors are surely aware, observationally equivalent models would explain what happened to prices as deteriorating fundamentals. I am not sure how one rejects the notion the large discounts of mortgage-backed securities reflect the sense that, if returns were exceptionally low, it would be a very bad state of the world. Mr. Alexander: Chairman Bernanke yesterday talked a lot about improving infrastructure and settlement systems for securities markets. I wondered if the authors and Peter could comment on the degree to which (if we expanded those things like having central counterparties or pushing more trading onto exchanges) you think that would mitigate some of these problems? Mr. Landau: My question is about liquidity holding on the interbank market. A lot of people would agree that this is the reason, rather than counterparty risk, why interbank markets were disrupted in the last year. It is only fair to say that nobody expected that to happen beforehand. So, I was wondering whether we have some kind of fundamental explanation of this behavior, why liquidity demand can increase so fast up to almost an infinite amount or whether we have to accept that as a fact of life that there are jumps between different kinds of regime shifts where liquidity demand jumps up and down. It seems to me that it is very important to get to a kind of deep understanding and that before you even start thinking about what central banks should be doing in those situations.

08 Book.indb 425

2/13/09 3:58:53 PM

426

Chair: Stanley Fischer

Mr. McCulley: I have the same asset or liability as Peter; I am not sure which in being a practitioner. Theory is theory and practice is practice, and I confess that I was a very large liquidity hoarder, even though I was a net lender to the System last fall. My serious question is actually to the authors of the paper, which is that, while I enjoyed your paper, I felt a huge vacuum in that you did not discuss the framework of Hyman Minsky at all in explaining this phenomenon. My question is, Why? Mr. Bullard: Since I am not European, I’ll comment on UBS. One argument would be that the problems of UBS are well-known—the problems described are well-known—and the markets are well-aware of these problems. What they are doing is anybody who is doing business with UBS is pricing in this information and taking into account the firm might fail. For this reason, should they actually fail, the probability of contagion is not very high. But maybe Professor Allen thinks the markets aren’t pricing this in there. Either they are unwilling or unable to do so. Mr. Allen: Let me first of all thank Peter for his comments. They are very interesting, and I don’t think that we disagree with anything he said. So, let me turn to the questions and discuss some of the points raised there. The first question was about this issue of should the Fed supply more Treasuries and supply collateral to make things easy because there seems to be a shortage? Again, this gets back to this windowdressing issue. Peter was saying this is indeed what is going on; there is window-dressing. But that is a serious problem because one of the ways the market disciplines financial institutions is to see what risks they are taking in order to get the returns they’re earning. If everyone looks pretty good because they’re holding these Treasuries while the central banks have the junk stuff they’re holding most of the rest of the quarter, that is not a very good way of investors or regulators being able to figure out what is going on. One way to solve this would be to make it, for example, random, which day you had to declare your holdings of securities. So, instead

08 Book.indb 426

2/13/09 3:58:53 PM

General Discussion

427

of making it a specific day, you would say, “We’re going to draw a number of an urn, and then you have to tell us what you held that day.” It would be different for different things, and we would get rid of these effects. So there are other ways of dealing with these window-dressings. Jeff Lacker was talking about the electric chair and no convincing rationale for intervention. Let me make a couple of points here. He was talking about the tax argument and there being these redistributions, so it’s redistribution rather than efficiency. One of the key points is that what goes wrong is that if you look at the completemarkets case, what in fact is going on is you are having redistributions. That is what the complete markets are doing. They are allowing risk-sharing, by transferring funds from people. That is what is breaking down, I would argue, in many of these cases. The central bank has a role to play in correcting that problem. Let me also make a point, which I don’t think I made clearly enough in the talk, which is contagion is a big problem. Because if you go through this sequence of events that Chairman Bernanke described yesterday with a chain of bankruptcies, those are very costly. There are an awful lot of deadweight costs in the bankruptcy of financial institutions. If I were to say what’s the most important reason that we need intervention, I would use the contagion argument because there are real efficiency issues there. Now, a question about how these actual cash-in-the-market effects work and can we supply liquidity. It is very difficult to get liquidity into the right place. These markets are fairly segmented. For these kinds of fairly exotic securities, there aren’t huge numbers of traders in them, and it’s difficult to get cash in there quickly because they have capital constraints. You have to go back and say, “Look, there is a problem in this market. We need more capital so we can arbitrage and we can make a lot of money.” That all takes time because of the kinds of agency problems we discussed yesterday. That creates the problem that once you get these links broken, we are into this risky arbitrage. That is so important. Do people anticipate these changes? We will have periods where prices do deviate from fundamentals. Gary was saying yesterday—

08 Book.indb 427

2/13/09 3:58:53 PM

428

Chair: Stanley Fischer

this was kind of unique because of the subprime mortgage-backed securities—these problems can occur with many kinds of securities. Take the Long-Term Capital Management (LTCM) crisis. LTCM was doing the convergent trade, where they were shorting the lowyield liquid securities and going long in the high-yield illiquid securities. Arguably what happened there—I haven’t gone back and looked at the data, but I will do this in the next few months—is we got liquidity pricing in those markets with the default of the Russians. That caused prices to move in the wrong way. Liquidity pricing kicked in, and that caused the problem. We know in the end it worked out. We didn’t discuss Hyman Minsky. I guess I am not a great believer in behavioral kinds of explanations of these kinds of phenomena. I believe in highly rational people driven to make money. I like to look for frictions for why things don’t work and rational expectations of that. That would be my justification. The question with UBS, why isn’t everything priced in? There is a lot of inertia. One of the things that has astonished me is that they haven’t had more outflows. People in general don’t realize that if they were to go down, there wouldn’t be anybody to step in. Maybe the Swiss would save the Swiss citizens, but other citizens I’m not sure they would. That is rational expectations because there are costs of discovering this and that is what inertia is. It’s cost of discovering the issues. Anyway, I will close there. Mr. Fisher: We are in agreement on remedies to window-dressing. It somewhat depresses me to realize it was in 1994 I first proposed that with leverage and other risk measures we should get them out of financial firms on an intraperiod basis, where they would disclose the mode, the high, and the low observation over a 90-day period rather than the March 31, so we are certainly in agreement on the direction there. Maybe we can move that idea along in the coming decades. Should the Fed expand its balance sheets?—John Makin’s question. Let me say I guess I would gently urge them not to bother. There will be enough Treasury borrowing coming along soon enough to deal with this. The FDIC fund, Congress, and stimulus in general eventually

08 Book.indb 428

2/13/09 3:58:53 PM

General Discussion

429

will provide enough Treasuries, and so that would be a very brief intervention that might be necessary. It was Jeffrey Lacker’s question on cash in the market and all the money on the sidelines: First, a number of investors are aware of the limits to arbitrage and the downward pressure on prices brought about by collapsing balance sheets. Just as I think a necessary condition for house prices to stabilize in America is stabilizing some measures of debt to income for the household sector, stabilizing the balance sheets of financial intermediaries is going to have something to do with stabilizing their income-to-debt ratios. The revenue aperture coming into their balance sheets will still be contracting because that revenue is contracting and the investors are aware there is a downward cycle yet ahead of them. But more precisely probably in your vein is that there is inertia in hurdle rates, and these investors are looking for one or two turns of leverage in order to get the hurdle rates they want. They are optimists about central banks’ ability to get us back to an economy growing at trend, so they want to get a mid-teens return. They need one or two turns to leverage to do that. They have to go to some intermediary, someone else to provide them the leverage, and that is the connection. The cash in the market is he who is lending. As I meant to make clear in my remarks, the repo market is the cash in the market. Lew Alexander asked about infrastructure and if I am optimistic about infrastructure improvements making a difference. The answer is yes and no. A number of infrastructure improvements will make a difference. Clearly, 20 years of work on the foreign exchange settlement process reduced our anxieties about foreign exchange settlement mechanisms being the unzippering process for a banking crisis, but it did take decades, not a couple of years. I forget who yesterday made the comment that improving risk management is all about increasing your ability to take more risk. There are a number of areas where the banks and major dealers’ appetite for these improvements is principally because they want to take more risk. They want to be able to do more activities, generate more volume.

08 Book.indb 429

2/13/09 3:58:53 PM

430

Chair: Stanley Fischer

I don’t know this for a fact but I know it anecdotally, from the people who would be in a position to have an opinion that the credit default swap market is probably already more like $60 trillion to $70 trillion, much of that growth since the middle of March. That the major dealers are now wards of the state and, therefore, counterparty risk is down, might be contributing to an acceleration in the writing of this product. I have a particular anxiety about cleaning up the infrastructure of the CDS market, whether there will be delivery of bonds. There are trillions of dollars of contracts that have been written on the premise— if you read their terms—that a bond will be delivered to the writer of protection. And the writer of protection, then, only covers the difference with that and the original covered price. The dealers don’t want to bother with this; that would be a nuisance to have to go buy all those bonds and deliver them—it would cause a big settlement headache, and we couldn’t clear up the system. So they’d rather tear up all these contracts. But they’ve found a polite lawyerlike way to say this: We are going to have a credit event default protocol with an auction with a limited number of participants to determine the value of the collateral. If that goes through without the authorities coming up with the appropriate capital charge for this insurance industry, it will be a very risky thing. So—yes and no. There are some improvements in infrastructure that make a very big difference, and there are some that give me some pause.

08 Book.indb 430

2/13/09 3:58:53 PM

Rethinking Capital Regulation Anil K. Kashyap, Raghuram G. Rajan and Jeremy C. Stein

I.

Introduction

Recent estimates suggest that U.S. banks and investment banks may lose up to $250 billion from their exposure to residential mortgage securities.1 The resulting depletion of capital has led to unprecedented disruptions in the market for interbank funds and to sharp contractions in credit supply, with adverse consequences for the larger economy. A number of questions arise immediately. Why were banks so vulnerable to problems in the mortgage market? What does this vulnerability say about the effectiveness of current regulation? How should regulatory objectives and actual regulation change to minimize the risks of future crises? These are the questions we focus on in this paper. Our brief answers are as follows. The proximate cause of the credit crisis (as distinct from the housing crisis) was the interplay between two choices made by banks. First, substantial amounts of mortgagebacked securities with exposure to subprime risk were kept on bank balance sheets even though the “originate and distribute” model of securitization that many banks ostensibly followed was supposed to transfer risk to those institutions better able to bear it, such as unleveraged pension funds.2 Second, across the board, banks financed these and other risky assets with short-term market borrowing. 431

08 Book.indb 431

2/13/09 3:58:53 PM

432

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

This combination proved problematic for the system. As the housing market deteriorated, the perceived risk of mortgage-backed securities increased, and it became difficult to roll over short-term loans against these securities. Banks were thus forced to sell the assets they could no longer finance, and the value of these assets plummeted, perhaps even below their fundamental values—i.e., funding problems led to fire sales and depressed prices. And as valuation losses eroded bank capital, banks found it even harder to obtain the necessary short-term financing—i.e., fire sales created further funding problems, a feedback loop that spawned a downward spiral.3 Bank funding difficulties spilled over to bank borrowers, as banks cut back on loans to conserve liquidity, thereby slowing the whole economy. The natural regulatory reaction to prevent a future recurrence of these spillovers might be to mandate higher bank capital standards, so as to buffer the economy from financial-sector problems. But this would overlook a more fundamental set of problems relating to corporate governance and internal managerial conflicts in banks— broadly termed agency problems in the finance literature. The failure to offload subprime risk may have been the leading symptom of these problems during the current episode, but they are a much more chronic and pervasive issue for banks—one need only to think back to previous banking troubles involving developing country loans, highly-leveraged transactions, and commercial real estate to reinforce this point. In other words, while the specific manifestations may change, the basic challenges of devising appropriate incentive structures and internal controls for bank management have long been present. These agency problems play an important role in shaping banks’ capital structures. Banks perceive equity to be an expensive form of financing and take steps to use as little of it as possible; indeed, a primary challenge for capital regulation is that it amounts to forcing banks to hold more equity than they would like. One reason for this cost-of-capital premium is the high level of discretion that an equity-rich balance sheet grants to bank management. Equity investors in a bank must constantly worry that bad decisions by management will dissipate the value of their shareholdings. By contrast, secured short-term creditors are better

08 Book.indb 432

2/13/09 3:58:53 PM

Rethinking Capital Regulation

433

protected against the actions of wayward bank management. Thus, the tendency for banks to finance themselves largely with short-term debt may reflect a privately optimal response to governance problems. This observation suggests a fundamental dilemma for regulators as they seek to prevent banking problems from spilling over onto the wider economy. More leverage, especially short-term leverage, may be the market’s way of containing governance problems at banks; this is reflected in the large spread between the costs to banks of equity and of short-term debt. But when governance problems actually emerge, as they invariably do, bank leverage becomes the mechanism for propagating bank-specific problems onto the economy as a whole. A regulator focused on the proximate causes of the crisis would prefer lower bank leverage, imposed for example through a higher capital requirement. This will reduce the risk of bank defaults. However, the higher capital ratio will also increase the overall cost of funding for banks, especially if higher capital ratios in good times exacerbate agency problems. Moreover, given that the higher requirement holds in both good times and bad, a bank faced with large losses will still face an equally unyielding tradeoff—either liquidate assets or raise fresh capital. As we have seen during the current crisis, and as we document in more detail below, capital-raising tends to be sluggish. Not only is capital a relatively costly mode of funding at all times, it is particularly costly for a bank to raise new capital during times of great uncertainty. Moreover, at such times many of the benefits of building a stronger balance sheet accrue to other banks and to the broader economy and thus are not properly internalized by the capital-raising bank. Here is another way of seeing our point. Time-invariant capital requirements are analogous to forcing a homeowner to hold a fixed fraction of his house’s value in savings as a hedge against storm damage—and then not letting him spend down these savings when a storm hits. Given this restriction, the homeowner will have no choice but to sell the damaged house and move to a smaller place—i.e., to suffer an economic contraction.

08 Book.indb 433

2/13/09 3:58:53 PM

434

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

This analogy suggests one possible avenue for improvement. One might raise the capital requirement to, say, 10% of risk-weighted assets in normal times, but with the understanding that it will be relaxed back to 8% in a crisis-like scenario. This amounts to allowing some of the rainy-day fund to be spent when it rains, which clearly makes sense—it will reduce the pressure on banks to liquidate assets and the associated negative spillovers for the rest of the economy. Thus, time-varying capital requirements represent a potentially important improvement over the current time-invariant approach in Basel II. Still, even time-varying capital requirements continue to be problematic on the cost dimension. If banks are asked to hold significantly more capital during normal times—which, by definition, is most of the time—their expected cost of funds will increase, with adverse consequences for economic activity. This is because the fundamental agency problem described above remains unresolved. Investors will continue to charge a premium for supplying banks with large amounts of equity financing during normal times because they fear that this will leave them vulnerable to the consequences of poor governance and mismanagement. Pushing our storm analogy a little further, a natural alternative suggests itself, namely disaster insurance. In the case of a homeowner who faces a small probability of a storm that can cause $500,000 of damage, the most efficient solution is not for the homeowner to keep $500,000 in a cookie jar as an unconditional buffer stock—especially if, in a crude form of internal agency, the cookie jar is sometimes raided by the homeowner’s out-of-control children. Rather, a better approach is for the homeowner to buy an insurance policy that pays off only in the contingency when it is needed, i.e. when the storm hits. Similarly, for a bank, it may be more efficient to arrange for a contingent capital infusion in the event of a crisis, rather than keeping permanent idle (and hence agency-prone) capital sitting on the balance sheet.4 To increase flexibility, the choice could be left to the individual banks themselves. A bank with $500 billion in risk-weighted assets could be given the following option by regulators: it could either

08 Book.indb 434

2/13/09 3:58:53 PM

Rethinking Capital Regulation

435

accept a capital requirement that is 2% higher, meaning that the bank would have to raise $10 billion in new equity. Or, it could acquire an insurance policy that pays off $10 billion upon the occurrence of a systemic “event”—defined perhaps as a situation in which the aggregate write-offs of major financial institutions in a given period exceed some trigger level. In terms of cushioning the impact of a systemic event on the economy, the insurance option is just as effective as higher capital requirements. To make the policy default-proof, the insurer (say a sovereign wealth fund, a pension fund, or even market investors) would, at inception, put $10 billion in Treasuries into a “lock box.” If there is no event over the life of the policy, the $10 billion would be returned to the insurer, who would also receive the insurance premium from the bank as well as the interest paid by the Treasuries. If there is an event, the $10 billion would transfer to the balance sheet of the insured bank. From the bank’s perspective, the premium paid in insuring a systemic event triggered by aggregate bank losses may be substantially smaller than the high cost it has to pay for additional unconditional capital on balance sheet. This reduced cost of additional capital would in turn dampen the bank’s incentive to engage in regulatory arbitrage. Note that the insurance approach does not strain the aggregate capacity of the market any more than the alternative approach of simply raising capital requirements. In either case, we must come up with $10 billion when the new regulation goes into effect. Nevertheless, there may be some concern about whether a clientele will emerge to supply the required insurance on reasonable terms. In this regard, it is reassuring to observe that the return characteristics associated with writing such insurance have been much sought after by investors around the world—a higher-than-risk free return most of the time, in exchange for a small probability of a serious loss. Also, given the opt-in feature, if the market is slow to develop or proves to be too expensive, banks will always have the choice of raising more equity instead of relying on insurance.

08 Book.indb 435

2/13/09 3:58:53 PM

436

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

To be clear, capital insurance is not intended to solve all the problems associated with regulating banks. For example, to the extent that the trigger is only breached when a number of large institutions experience losses at the same time, the issue of dealing with a single failing firm that is very inter-connected to the financial system would remain. The opt-in aspect of our proposal also underscores the fact that one should not view capital insurance as a replacement for traditional capital regulation, but rather, as one additional element of the capital-regulation toolkit. What makes this one particular tool potentially valuable is that it is designed with an eye towards mitigating the underlying frictions that make bank equity expensive—namely the governance and internal agency problems that are pervasive in this industry. The added flexibility associated with the insurance option may therefore help to reduce the externalities associated with bank distress, while at the same time minimizing the potential costs of public bailouts during crises, as well as the drag on intermediation in normal times. More generally, our proposal reflects some pessimism that regulators can ever make the financial system fail-safe. Rather than placing the bulk of the emphasis on preventative measures, more attention should be paid to reducing the costs of a crisis. Or, using an analogy from Hoenig (2008), instead of attempting to write the most comprehensive fire code possible, we should give some thought to installing more sprinklers. The rest of the paper is organized as follows. In Section II, we describe the causes of the current financial crisis and its spillover effects onto the real economy. In Section III, we discuss capital regulation, with a particular focus on the limitations of the current system. In Section IV, we use our analysis to draw out some general principles for reform. In Section V, we develop our specific capital-insurance proposal. Section VI concludes. II.

The Credit-Market Crisis: Causes and Consequences

We begin our analysis by asking why so many mortgage-related securities ended up on bank balance sheets and why banks funded these assets with so much short-term borrowing.

08 Book.indb 436

2/13/09 3:58:53 PM

Rethinking Capital Regulation

437

II. A. Agency problems and the demand for low-quality assets Our preferred explanation for why bank balance sheets contained problematic assets, ranging from exotic mortgage-backed securities to covenant-light loans, is that there was a breakdown of incentives and risk-control systems within banks.5 A key factor contributing to this breakdown is that, over short periods of time, it is very hard, especially in the case of new products, to tell whether a financial manager is generating true excess returns adjusting for risk, or whether the current returns are simply compensation for a risk that has not yet shown itself but that will eventually materialize. Consider the following specific manifestations of the problem. Incentives at the top The performance of CEOs is evaluated based in part on the earnings they generate relative to their peers. To the extent that some leading banks can generate legitimately high returns, this puts pressure on other banks to keep up. Follower-bank bosses may end up taking excessive risks in order to boost various observable measures of performance. Indeed, even if managers recognize that this type of strategy is not truly value-creating, a desire to pump up their stock prices and their personal reputations may nevertheless make it the most attractive option for them (Stein, 1989; Rajan, 1994). There is anecdotal evidence of such pressure on top management. Perhaps most famously, Citigroup Chairman Chuck Prince, describing why his bank continued financing buyouts despite mounting risks, said: “When the music stops, in terms of liquidity, things will be complicated. But as long as the music is playing, you’ve got to get up and dance. We’re still dancing.” 6 Flawed internal compensation and control Even if top management wants to maximize long-term bank value, it may find it difficult to create incentives and control systems that steer subordinates in this direction. Retaining top traders, given the competition for talent, requires that they be paid generously based on performance. But high-powered pay-for-performance schemes create

08 Book.indb 437

2/13/09 3:58:53 PM

438

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

an incentive to exploit deficiencies in internal measurement systems. For instance, at UBS, AAA-rated mortgage-backed securities were apparently charged a very low internal cost of capital. Traders holding these securities were allowed to count any spread in excess of this low hurdle rate as income, which then presumably fed into their bonuses.7 No wonder that UBS loaded up on mortgage-backed securities. More generally, traders have an incentive to take risks that are not recognized by the system, so they can generate income that appears to stem from their superior abilities, even though it is in fact only a market risk premium.8 The classic case of such behavior is to write insurance on infrequent events, taking on what is termed “tail” risk. If a trader is allowed to boost her bonus by treating the entire insurance premium as income, instead of setting aside a significant fraction as a reserve for an eventual payout, she will have an excessive incentive to engage in this sort of trade. This is not to say that risk managers in a bank are unaware of such incentives. However, they may be unable to fully control them, because tail risks are by their nature rare, and therefore hard to quantify with precision before they occur. Absent an agreed-on model of the underlying probability distribution, risk managers will be forced to impose crude and subjective-looking limits on the activities of those traders who are seemingly the bank’s most profitable employees. This is something that is unlikely to sit well with a top management that is being pressured for profits.9 As a run of good luck continues, risk managers are likely to become increasingly powerless, and indeed may wind up being most ineffective at the point of maximum danger to the bank. II. B. Agency problems and the (private) appeal of short-term borrowing We have described specific manifestations of what are broadly known in the finance literature as managerial agency problems. The poor investment decisions that result from these agency problems would not be so systemically threatening if banks were not also highly levered, and if such a large fraction of their borrowing was not short-term in nature.

08 Book.indb 438

2/13/09 3:58:53 PM

Rethinking Capital Regulation

439

Why is short-term debt such an important source of finance for banks? One answer is that short-term debt is an equilibrium response to the agency problems described above.10 If instead banks were largely equity financed, this would leave management with a great deal of unchecked discretion, and shareholders with little ability to either restrain value-destroying behavior or to ensure a return on their investment. Thus, banks find it expensive to raise equity financing, while debt is generally seen as cheaper.11 This is particularly true if the debt can be collateralized against a specific asset, since collateral gives the investor powerful protection against managerial misbehavior. The idea that collateralized borrowing is a response to agency problems is a common theme in corporate finance (see, e.g., Hart and Moore, 1998), and of course this is how many assets—from real estate to plant and equipment—are financed in operating firms. What distinguishes collateralized borrowing in the banking context is that it tends to be very short-term in nature. This is likely due to the highly liquid and transformable nature of banking firms’ assets, a characteristic emphasized by Myers and Rajan (1998). For example, unlike with a plot of land, it would not give a lender much comfort to have a long-term secured interest in a bank’s overall trading book, given that the assets making up this book can be completely reshuffled overnight. Rather, any secured interest will have to be in the individual components of the trading book, and given the easy resale of these securities, will tend to short-term in nature. This line of argument helps to explain why short-term, often secured, borrowing is seen as significantly cheaper by banks than either equity or longer-term (generally unsecured) debt. Of course, shortterm borrowing has the potential to create more fragility as well, so there is a tradeoff. However, the costs of this fragility may in large part be borne systemically, during crisis episodes, and hence not fully internalized by individual banks when they pick an optimal capital structure.12 It is to these externalities that we turn next. II.C. Externalities during a crisis episode When banks suffer large losses, they are faced with a basic choice: Either they can shrink their (risk-weighted) asset holdings so that

08 Book.indb 439

2/13/09 3:58:54 PM

440

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

they continue to satisfy their capital requirements with their nowdepleted equity bases, or they can raise fresh equity. For a couple of reasons, equity-raising is likely to be sluggish, leaving a considerable fraction of the near-term adjustment to be taken up by asset liquidations. One friction comes from what is known as the debt overhang problem (Myers, 1977): By bolstering the value of existing risky debt, a new equity issue results in a transfer of value from existing shareholders. A second difficulty is that equity issuance may send a negative signal, suggesting to the market that there are more losses to come (Myers and Majluf, 1984). Thus, banks may be reluctant to raise new equity when under stress. It may also be difficult for them to cut dividends to stem the outflow of capital, for such cuts may signal management’s lack of confidence in the firm’s future. And a loss of confidence is the last thing a bank needs in the midst of a crisis. Chart 1 plots both cumulative disclosed losses and new capital raised by global financial institutions (these include banks and brokerage firms) over the last four quarters. As can be seen, while there has been substantial capital raising, it has trailed far behind aggregate losses. The gap was most pronounced in the fourth quarter of 2007 and the first quarter of 2008, when cumulative capital raised was only a fraction of cumulative losses. For example, through 2008Q1, cumulative losses stood at $394.7 billion, while cumulative capital raised was only $149.1 billion, leaving a gap of $245.6 billion. The situation improved in the second quarter of 2008, when reported losses declined, while the pace of capital raising accelerated. While banks may have good reasons to move slowly on the capitalraising front, this gradual recapitalization process imposes externalities on the rest of the economy. The fire-sale externality If a bank does not want to raise capital, the obvious alternative will be to sell assets, particularly those that have become hard to finance on a short-term basis.13 This creates what might be termed a fire-sale externality. Elements of this mechanism have been described in theoretical work by Allen and Gale (2005), Brunnermeier and Pedersen (2008), Kyle and Xiong (2001), Gromb and Vayanos (2002), Morris

08 Book.indb 440

2/13/09 3:58:54 PM

Rethinking Capital Regulation

441

Chart 1 Progress Towards Recapitalization by Global Financial Firms 600

600 Cumulative Writedowns Cumulative Capital Raised

500

500

Gap

400

400

300

300

200

200

100

100

0 2007 Q1

0 2007 Q2

2007 Q3

2007 Q4

2008 Q1

2008 Q2

Source: Bloomberg, WDCI , accessed August 6, 2008

and Shin (2004), and Shleifer and Vishny (1992, 1997) among others, and it has occupied a central place in accounts of the demise of Long-Term Capital Management in 1998. When bank A adjusts by liquidating assets—e.g., it may sell off some of its mortgage-backed securities—it imposes a cost on another bank B who holds the same assets: The mark-to-market price of B’s assets will be pushed down, putting pressure on B’s capital position and in turn forcing it to liquidate some of its positions. Thus, selling by one bank begets selling by others, and so on, creating a vicious circle. This fire-sale problem is further exacerbated when, on top of capital constraints, banks also face short-term funding constraints. In the example above, even if bank B is relatively well-capitalized, it may be funding its mortgage-backed securities portfolio with short-term secured borrowing. When the mark-to-market value of the portfolio falls, bank B will effectively face a margin call, and may be unable to roll over its loans. This too can force B to unwind some of its holdings. Either way, the end result is that bank A’s initial liquidation— through its effect on market prices and hence its impact on bank B’s

08 Book.indb 441

2/13/09 3:58:54 PM

442

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

price-dependent financing constraints—forces bank B to engage in a second round of forced selling, and so on. The credit-crunch externality What else can banks do to adjust to a capital shortage? Clearly, other more liquid assets (e.g. Treasuries) can be sold, but this will not do much to ease the crunch since these assets do not require much capital in the first place. The weight of the residual adjustment will fall on other assets that use more capital, even those far from the source of the crisis. For instance, banks may cut back on new lending to small businesses. The externality here stems from the fact that a constrained bank does not internalize the lost profits from projects the small businesses terminate or forego, and the bank-dependent enterprises cannot obtain finance elsewhere (see, e.g., Diamond and Rajan, 2005). Adrian and Shin (2008b) provide direct evidence that these balance sheet fluctuations affect various measures of aggregate activity, even controlling for short-term interest rates and other financial market variables. Recapitalization as a public good From a social planner’s perspective, what is going wrong in both the fire-sale and credit-crunch cases is that bank A should be doing more of the adjustment to its initial shock by trying to replenish its capital base, and less by liquidating assets or curtailing lending. When bank A makes its privately-optimal decision to shrink, it fails to take into account the fact that were it to recapitalize instead, this would spare others in the chain the associated costs. It is presumably for this reason that Federal Reserve officials, among others, have been urging banks to take steps to boost their capital bases, either by issuing new equity or by cutting dividends.14 A similar market failure occurs when bank A chooses its initial capital structure up front and must decide how much, if any, “dry powder” to keep. In particular, one might hope that bank A would choose to hold excess capital well above the regulatory minimum, and not to have too much of its borrowing be short-term, so that when losses hit, it would not be forced to impose costs on others. Unfortunately, to the extent that a substantial portion of the costs are

08 Book.indb 442

2/13/09 3:58:54 PM

Rethinking Capital Regulation

443

social, not private costs, any individual bank’s incentives to keep dry powder may be too weak. II.D. Alternatives for regulatory reform Since the banking crisis (as distinct from the housing crisis) has roots in both bank governance and capital structure, reforms could be considered in both areas. Start first with governance. Regulators could play a coordinating role in cases where action by individual banks is difficult for competitive reasons—for example, in encouraging the restructuring of employee compensation so that some performance pay is held back until the full consequences of an investment strategy play out, thus reducing incentives to take on tail risk. More difficult, though equally worthwhile, would be to find ways to present a risk-adjusted picture of bank profits, so that CEOs do not have an undue incentive to take risk to boost reported profits. But many of these problems are primarily for corporate governance, not regulation, to deal with, and given the nature of the modern financial system, impossible to fully resolve. For example, reducing high-powered incentives may curb excessive risk taking but will also diminish the constant search for performance that allows the financial sector to allocate resources and risk. Difficult decisions on tradeoffs are involved, and these are best left to individual bank boards rather than centralized through regulation. At best, supervisors should have a role in monitoring the effectiveness of the decision-making process. This means that the bulk of regulatory efforts to reduce the probability and cost of a recurrence might have to be focused on modifying capital regulation. III.

The Role of Capital Regulation

To address this issue, we begin by describing the “traditional view” of capital regulation—the mindset that appears to inform the current regulatory approach, as in the Basel I and II frameworks. We then discuss what we see to be the main flaws in the traditional view. For reasons of space, our treatment has elements of caricature: It is admittedly simplistic and probably somewhat unfair. Nevertheless,

08 Book.indb 443

2/13/09 3:58:54 PM

444

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

it serves to highlight what we believe to be the key limitations of the standard paradigm. III.A. The traditional view In our reading, the traditional view of capital regulation rests largely on the following four premises. Protect the deposit insurer (and society) from losses due to bank failures Given the existence of deposit insurance, when a bank defaults on its obligations, losses are incurred that are not borne by either the bank’s shareholders or any of its other financial claimholders. Thus, bank management has no reason to internalize these losses. This observation yields a simple and powerful rationale for capital regulation: A bank should be made to hold a sufficient capital buffer such that, given realistic lags in supervisory intervention, etc., expected losses to the government insurer are minimized. One can generalize this argument by noting that, beyond just losses imposed on the deposit insurer, there are other social costs that arise when a bank defaults—particularly when the bank in question is large in a systemic sense. For example, a default by a large bank can raise questions about the solvency of its counterparties, which in turn can lead to various forms of gridlock. In either case, however, the reduced-form principle is this: Bank failures are bad for society, and the overarching goal of capital regulation—and the associated principle of prompt corrective action—is to ensure that such failures are avoided. Align incentives A second and related principle is that of incentive alignment. Simply put, by increasing the economic exposure of bank shareholders, capital regulation boosts their incentives to monitor management and to ensure that the bank is not taking excessively risky or otherwise valuedestroying actions. A corollary is that any policy action that reduces the losses of shareholders in a bad state is undesirable from an ex ante incentive perspective—this is the usual moral hazard problem.

08 Book.indb 444

2/13/09 3:58:54 PM

Rethinking Capital Regulation

445

Higher capital charges for riskier assets To the extent that banks view equity capital as more expensive than other forms of financing, a regime with “flat” (non-risk-based) capital regulation inevitably brings with it the potential for distortion, because it imposes the same cost-of-capital markup on all types of assets. For example, relatively safe borrowers may be driven out of the banking sector and forced into the bond market, even in cases where a bank would be the economically more efficient provider of finance. The response to this problem is to tie the capital requirement to some observable proxy for an asset’s risk. Under the so-called IRB (internal-ratings-based) approach of the Basel II accord, the amount of capital that a bank must hold against a given exposure is based in part on an estimated probability of default, with the estimate coming from the bank’s own internal models. These internal models are sometimes tied to those of the rating agencies. In such a case, riskbased capital regulation amounts to giving a bank with a given dollar amount of capital a “risk budget” that can be spent on either AAArated assets (at a low price), on A-rated assets (at a higher price), or on B-rated assets (at an even higher price). Clearly, a system of risk-based capital works well only insofar as the model used by the bank (or its surrogate, the rating agency) yields an accurate and not-easily-manipulated estimate of the underlying economic risks. Conversely, problems are more likely to arise when dealing with innovative new instruments for which there exists little reliable historical data. Here the potential for mischaracterizing risks—either by accident, or on purpose, in a deliberate effort to subvert the capital regulations—is bound to be greater. License to do business A final premise behind the traditional view of capital regulation is that it forces troubled banks to seek re-authorization from the capital market in order to continue operating. In other words, if a bank suffers an adverse shock to its capital, and it cannot convince the equity market to contribute new financing, a binding capital requirement will necessarily compel it to shrink. Thus, capital requirements can be said to impose a type of market discipline on banks.

08 Book.indb 445

2/13/09 3:58:54 PM

446

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

III.B. Problems with the traditional mindset The limits of incentive alignment Bear Stearns’ CEO Jim Cayne sold his 5,612,992 shares in the company on March 25, 2008, at price of $10.84, meaning that the value of his personal equity stake fell by over $425 million during the prior month. Whatever the reasons for Bear’s demise, it is hard to imagine that the story would have had a happier ending if only Cayne had had an even bigger stake in the firm, and hence higherpowered incentives to get things right. In other words, ex ante incentive alignment, while surely of some value, is far from a panacea—no matter how well incentives are aligned, disasters can still happen. Our previous discussion highlights a couple of specific reasons why even very high-powered incentives at the top of a hierarchy may not solve all problems. First, in a complex environment with rapid innovation and short histories on some of the fastest-growing products, even the best-intentioned people are sometimes going to make major mistakes. And second, the entire hierarchy is riddled with agency conflicts that may be difficult for a CEO with limited information to control. A huge bet on a particular product that looks, in retrospect, like a mistake from the perspective of Jim Cayne may have represented a perfectly rational strategy from the perspective of the individual who actually put the bet on—perhaps he had a bonus plan that encouraged risk taking, or his prospects for advancement within the firm were dependent on a high volume of activity in that product. Fire sales and large social costs outside of default Perhaps the biggest problem with the traditional capital-regulation mindset is that it places too much emphasis on the narrow objective of averting defaults by individual banks, while paying too little attention to the fire-sale and credit-crunch externalities discussed earlier.15 Consider a financial institution, which, when faced with large losses, immediately takes action to bring its capital ratio back into line by liquidating a substantial fraction of its asset holdings.16 On the one hand, this liquidation-based adjustment process can be seen as precisely the kind of “prompt corrective action” envisioned by fans of capital regulation with a traditional mindset. And there is no

08 Book.indb 446

2/13/09 3:58:54 PM

Rethinking Capital Regulation

447

doubt that from the perspective of avoiding individual bank defaults, it does the trick. Unfortunately, as we have described above, it also generates negative spillovers for the economy: Not only is there a reduction in credit to customers of the troubled bank, there is also a fire-sale effect that depresses the value of other institutions’ assets, thereby forcing them into a similarly contractionary adjustment. Thus, liquidation-based adjustment may spare individual institutions from violating their capital requirements or going into default, but it creates a suboptimal outcome for the system as a whole. Regulatory arbitrage and the viral nature of innovation Any command-and-control regime of regulation creates incentives for getting around the rules, i.e., for regulatory arbitrage. Compared to the first Basel accord, Basel II attempts to be more sophisticated in terms of making capital requirements contingent on fine measures of risk; this is an attempt to cut down on such regulatory arbitrage. Nevertheless, as recent experience suggests, this is a difficult task, no matter how elaborate a risk-measurement system one builds into the regulatory structure. One complicating factor is the viral nature of financial innovation. For example, one might argue that AAA-rated CDOs were a successful product precisely because they filled a demand on the part of institutions for assets that yielded unusually high returns, given their low regulatory capital requirements.17 In other words, financial innovation created a set of securities that were highly effective at exploiting skewed incentives and regulatory loopholes. (See, e.g., Coval, Jurek and Stafford, 2008a, b; and Benmelech and Dlugosz, 2008.) Insufficient attention paid to cost of equity A final limitation of the traditional capital-regulation mindset is that it simply takes as given that equity capital is more expensive than debt, but does not seek to understand the root causes of this wedge. However, if we had a better sense of why banks viewed equity capital as particularly costly, we might have more success in designing policies that moderated these costs. This in turn would reduce the drag

08 Book.indb 447

2/13/09 3:58:54 PM

448

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

on economic growth associated with capital regulation, as well as lower the incentives for regulatory arbitrage. Our discussion above has emphasized the greater potential for governance problems in banks relative to non-financial firms. This logic suggests that equity or long-term debt financing may be much more expensive than short-term debt, not only because long-term debt or equity has little control over governance problems, it is also more exposed to the adverse consequences. If this diagnosis is correct, it suggests that rather than asking banks to carry expensive additional capital all the time, perhaps we should consider a conditional capital arrangement that only channels funds to the bank in those bad states of the world where capital is particularly scarce, where the market monitors bank management carefully, and hence where excess capital is least likely to be a concern. We will elaborate on one such idea shortly. IV.

Principles for Reform

Having discussed what we see to be the limitations of the current regulatory framework for capital, we now move on to consider potential reforms. We do so in two parts. First, in this section, we articulate several broad principles for reform. Then, in Section V, we offer one specific, fleshed-out recommendation. IV.A. Don’t just fight the last war In recent months, a variety of policy measures have been proposed that are motivated by specific aspects of the current crisis. For example, there have been calls to impose new regulations on the rating agencies, given the large role generally attributed to their perceived failures. Much scrutiny has also been given to the questionable incentives underlying the “originate to distribute” model of mortgage securitization (Keys, et al., 2008). And there have been suggestions for modifying aspects of the Basel II risk-weighting formulas, e.g., to increase the capital charges for highly-rated structured securities. While there may well be important benefits to addressing these sorts of issues, such an approach is inherently limited in terms of its ability to prevent future crises. Even without any new regulation, the one thing we can be almost certain of is that when the next

08 Book.indb 448

2/13/09 3:58:54 PM

Rethinking Capital Regulation

449

crisis comes, it won’t involve AAA-rated subprime mortgage CDOs. Rather, it will most likely involve the interplay of some new investment vehicles and institutional arrangements that cannot be fully envisioned at this time. This is the most fundamental message that emerges from taking a viral view of the process of financial innovation—the problem one is trying to fight is always mutating. Indeed, a somewhat more ominous implication of this view is that the seeds of the next crisis may be unwittingly planted by the regulatory responses to the current one: Whatever new rules are written in the coming months will spawn a new set of mutations whose properties are hard to anticipate. IV.B. Recognize the costs of excessive reliance on ex ante capital Another widely discussed approach to reform is to simply raise the level of capital requirements. We see several possible limitations to this strategy. In addition to the fact that it would chill intermediation activity generally by increasing banks’ cost of funding, it would also increase the incentives for regulatory arbitrage. While any system of capital regulation inevitably creates some tendency towards regulatory arbitrage, basic economics suggests that the volume of this activity is likely to be responsive to incentives—the higher the payoff to getting around the rules, the more creative energy will be devoted to doing so. In the case of capital regulation, the payoff to getting around the rules is a function of two things: i) the level of the capital requirement; and ii) the wedge between the cost of equity capital (or whatever else is used to satisfy the requirement) and banks’ otherwise preferred form of financing. Simply put, given the wedge, capital regulation will be seen as more cumbersome and will elicit a more intense evasive response when the required level of capital is raised. A higher capital requirement also does not eliminate the fire-sale and credit-crunch externalities identified above. If a bank faces a binding capital requirement—with its assets being a fixed multiple of its capital base—then when a crisis depletes a large chunk of its capital, it must either liquidate a corresponding fraction of its assets or raise new capital. This is true whether the initial capital requirement is 8% or 10%.18

08 Book.indb 449

2/13/09 3:58:54 PM

450

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

A more sophisticated variant involves raising the ex-ante capital requirement, but at the same time pre-committing to relax it in a bad state of the world.19 For example, the capital requirement might be raised to 10% with a provision that it would be reduced to 8% conditional on some publicly observable crisis indicator.20 Leaving aside details of implementation, this design has the appeal that it helps to mitigate the fire-sale and credit-crunch effects: Because banks face a lower capital requirement in bad times, there is less pressure on them to shrink their balance sheets at such times (provided, of course, that the market does not hold them to a higher standard than regulators). In light of our analysis above, this is clearly a helpful feature. At the same time, since crises are by definition rare, this approach has roughly the same impact on the expected cost of funding to banks as one of simply raising capital requirements in an un-contingent fashion. In particular, if a crisis only occurs once every ten years, then in the other nine years this looks indistinguishable from a regime with higher un-contingent capital requirements. Consequently, any adverse effects on the general level of intermediation activity, or on incentives for regulatory arbitrage, are likely to be similar. Thus if one is interested in striking a balance between: i) improving outcomes in crisis states, and ii) fostering a vibrant and non-distortionary financial sector in normal times, then even time-varying capital requirements are an imperfect tool. If one raises the requirement in good times high enough, this will lead to progress on the first objective, but only at the cost of doing worse on the second. IV.C. Anticipate ex post cleanups; encourage private-sector recapitalization Many of the considerations that we have been discussing throughout this paper lead to one fundamental conclusion: It is very difficult—probably impossible—to design a regulatory approach that reduces the probability of financial crises to zero without imposing intolerably large costs on the process of intermediation in normal times. First of all, the viral nature of financial innovation will tend to frustrate attempts to simply ban whatever “bad” activity was the proximate cause of the previous crisis. Second, given the complexity

08 Book.indb 450

2/13/09 3:58:54 PM

Rethinking Capital Regulation

451

of both the instruments and the organizations involved, it is probably naïve to hope that governance reforms will be fully effective. And finally, while one could in principle force banks to hold very large buffer stocks of capital in good times, this has the potential to sharply curtail intermediation activity, as well as to lead to increased distortions in the form of regulatory arbitrage. It follows that an optimal regulatory system will necessarily allow for some non-zero probability of major adverse events, and focus on reducing the costs of these events. At some level this is an obvious point. The more difficult question is what the policy response should then be once an event hits. On the one hand, the presence of systemic externalities suggests a role for government intervention in crisis states. We have noted that, in a crisis, private actors do too much liquidation and too little recapitalization relative to what is socially desirable. Based on this observation, one might be tempted to argue that the government ought to help engineer a recapitalization of the banking system or of individual large players. This could be done directly, through fiscal means, or more indirectly, e.g., via extremely accommodative monetary policy that effectively subsidizes the profits of the banking industry. Of course, ad hoc government intervention of this sort is likely to leave many profoundly uncomfortable, and for good reason, even in the presence of a well-defined externality. Beyond the usual moral hazard objections, there are a variety of political-economy concerns. If, for example, there are to be meaningful fiscal transfers in an effort to recapitalize a banking system in crisis, there will inevitably be some level of discretion in the hands of government officials regarding how to allocate these transfers. And such discretion is, at a minimum, potentially problematic. In our view, a better approach is to recognize up front that there will be a need for recapitalization during certain crisis states, and to “pre-wire” things so that the private sector—rather than the government—is forced to do the recapitalization. In other words, if the fundamental market failure is insufficiently aggressive recapitalization during crises, then regulation should seek to speed up the process of private-sector recapitalization. This is distinct from both: i) the

08 Book.indb 451

2/13/09 3:58:54 PM

452

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

government being directly involved in recapitalization via transfers; ii) requiring private firms to hold more capital ex ante. V.

A Specific Proposal: Capital Insurance

V.A. The basic idea As an illustration of some of our general principles and building on the logic we have developed throughout the paper, we now offer a specific proposal. The basic idea is to have banks buy capital insurance policies that would pay off in states of the world when the overall banking sector is in sufficiently bad shape.21 In other words, these policies would be set up so as to transfer more capital onto the balance sheets of banking firms in those states when aggregate bank capital is, from a social point of view, particularly scarce. Before saying anything further about this proposal, we want to make it clear that it is only meant to be one element in what we anticipate will be a broader reform of capital regulation in the coming years. For example, the scope of capital regulation is likely to be expanded to include investment banks. And it may well make sense to control liquidity ratios more carefully going forward—i.e., to require, for example, banks’ ratio of short-term borrowings to total liabilities not to exceed some target level (though clearly, any new rules of this sort will be subject to the kind of concerns we have raised about higher capital requirements). Our insurance proposal is in no way intended to be a substitute for these other reforms. Instead, we see it as a complement—as a way to give an extra degree of flexibility to the system so that the overall costs of capital regulation are less burdensome. More specifically, we envision that capital insurance would be implemented on an opt-in basis in conjunction with other reforms as follows. A bank with $500 billion in risk-weighted assets could be given the following choice by regulators: It could either accept an upfront capital requirement that is, say, 2% higher, meaning that the bank would have to raise $10 billion in new equity. Or it could acquire an insurance policy that pays off $10 billion upon the occurrence of a systemic “event”—defined perhaps as a situation in which

08 Book.indb 452

2/13/09 3:58:54 PM

Rethinking Capital Regulation

453

the aggregate write-offs of major financial institutions in a given period exceed some trigger level. To make the policy default-proof, the insurer (we have in mind a pension fund or a sovereign wealth fund) would at inception put $10 billion in Treasuries into a custodial account, i.e., a “lock box.” If there is no event over the life of the policy, the $10 billion would be returned to the insurer, who would also receive the insurance premium from the bank as well as the interest paid by the Treasuries. If there is an event, the $10 billion would transfer to the balance sheet of the insured bank. Thus from the perspective of the insurer, the policy would resemble an investment in a defaultable “catastrophe” bond. V.B. The economic logic This proposal obviously raises a number of issues of design and implementation, and we will attempt to address some of these momentarily. Before doing so, however, let us describe the underlying economic logic. One way to motivate our insurance idea is as a form of “recapitalization requirement.” As discussed above, the central market failure is that, in a crisis, individual financial institutions are prone to do too much liquidation and too little new capital raising relative to the social optimum. In principle, this externality could be addressed by having the government inject capital into the banking sector, but this is clearly problematic along a number of dimensions. The insurance approach that we advocate can be thought of as a mechanism for committing the private sector to come up with the fresh capital injection on its own, without resorting to government transfers. An important question is how this differs from simply imposing a higher capital requirement ex ante—albeit one that might be relaxed at the time of a crisis. In the context of the example above, one might ask: What is the difference between asking a pension fund to invest $10 billion in what amounts to a catastrophe bond, versus asking it to invest $10 billion in the bank’s equity, so that the bank can satisfy an increased regulatory capital requirement? Either way, the pension fund has put $10 billion of its money at risk, and either way, the

08 Book.indb 453

2/13/09 3:58:54 PM

454

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

bank will have access to $10 billion more in the event of an adverse shock that triggers the insurance policy. The key distinction has to do with the state-contingent nature of the insurance policy. In the case of the straight equity issue, the $10 billion goes directly onto the bank’s balance sheet right away, giving the bank full access to these funds immediately, independent of how the financial sector subsequently performs. In a world where banks are prone to governance problems, the bank will have to pay a costof-capital premium for the unconditional discretion that additional capital brings.22 By contrast, with the insurance policy, the $10 billion goes into a custodial account. It is only taken out of the account, and made available to the bank, in a crisis state. And crucially, in such states, the bank’s marginal investments are much more likely to be value-creating, especially when evaluated from a social perspective. In particular, a bank that has an extra $10 billion available in a crisis will be able to get by with less in the way of socially-costly asset liquidations.23 This line of argument is an application of a general principle of corporate risk management, developed in Froot, Scharfstein and Stein (1993). A firm can in principle always manage risk via a simple non-contingent “war chest” strategy of having a less leveraged capital structure and more cash on hand. But this is typically not as efficient as a state-contingent strategy that also uses insurance and/or derivatives to more precisely align resources with investment opportunities on a state-by-state basis, so that, to the extent possible, the firm never has “excess” capital at any point in time. In emphasizing the importance of a state-contingent mechanism, we share a key common element with Flannery’s (2005) proposal for banks to use reverse-convertible securities in their capital structure.24 However, we differ substantially from Flannery on a number of specific design issues. We sketch some of the salient features of our proposal below, acknowledging that many details will have to be filled in after more analysis.

08 Book.indb 454

2/13/09 3:58:54 PM

Rethinking Capital Regulation

V.C.

455

Design

We first review some basic logistical issues and then offer an example to illustrate how capital insurance might work. Who participates? Capital insurance is primarily intended for entities that are big enough to inflict systemic externalities during a crisis. It may, however, be unwise for regulatory authorities to identify ahead of time those whom they deem to be of systemic importance. Moreover, even smaller banks could contribute to the credit-crunch and the fire-sale externalities. Thus we recommend that any entity facing capital requirements be given the option to satisfy some fraction of the requirement using insurance. Suppliers Although the natural providers of capital insurance may include institutions such as pension funds and sovereign wealth funds, the securitized design we propose means that policies can be supplied by any investor who is willing to receive a higher-than-risk-free return in exchange for a small probability of a large loss.25 The experience of the last several years suggests that such a risk profile can be attractive to a range of investors. While the market should be allowed to develop freely, one category of investor should be excluded, namely those that are themselves subject to capital requirements. It makes no sense for banks to simultaneously purchase protection with capital insurance, only to suffer losses from writing similar policies. Of course, banks should be allowed to design and broker such insurance so long as they do not take positions. Trigger The trigger for capital insurance to start paying out should be based on losses that affect aggregate bank capital (where the term “bank” should be understood to mean any institution facing capital requirements). In this regard, a key question is the level of geographic aggregation. There are two concerns here. First, banks could suffer

08 Book.indb 455

2/13/09 3:58:54 PM

456

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

losses in one country and withdraw from another.26 Second, international banks may have some leeway in transferring operations to unregulated territories.27 These considerations suggest two design features: First, each major country or region should have its own contingent capital regime meeting uniform international standards so that if, say, losses in the U.S. are severe, multinational banks with significant operations in the U.S. do not spread the pain to other countries. Second, multinational banks should satisfy their primary regulator that a significant proportion of their global operations (say 90 percent) are covered by capital insurance. With these provisos, the trigger for capital insurance could be that the sum of losses of covered entities in the domestic economy (which would include domestic banks and local operations of foreign banks) exceeds some significant amount. To avoid concerns of manipulation, especially in the case of large banks, the insurance trigger for a specific bank should be based on losses of all other banks except the covered bank. The trigger should be based on aggregate bank losses over a certain number of quarters.28 This horizon needs to be long enough for substantial losses to emerge, but short enough to reflect a relatively sudden deterioration in performance, rather than a long, slow downturn. In our example below, we consider a four-quarter benchmark, which means that if there were two periods of large losses that were separated by more than a year, the insurance might not be triggered. An alternative to basing the trigger on aggregate bank losses would be to base it on an index of bank stock prices, in which case the insurance policy would be no more than a put option on a basket of banking stocks. However, this alternative raises a number of further complications. For example, with so many global institutions, creating the appropriate country-level options would be difficult, since there are no share prices for many of their local subsidiaries. Perhaps more importantly, the endogenous nature of stock prices—the fact that stock prices would depend on insurance payouts and vice-versa—could create various problems with indeterminacy or multiple

08 Book.indb 456

2/13/09 3:58:55 PM

Rethinking Capital Regulation

457

equilibria. For these reasons, it is better to link insurance payouts to a more exogenous measure of aggregate bank health. Payout profile A structure that offers large discrete payouts when a threshold level of losses is hit might create incentives for insured banks to artificially inflate their reported losses when they find themselves near the threshold. To deter such behavior, the payout on a policy should increase continuously in aggregate losses once the threshold is reached. Below, we give a concrete example of a policy with this kind of payout profile. Staggered maturities An important question is how long a term the insurance policies would run for. Clearly, the longer the term, the harder it would be to price a policy and the more unanticipated risk the insurer would be subject to, while the shorter the term, the higher the transactions costs of repeated renewal. Perhaps a five-year term might be a reasonable compromise. However, with any finite term length, there is the issue of renewal under stress: What if a policy is expiring at a time when large losses are anticipated, but have not yet been realized? In this case, the bank will find it difficult to renew the policy on attractive terms. To partially mitigate this problem, it may be helpful for each bank to have in place a set of policies with staggered maturities, so that each year only a fraction of the insurance needs to be replaced. Another point to note is that if renewal ever becomes prohibitively expensive, there is always the option to switch back to raising capital in a conventional manner, i.e., via equity issues. An example To illustrate these ideas, Table 1 provides a detailed example of how the proposal might work for a bank seeking $10 billion in capital insurance. We assume that protection is purchased via five policies of $2 billion each that expire at year end for each of the next five

08 Book.indb 457

2/13/09 3:58:55 PM

458

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

Table 1 Hypothetical Capital Insurance Payout Structure In this example, Bank X purchases $10 billion in total coverage. It does so by buying five policies of $2 billion each, with expiration dates of 12/31/2009, 12/31/2010, 12/31/2011, 12/31/2012, and 12/31/2013. The payout on each policy is given by: Payout=

4 quarter loss - max (high watert-1 , trigger) * (Policy face) if 4 quarter loss > high water loss t−1 Full Payout - trigger =0 otherwise

The trigger on each policy is $100 billion in aggregate losses for all banks other than X, and full payout is reached when losses by all banks other than X reach $200 billion. Dollars (billions) 2008Q4

2009Q1

2009Q2

2009Q3

2009Q4

Current quarter loss

50

40

20

0

140

Cumulative 4 quarter loss

80

120

140

110

200

High water mark on losses

80

120

140

140

200

Payout per policy

0

0.4

0.4

0

1.2

Payout total

0

2

2

0

6

Cumulative payout

0

2

4

4

10

years. There are three factors that shape the payouts on the policies: the trigger points for both the initiation of payouts and the capping of payouts, the pattern of bank losses, and the function that governs how losses are translated into payouts. In the example, the trigger for initiating payouts is hit once cumulative bank losses over the last four quarters reach $100 billion. And payouts are capped once cumulative losses reach $200 billion. In between, payouts are linear in cumulative losses. This helps to ensure that, aside from the time value of earlier payments, banks have no collective benefit to pulling forward large loss announcements. The payout function also embeds a “high-water” test, so that—given the four-quarter rolling window for computing losses—only incremental losses in a given quarter lead to further payouts. In the example, this feature comes into play in the third quarter of 2009, when current losses are zero. Because of the high-water feature, payouts in this quarter are zero also, even though cumulative losses over the prior four

08 Book.indb 458

2/13/09 3:58:55 PM

Rethinking Capital Regulation

459

quarters continue to be high. Put simply, the high-water feature allows us to base payouts on a four-quarter window, while at the same time avoiding double-counting of losses. These and other details of contract design are important, and we offer the example simply as a starting point for further discussion. However, given that the purpose of the insurance is to guarantee relatively rapid recapitalization of the banking sector, one property of the example that we believe should carry over to any real-world structure is that it be made to pay off promptly. V.D. Comparisons with alternatives An important precursor to our proposal, and indeed the starting point for our thinking on this, is Flannery (2005). Flannery proposes that banks issue reverse convertible debentures, which convert to equity when a bank’s share price falls below a threshold. Such an instrument can be thought of as a type of firm-specific capital insurance. One benefit of a firm-specific trigger is that it provides the bank with additional capital in any state of the world when it is in trouble—unlike our proposal where a bank gets an insurance payout only when the system as a whole is severely stressed. In the spirit of the traditional approach to capital regulation, the firm-specific approach does a more complete job of reducing the probability of distress for each individual institution. The firm-specific trigger also should create monitoring incentives for the bond holders, which could be useful. Finally, to the extent that one firm’s failure could be systemically relevant, this proposal resolves that problem, whereas ours does not. However, a firm-specific trigger also has disadvantages. First, given that a reverse convertible effectively provides a bank with debt forgiveness if it performs poorly enough, it could exacerbate problems of governance and moral hazard. Moreover, the fact that the trigger is based on the bank’s stock price may be particularly problematic here. One can imagine that once a bank begins to get into trouble, there may be the ingredients in place for a self-fulfilling downwards spiral: As existing shareholders anticipate having their stakes diluted via the

08 Book.indb 459

2/13/09 3:58:55 PM

460

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

conversion of the debentures, stock prices decline further, making the prospect of conversion even more likely, and so on.29 Our capital insurance structure arguably does better than reverse convertibles on bank-specific moral hazard, given that payouts are triggered by aggregate losses rather by poor individual performance. With capital insurance, not only is a bank not rewarded for doing badly, it gets a payout in precisely those states of the world when access to capital is most valuable, i.e., when assets are cheap and profitable lending opportunities abound. Therefore, banks’ incentives to preserve their own profits are unlikely to diminished by capital insurance. Finally, ownership of the banking system brings with it important political-economy considerations. Regulators may be unwilling to allow certain investors to accumulate large control stakes in a banking firm. To the extent that holders of reverse convertibles get a significant equity stake upon conversion, regulators may want to restrict investment in these securities to those who are fit and proper, or alternatively, remove their voting rights. Either choice would further limit the attractiveness of the reverse convertible. By contrast, our proposal does not raise any knotty ownership issues: When the trigger is hit, the insured bank simply gets a cash payout with no change in the existing structure of shareholdings. The important common element of the Flannery (2005) proposal and ours is the contingent nature of the financing. There are other contingent schemes that could also be considered; Culp (2002) offers an introductory overview of these types of securities and a description of some that have been issued. Security design could take care of a variety of concerns. For example, if investors do not like the possibility of losing everything on rare occasions, the insurance policies could be over-collateralized: The insurer would put $10 billion into the lock box, but only a maximum of $5 billion could be transferred to the insured policy in the event the trigger is breached. This is a transparent change that might get around problems arising because some buyers (such as pension funds or insurance companies) face restrictions on buying securities with low ratings.

08 Book.indb 460

2/13/09 3:58:55 PM

Rethinking Capital Regulation

461

A security that has some features of Flannery’s proposal (it is tied to firm-specific events) and some of ours (it is tied to losses, not stock prices) is the hybrid security issued in 2000 by the Royal Bank of Canada (RBC). RBC sold a privately placed bond to Swiss RE that, upon a trigger event, converted into preferred shares with a given dividend yield. The conversion price was negotiated at date of the bond issue, and the trigger for conversion was tied to a large drop in RBC’s general reserves. The size of the issue (C$200 million) was set to deliver an equity infusion of roughly one percent of RBC’s tier capital requirement. Of particular interest is the rationale RBC had for this transaction. Culp (2002, p. 51) quotes RBC executive David McKay as follows: “It costs the same to fund your reserves whether they’re geared for the first amount of credit loss or the last amount of loss… What is different is the probability of using the first loss amounts versus the last loss amounts. Keeping capital on the balance sheet for a last loss amount is not very efficient.” The fact that this firm-specific security could be priced and sold suggests the industry-linked one that we are proposing need not present insurmountable practical difficulties. Before concluding, let us turn to a final concern about our insurance proposal that it might create the potential for a different kind of moral hazard. Even though banks do not get reimbursed for their own losses, the fact that they get a cash infusion in a crisis might reduce their incentives to hedge against the crisis, to the extent that they are concerned about not only expected returns, but also the overall variance of their portfolios. In other words, banks might negate some of the benefits of the insurance by taking on more systematic risk. To see the logic most transparently, consider a simple case where a bank sets a fixed target on the net amount of money it is willing to lose in the bad state (i.e., it implements a value-at-risk criterion). If it knows that it will receive a $10 billion payoff from an insurance policy in the crisis, it may be willing to tolerate $10 billion more of pre-insurance losses in the crisis. If all banks behave in this way, they may wind up with more highly correlated portfolios than they would absent capital insurance.

08 Book.indb 461

2/13/09 3:58:55 PM

462

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

This concern is clearly an important one.  However, there are a couple of potentially mitigating factors. First, what is relevant is not whether our insurance proposal creates any moral hazard, but whether it creates more or less than the alternative of raising capital requirements.  One could equally well argue that, in an effort to attain a desired level of return on equity, banks target the amount of systematic risk borne by their stockholders, i.e., their equity betas. If so, when the capital requirement is raised, banks would offset this by simply raising the systematic risk of their asset portfolios, so as to keep constant the amount of systematic risk borne per unit of equity capital. In this sense, any form of capital regulation faces a similar problem. Second, the magnitude of the moral hazard problem associated with capital insurance is likely to depend on how the trigger is set, i.e. on the likelihood that the policy will pay off. Suppose that the policy only pays off in an extremely bad state which occurs with very low probability a true financial crisis. Then a bank that sets out to take advantage of the system by holding more highly correlated assets faces a tradeoff: This strategy makes sense to the extent that the crisis state occurs and the insurance is triggered, but will be regretted in the much more likely scenario that things go badly, but not sufficiently badly to trigger a payout. This logic suggests that with an intelligently designed trigger, the magnitude of the moral hazard problem need not be prohibitively large. This latter point is reinforced by the observation that, because of the agency and performance-measurement problems described above, bank managers likely underweight very low probability tail events when making portfolio decisions. On the one hand, this means that they do not take sufficient care to avoid assets that have disastrous returns with very low probability, hence the current crisis. At the same time, it also means that they do not go out of their way to target any specific pattern of cashflows in such crisis states. Rather, they effectively just ignore the potential for such states ex ante and focus on optimizing their portfolios over the more normal parts of the distribution.  If this is the case, insurance with a sufficiently low-probability trigger will not have as much of an adverse effect on behavior.

08 Book.indb 462

2/13/09 3:58:55 PM

Rethinking Capital Regulation

VI.

463

Conclusions

Our analysis of the current crisis suggests that governance problems in banks and excessive short-term leverage were at its core. These two causes are related. Any attempt at preventing a recurrence should recognize that it is difficult to resolve governance problems, and, consequently, to wean banks from leverage. Direct regulatory interventions, such as mandating more capital, could simply exacerbate private sector attempts to get around them, as well as chill intermediation and economic growth. At the same time, it is extremely costly for society to either continue rescuing the banking system or to leave the economy to be dragged into the messes that banking crises create. If despite their best efforts, regulators cannot prevent systemic problems, they should focus on minimizing their costs to society without dampening financial intermediation in the process. We have offered one specific proposal, capital insurance, which aims to reduce the adverse consequences of a crisis, while making sure the private sector picks up the bill. While we have sketched the broad outlines of how a capital insurance scheme might work, there is undoubtedly much more work to be done before it can be implemented. We hope that other academics, policymakers and practitioners will take up this challenge.

Authors’ note: We thank Alan Boyce, Chris Culp, Doug Diamond, Martin Feldstein, Benjamin Friedman, Kiyohiko Nishimura, Eric Rosengren, Hyun Shin, Andrei Shleifer and Tom Skwarek for helpful conversations. We also thank Olivier Blanchard, Steve Cecchetti, Darrell Duffie, Bill English, Jean-Charles Rochet, Larry Summers, Paul Tucker and seminar participants at the Bank of Canada, NBER Summer Institute, the Chicago GSB Micro Lunch, the University of Michigan, the Reserve Bank of Australia, and the Australian Prudential Regulatory Authority for valuable comments. Yian Liu provided expert research assistance. Kashyap and Rajan thank the Center for Research on Security Prices and the Initiative on Global Markets for research support. Rajan also acknowledges support from the NSF. All mistakes are our own.

08 Book.indb 463

2/13/09 3:58:55 PM

464

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

Endnotes See Bank for International Settlements (2008, chapter 6), Bank of England (2008), Bernanke (2008), Borio (2008), Brunnermeier (2008), Dudley (2007, 2008), Greenlaw et al (2008), IMF (2008), and Knight (2008) for comprehensive descriptions of the crisis. 1

Throughout this paper, we use the word “bank” to refer to both commercial and investment banks. We say “commercial bank” when we refer to only the former. 2

3 See Brunnermeier and Pedersen (2008) for a detailed analysis of these kinds of spirals and Adrian and Shin (2008b) for empirical evidence on the spillovers.

The state-contingent nature of such an insurance scheme makes it similar in some ways to Flannery’s (2005) proposal for the use of reverse convertible securities in banks’ capital structures. We discuss the relationship between the two ideas in more detail below. 4

See Hoenig (2008) and Rajan (2005) for a similar diagnosis.

5

Financial Times, July 9, 2007.

6

Shareholder Report on UBS Writedowns, April 18, 2008, http://www.ubs.com/1/e/ investors/agm.html. 7

8 Another example of the effects of uncharged risk is described in the Shareholder Report on UBS Writedowns on page 13: “The CDO desk received structuring fees on the notional value of the deal, and focused on Mezzanine (“Mezz”) CDOs, which generated fees of approximately 125 to 150 bp (compared with high-grade CDOs, which generated fees of approximately 30 to 50 bp).” The greater fee income from originating riskier, lower quality mortgages fed directly to the originating unit’s bottom line, even though this fee income was, in part, compensation for the greater risk that UBS would be stuck with unsold securities in the event that market conditions turned.

As the Wall Street Journal (April 16, 2008) reports, “Risk controls at [Merrill Lynch], then run by CEO Stan O’Neal, were beginning to loosen. A senior risk manager, John Breit, was ignored when he objected to certain risks…Merrill lowered the status of Mr. Breit’s job...Some managers seen as impediments to the mortgage-securities strategy were pushed out. An example, some former Merrill executives say, is Jeffrey Kronthal, who had imposed informal limits on the amount of CDO exposure the firm could keep on its books ($3 billion to $4 billion) and on its risk of possible CDO losses (about $75 million a day). Merrill dismissed him and two other bond managers in mid-2006, a time when housing was still strong but was peaking. To oversee the job of taking CDOs onto Merrill’s own books, the firm tapped …a senior trader but one without much experience in mortgage securities. CDO holdings on Merrill’s books were soon piling up at a rate of $5 billion to $6 billion per quarter.” Bloomberg (July 22, 2008, “Lehman Fault-Finding 9

08 Book.indb 464

2/13/09 3:58:55 PM

Rethinking Capital Regulation

465

Points to Last Man Fuld as Shares Languish”) reports a similar pattern at Lehman Brothers whereby “at least two executives who urged caution were pushed aside.” The story quotes Walter Gerasimowicz, who worked at Lehman from 1995 to 2003, as saying “Lehman at one time had very good risk management in place. They strayed in search of incremental profit and market share.” 10 The insight that agency problems lead banks to be highly levered goes back to Diamond’s (1984) classic paper.

By analogy, it appears that the equity market penalizes too much financial slack in operating firms with poor governance. For example, Dittmar and Mahrt-Smith (2007) estimate that $1.00 of cash holdings in a poorly-governed firm is only valued by the market at between $0.42 and $0.88. 11

A more subtle argument is that the fragile nature of short-term debt financing is actually part of its appeal to banks: Precisely because it amplifies the negative consequences of mismanagement, short-term debt acts as a valuable ex ante commitment mechanism for banks. See Calomiris and Kahn (1991). However, when thinking about capital regulation, the critical issue is whether short-term debt has some social costs that are not fully internalized by individual banks. 12

In a Basel II regime, the pressure to liquidate assets is intensified in crisis periods because measured risk levels—and hence risk-weighted capital requirements—go up. One can get a sense of magnitudes from investment banks, who disclose firm-wide “value at risk” (VaR) numbers. Greenlaw et al (2008) calculate a simple average of the reported VaR for Morgan Stanley, Goldman Sachs, Lehman Brothers and Bear Stearns, and find that it rose 34% between August 2007 and February 2008. 13

For instance, Bernanke (2008) says: “I strongly urge financial institutions to remain proactive in their capital-raising efforts. Doing so not only helps the broader economy but positions firms to take advantage of new profit opportunities as conditions in the financial markets and the economy improve.” 14

15 Kashyap and Stein (2004) point out that the Basel II approach can be thought of as reflecting the preferences of a social planner who cares only about avoiding bank defaults, and who attaches no weight to other considerations, such as the volume of credit creation.

See Adrian and Shin (2008a) for systematic evidence on this phenomenon.

16

Subprime mortgage originations seemed to take off to supply this market. For instance, Greenlaw et al show that subprime plus Alt-A loans combine represented fewer than 10% of all mortgage originations in 2001, 2002 and 2003, but then jumped to 24% in 2004 and further to 33% in 2005 and 2006; by the end of 2007 they were back to 9%. As Mian and Sufi (2008) and Keys et al (2008) suggest, the quality of underlying mortgages deteriorated considerably with increased demand for mortgagedbacked securities. See European Central Bank (2008) for a detailed description of the role of structured finance products in propagating the initial subprime shock. 17

08 Book.indb 465

2/13/09 3:58:55 PM

466

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

It should be noted, however, that higher ex ante capital requirements do have one potentially important benefit. If a bank starts out with a high level of capital, it will find it easier to recapitalize once a shock hits, because the lower is its postshock leverage ratio, the less of a debt overhang problem it faces, and hence the easier it is issue more equity. Hence the bank will do more recapitalization, and less liquidation, which is a good thing. 18

19 See Tucker (2008) for further thoughts on this. For instance, capital standards could also be progressively increased during a boom to discourage risk-taking.

Starting in 2000 Spain has run a system based on “dynamic provisioning” whereby provisions are built up during times of low reported losses that are to be applied when losses rise. According to Fernández-Ordóñez (2008), Spanish banks “had sound loan loss provisions (1.3% of total assets at the end of 2007, and this despite bad loans being at historically low levels).” In 2008 the Spanish economy has slowed, and loan losses are expected to rise, so time will tell whether this policy changes credit dynamics. 20

Our proposal is similar in the spirit to Caballero’s (2001) contingent insurance plan for emerging market economies. 21

There may be a related cosmetic benefit of the insurance policy. Since the bank takes less equity onto its balance sheet, it has fewer shares outstanding, and various measures of performance, such as earnings per share and return on equity, may be less adversely impacted than by an increase in the ex ante capital requirement. Of course, this will also depend on how the bank is allowed to amortize the cost of the policy. 22

To illustrate, suppose a bank has 100 in book value of loans today; these will yield a payoff of either 90 or 110 next period, with a probability ½ of either outcome. One way for the bank to insure against default would be to finance itself with 90 of debt and 10 of equity. But this approach leaves the bank with 20 of free cash in the good state. If investors worry that this cash in good times will lead to mismanagement and waste, they will discount the bank’s stock. Now suppose instead that the bank seeks contingent capital. It could raise 105, with 100 of this in debt and 5 in equity, and use the extra 5 to finance, in addition to the 100 of loans, the purchase of an insurance policy that pays off 10 only in the bad state. From a regulator’s perspective, the bank should be viewed as just as well-capitalized as before, since it is still guaranteed not to default in either state. At the same time, the agency problem is attenuated, because after paying off its debt, the bank now has less cash to be squandered in the good state (10, rather than 20). 23

24 See also Stein (2004) for a discussion of state-contingent securities in a banking context.

There may be some benefit to having the insurance provided by passive investors. Not only do they have pools of assets that are idle and can profitably serve as collateral (in contrast to an insurance company that might be reluctant to see 25

08 Book.indb 466

2/13/09 3:58:55 PM

Rethinking Capital Regulation

467

its assets tied up in a lock box), they also have the capacity to bear losses without attempting to hedge them (again, unlike a more active financial institution). Individual investors, pension funds, and sovereign wealth funds would be important providers. See Organization for Economic Cooperation and Development (2008) for a list of major investments, totaling over $40 billion, made by sovereign wealth funds in the financial sector from 2007 through early 2008. 26 Indeed, Peek and Rosengren (2000) document the withdrawal of Japanese banks from lending in California in response to severe losses in Japan.

The trigger might also be stated in terms of the size of the domestic market so that firms entering a market do not mechanically change the likelihood of a payment. 27

Because this insurance pays off only in systemically bad states of nature, it will be expensive, but not relative to pure equity financing. For example, suppose that there are 100 different future states of the world for each bank and that the trigger is breached only in 1 of the 100 scenarios. Because equity returns are low both in the trigger state and in many others (with either poor bank-specific outcomes or bad but not disastrous aggregate outcomes), the cost of equity must be higher than the cost of the insurance. 28

Relatedly, such structures can create incentives for speculators to manipulate bank stock prices. For example, it may pay for a large trader to take a long position in reverse convertibles, then try to push down the price of the stock via short-selling in order to force conversion and thereby acquire an equity stake on favorable terms. 29

08 Book.indb 467

2/13/09 3:58:55 PM

468

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

References Adrian, Tobias, and Hyun Song Shin, (2008a), Liquidity, Financial Cycles and Monetary Policy, Current Issues in Economics and Finance, Federal Reserve Bank of New York, 14(1). Adrian, Tobias, and Hyun Shin, (2008b), Financial Intermediaries, Financial Stability and Monetary Policy, paper prepared for Federal Reserve Bank of Kansas City Symposium on Maintaining Stability in a Changing Financial System, Jackson Hole, Wyoming, August 21-23, 2008. Allen, Franklin, and Douglas Gale, (2005), From Cash-in-the-Market Pricing to Financial Fragility, Journal of the European Economic Association 3, 535-546. Bank for International Settlements, (2008), 78th Annual Report: 1 April 2007 31 March 2008, Basel, Switzerland. Bank of England, (2008), Financial Stability Report, April 2008, Issue Number 23, London. Benmelech, Efraim, and Jennifer Dlugosz, (2008), The Alchemy of CDO Credit Ratings, Harvard University working paper. Bernanke, Ben S., (2008), Risk Management in Financial Institutions, Speech delivered at the Federal Reserve Bank of Chicago’s Annual Conference on Bank Structure and Competition, Chicago, Illinois, May 15. Borio, Claudio (2008), The Financial Turmoil of 2007-?: A Preliminary Assessment and Some Policy Considerations, BIS working paper no. 251. Brunnermeier, Markus K., (2008), Deciphering the 2007-08 Liquidity and Credit Crunch, Journal of Economic Perspectives, forthcoming. Brunnermeier, Markus K., and Lasse Pedersen, (2008), Market Liquidity and Funding Liquidity, Review of Financial Studies, forthcoming. Caballero, Ricardo J., (2001), Macroeconomic Volatility in Reformed Latin America: Diagnosis and Policy Proposal, Inter-American Development Bank, Washington, D.C., 2001. Calomiris, Charles W., and Charles M. Kahn, (1991), The Role of Demandable Debt in Structuring Optimal Banking Arrangements, American Economic Review 81, 495-513. Coval, Joshua D., Jakub W. Jurek, and Erik Stafford, (2008a), Economic Catastrophe Bonds, American Economic Review, forthcoming. Coval, Joshua D., Jakub W. Jurek, and Erik Stafford, (2008b), Re-Examining The Role of Rating Agencies: Lessons From Structured Finance, Journal of Economic Perspectives, forthcoming.

08 Book.indb 468

2/13/09 3:58:55 PM

Rethinking Capital Regulation

469

Culp, Christopher, L., (2002), Contingent Capital: Integrating Corporate Financing and Risk Management Decisions, Journal of Applied Corporate Finance, 55(1), 46-56. Diamond, Douglas W., (1984), Financial Intermediation and Delegated Monitoring, Review of Economic Studies 51, 393-414. Diamond, Douglas W., and Raghuram G. Rajan, (2005), Liquidity Shortages and Banking Crises, Journal of Finance 60, 615-647. Dittmar, Amy, and Jan Mahrt-Smith, (2007), Corporate Governance and the Value of Cash Holdings, Journal of Financial Economics 83, 599-634. Dudley, William C., (2007), May You Live in Interesting Times, Remarks at the Federal Reserve Bank of Philadelphia, October 17. Dudley, William C., (2008), May You Live in Interesting Times: The Sequel, Remarks at the Federal Reserve Bank of Chicago’s 44th Annual Conference on Bank Structure and Competition, Chicago, Illinois, May 15. European Central Bank, (2008), Financial Stability Review, June 2008, Frankfurt. Fernández-Ordóñez, Miguel, (2008), Remarks at 2008 International Monetary Conference Central Bankers Panel, Barcelona, June 3. Flannery, Mark J., (2005), No Pain, No Gain? Effecting Market Discipline via Reverse Convertible Debentures, Chapter 5 of Hal S. Scott, ed., Capital Adequacy Beyond Basel: Banking Securities and Insurance, Oxford: Oxford University Press. Froot, Kenneth, David S. Scharfstein, and Jeremy C. Stein, (1993), Risk Management: Coordinating Corporate Investment and Financing Policies, Journal of Finance 48, 1629-1658. Greenlaw, David, Jan Hatzius, Anil K Kashyap, and Hyun Song Shin, (2008), Leveraged Losses: Lessons from the Mortgage Market Meltdown, U.S. Monetary Policy Forum Report No. 2, Rosenberg Institute, Brandeis International Business School and Initiative on Global Markets, University of Chicago Graduate School of Business. Gromb, Denis, and Dimitri Vayanos, (2002), Equilibrium and Welfare in Markets With Financially Constrained Arbitrageurs, Journal of Financial Economics 66, 361-407. Hart, Oliver and John Moore, (1998), Default and Renegotiation: A Dynamic Model of Debt, Quarterly Journal of Economics 113, 1-41. Hoenig, Thomas M., (2008), Perspectives on the Recent Financial Market Turmoil, Remarks at the 2008 Institute of International Finance Membership Meeting, Rio de Janeiro, Brazil, March 5. IMF, (2008), Global Financial Stability Report, April, Washington DC.

08 Book.indb 469

2/13/09 3:58:56 PM

470

Anil K. Kashyap, Raghuram G. Rajan, Jeremy C. Stein

Kashyap, Anil K. and Jeremy C. Stein, (2004), Cyclical Implications of the Basel-II Capital Standards, Federal Reserve Bank of Chicago Economic Perspectives 28, 18-31. Kelly, Kate, (2008), Lost Opportunities Haunt Final Days of Bear Stearns: Executives Bickered Over Raising Cash, Cutting Mortgages, Wall Street Journal, A1, May 27. Keys, Benjamin, Tanmoy Mukherjee, Amit Seru, and Vikrant Vig, (2008), Did Securitization Lead to Lax Screening? Evidence from Subprime Loans, Chicago GSB working paper. Knight, Malcolm, (2008), Now You See It, Now You Don’t: The Nature of Risk and the Current Financial Turmoil, speech delivered at the Ninth Annual Risk Management Convention of the Global Association of Risk Professionals, February 26-27. Kyle, Albert S., and Wei Xiong, (2001), Contagion as a Wealth Effect, Journal of Finance 56, 1401-1440. Mian, Atif, and Amir Sufi, (2008), The Consequences of Mortgage Credit Expansion: Evidence from the 2007 Mortgage Default Crisis, Chicago GSB working paper. Morris, Stephen, and Hyun Song Shin, (2004), Liquidity Black Holes, Review of Finance 8, 1-18. Myers, Stewart C., (1977), Determinants of Corporate Borrowing, Journal of Financial Economics 5, 147-175. Myers, Stewart C., and Nicholas S. Majluf, (1984), Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have, Journal of Financial Economics 13, 187-221. Myers, Stewart C., and Raghuram G. Rajan, (1998), The Paradox of Liquidity, Quarterly Journal of Economics 113, 733-771. Organization for Economic Cooperation and Development, (2008), Financial Market Highlights May 2008: The Recent Financial Market Turmoil, Contagion Risks and Policy Responses, Financial Market Trends, No 94, Volume 2008/1 June 2008, Paris. Peek, Joe, and Eric Rosengren (2000), Collateral Damage: Effects of the Japanese Bank Crisis on Real Activity in the United States, American Economic Review 90, 30-45. Rajan, Raghuram G., (1994), Why Bank Credit Policies Fluctuate: A Theory and Some Evidence, Quarterly Journal of Economics 109, 399-442. Rajan, Raghuram G. (2005), Has Financial Development Made the World Riskier? Proceedings of the Jackson Hole Conference organized by the Kansas City Fed.

08 Book.indb 470

2/13/09 3:58:56 PM

Rethinking Capital Regulation

471

Shleifer, Andrei, and Robert W. Vishny, (1992), Liquidation Values and Debt Capacity: A Market Equilibrium Approach, Journal of Finance 47. Shleifer, Andrei, and Robert W. Vishny, (1997), The Limits of Arbitrage, Journal of Finance 52, 35-55. Stein, Jeremy C., (1989), Efficient Capital Markets, Inefficient Firms: A Model of Myopic Corporate Behavior, Quarterly Journal of Economics 104, 655-669. Stein, Jeremy C., (2004), Commentary, Federal Reserve Bank of New York Economic Policy Review, 10, September, pp. 27-29. Tucker, Paul M. W., (2008), Monetary Policy and the Financial System, remarks at the Institutional Money Market Funds Association Annual Dinner, London, April 2, 2008.

08 Book.indb 471

2/13/09 3:58:56 PM

08 Book.indb 472

2/13/09 3:58:56 PM

Commentary: Rethinking Capital Regulation Jean-Charles Rochet

It is a privilege to be here today to discuss this stimulating article of my distinguished colleagues Kashyap, Rajan and Stein, and to participate in this very interesting conference on how to maintain financial stability after the current credit crisis. Many influential commentators1 have advocated for fundamental reforms of financial regulatory/supervisory systems as a necessary response to the crisis. Capital regulations are clearly a crucial element of these systems, and the article by Kashyap, Rajan and Stein offers several important insights and a specific proposal on how to improve these regulations. This article is therefore particularly timely. I will organize my comments in three parts: 1. The objectives of capital regulation. 2. The regulatory treatment of capital insurance. 3. Reorganizing the financial infrastructure. 1. The objectives of capital regulation Capital regulation is a fundamental component of the financial safety net, together with deposit insurance, supervisory intervention, liquidity support by central banks and in some cases capital 473

08 Book.indb 473

2/13/09 3:58:56 PM

474

Jean-Charles Rochet

injections by the Treasury. This financial safety net has officially two objectives: • To protect small depositors against the failure of their bank (microprudential objective), • And to protect the financial system as a whole against aggregate shocks (macro-prudential objective). As pointed out by Kashyap, Rajan and Stein, individual bank failures and systemic crises cannot be eliminated altogether, which raises two questions: • What should be their “optimal” frequency? • How should we manage individual failures and, more importantly, systemic crises when they occur? Existing capital regulations, notably Basel 2, have only offered a relatively precise answer to the first question, at least for individual bank failures. In particular, the IRB approach to credit risk in the pillar one of Basel 2 implies more or less explicitly a quantitative target for the maximum probability of default of commercial banks (0.1% over one year). This focus on the probability of default is consistent with traditional actuarial methods in insurance, with the practice of rating agencies and with the VaR approach to risk management developed by large banks (see also Gordy, 2003). However, I want to suggest that focusing on a exogenously given probability of default is largely arbitrary and has many undesirable consequences. For example, Kashyap and Stein (2003), among others, argue that it would make more sense to implement a flexible approach where the maximum probability of failure would not be constant but would instead vary along the business cycle (concretely, to allow banks to take more risks during recessions and less during booms). This is obviously related to the procyclicality debate. Moreover, the VaR approach can be easily manipulated and has led to many forms of regulatory arbitrage. In particular, it gives incentives for banks to shift their risks towards the upper tails of loss distributions, which increases systemic risk. In fact, VaR measures

08 Book.indb 474

2/13/09 3:58:56 PM

Commentary

475

may be appropriate from the perspective of a bank shareholder (who is protected by limited liability) but certainly not from that of public authorities (who will ultimately bear the costs of extreme losses). From a conceptual viewpoint, capital requirements should be seen as a component of an insurance contract between regulators and banks, whereby banks have access to the financial safety net, provided they satisfy certain conditions. The capital of the bank can be interpreted as the “deductible” in this insurance contract, namely the size of the first tranche of losses, that will be entirely borne by shareholders. The failure of the bank occurs exactly when incurred losses exceed this amount. In property casualty insurance, the level of deductibles on an insurance contract is not determined by a hypothetical target probability of claims (here bank failures), but instead by a trade-off between the expected cost of these claims (including transaction costs), the cost of self-financing the deductible (here the cost of equity for banks) and the benefit of insurance for customers that includes being able to increase the level of their risky activities (here the volume of lending). By analogy, the capital requirement (CR) for banks should not be computed as a “VaR” but as an expected shortfall (or Tail VaR), which takes fully into account the tail distribution of losses, and thus does not give perverse incentives to shift risks to the upper tail of the loss distributions. Moreover, this “economic” approach to CR is much more flexible than the dominant “actuarial” approach. As in the case of insurance (see Plantin and Rochet, 2008), optimal CRs can in this way be determined by trading off the social cost of bank failures against the social benefit of bank lending, which are both likely to vary across the business cycle. They can also incorporate incentive considerations, on which I will comment below. 2. The regulatory treatment of capital insurance As shown by Kashyap, Rajan and Stein (2008), the macro-prudential component of financial regulation is not sufficiently taken into account in existing capital regulations. They rightly point at the aggregate effects of the behavior of banks (especially large ones) during

08 Book.indb 475

2/13/09 3:58:56 PM

476

Jean-Charles Rochet

crises. When these large banks face binding solvency constraints, they tend to react by reducing too much (from a social welfare perspective) their volume of assets (lending less and selling securities, even at a depressed price), rather than by issuing the amount of new equity that would allow them to keep the same volume of assets. This is because banks do not internalize the negative impact of their fire sales on the prices of these assets, which may itself force other banks to liquidate some of their assets, provoking a credit crunch and a downward spiral for asset prices (Brunnermeier, 2008; Adrian and Shin, 2008). Kashyap, Rajan and Stein (2008) put forward a specific proposal for improving capital regulation: encouraging banks (on a voluntary basis) to purchase capital insurance contracts that would pay off in states of the world where the overall banking system is in bad shape. The idea behind this proposal is that whereas banks’ preferred form of financing during tranquil times is short-term debt (because it is a better disciplining tool than equity or long-term debt, given the complexity of banking activities), equity capital becomes too scarce during recessions and banking crises. Banks tend to respond to these negative shocks by reducing the size of their balance sheets rather than by issuing new equity, both because investors are reluctant to provide it during stress periods and because banks do not internalize the negative impact on the economy. In the capital insurance contracts proposed by Kashyap, Rajan and Stein, the insurer would commit to provide a given amount of cash when some aggregate measure of banks’ performance falls below a pre-specified threshold. Banks would be less inclined to sell assets, and the need for public authorities to step in would be reduced. This proposal (which resembles an earlier proposal put forward by Flannery, 2005) is a particular form of the new Alternative Risk Transfer (ART) methods that provide hybrid instruments (with both insurance and financing components) to large firms, not exclusively in the financial sector. These ART instruments (such as contingent capital, catastrophe bonds and options) have been promoted by several re-insurers (notably Swiss Re) but have not so far been used extensively in practice.

08 Book.indb 476

2/13/09 3:58:56 PM

Commentary

477

The proposal of Kashyap, Rajan and Stein is a good idea, but several questions have to be answered more precisely. For example, isn’t it too demanding to impose that the insurer post a 100% collateral deposit in a custodial account, considering that the probability of a claim is (hopefully) very small and the duration of the contract presumably quite long? On the other hand, how can regulators guarantee that the insurer will always fulfill its obligations, unless the insurer’s capital itself is also regulated? Also, the pricing of these capital insurance contracts is likely to be difficult, given that claims will have a low probability of occurrence, but will occur exactly when the overall economic situation is very bad. Finally, the authors should clarify whether they think the main reason why banks do not issue more capital during crises is that they cannot or that they do not want to. In the first case, capital insurance contracts make a lot of sense, but then why is it that the banks themselves have not already come up with the idea? In the second case (i.e. if banks do not want to issue more capital during crises), capital insurance can still be good from a regulatory perspective (if not from a private perspective), but regulators have to be given the power to prevent the banks from distributing dividends with the money collected from the capital insurance contract. I would like to put forward a similar proposal, inspired by Holmström and Tirole (1998), which could be viewed as a complement to the capital insurance proposal of Kashyap, Rajan and Stein. Suppose indeed that the Treasury issues a new type of security, namely a contingent bond that would pay off only conditionally on some trigger (that could be related to aggregate bank losses like in the proposal of Kashyap, Rajan and Stein, or more generally to other indicators of macroeconomic stress). The insurance properties of this security would be exactly the same as the one suggested by Kashyap, Rajan and Stein, but it would be provided by the Treasury and not by private investors such as sovereign funds or pension funds. The advantages would be that the solvency of the issuer would not have to be monitored and that liquidity would only be issued ex post (in the states of the world where it is needed) and would not be “wasted” in the states of the world where it is not needed. The superiority of the government over the market in providing ex-post liquidity comes from its unique ability to tax households and firms in the future.

08 Book.indb 477

2/13/09 3:58:56 PM

478

Jean-Charles Rochet

Let me address now the questions of incentives. There seems to be a consensus that agency problems have been prevalent at all stages of the securitization process. A recent study by Ashcraft and Schuermann (2008) gives a splendid illustration of this prevalence. An important empirical question is whether capital requirements can be really efficient for aligning incentives between bank managers and public authorities. Kashyap, Rajan and Stein argue that short-term finance may be a better tool for disciplining bankers, essentially because banks are too complex entities to be monitored by shareholders. They observe that even if managers have very large stakes in their banks, they are inclined to take huge risks. This may explain why equity financing is so expensive for banks. I believe this view is more appropriate for investment banks rather than commercial banks. In fact, since the implementation of Basel 1, commercial banks have traditionally held way more equity than the regulatory minimum, in response to market discipline. This seems to suggest that financial analysts and rating agencies consider that commercial banks need a sufficient amount of equity capital, above regulatory minimums. In fact, economic capital for a well-managed bank is often evaluated to a given multiple of regulatory capital. Therefore, regulation has to be designed in such a way that banks can save on their minimum capital charges (and thus on their economic capital, which allows them to increase return on equity) when they make investment decisions that are socially beneficial. More generally, if ones believes that capital regulation may have a sizable impact on bankers’ incentives, it is particularly important to design capital charges for securitization and other credit risk transfer operations in such a way that they align the incentives of bank shareholders with the regulator’s objective: encourage the transfer of “exogenous” risks (those that are not under the control of the bankers), limit the transfer of “endogenous” risks (the risks that are partially affected by bankers’ actions) to the maximum amount that preserves incentives. The current implications of securitization in terms of regulatory capital requirements (especially Basel 2) do not necessarily encourage banks to adopt this strategy.

08 Book.indb 478

2/13/09 3:58:56 PM

Commentary

479

3. Reorganizing the financial infrastructure As was clearly advocated by Tim Geithner, the president and CEO of the New York Fed, in a recent article (Financial Times, June 8, 2008), the important changes in the industrial organization of the financial industry that have been observed in the last decade make it necessary to “adapt the regulatory system to address the vulnerabilities exposed by the financial crisis.” In particular, he argues that “supervision has to ensure that counterparty risk management in the supervised institutions limits the risk of a rise in overall leverage outside the regulated institutions that could threaten the stability of the whole financial system.” The guiding principle here should be the absence of a “regulatory free lunch”: If investment banks want to have access to the liquidity provision facilities put in place by central banks, they should be required to satisfy more stringent conditions in terms of capital, liquidity and risk management. Similarly, if supervised institutions want to benefit from reductions in capital charges when they use new, complex credit risk transfer instruments, they should accept a certain degree of standardization and centralization in the issuance, clearing and settlement of these instruments. The management of systemic risk is obviously easier at the level of a central platform (exchange, clearing house or central depository) than when there exists a complex nexus of opaque, over the counter (OTC) transactions. An interesting innovation in this direction is the development by the Deposit Trust and Clearing Corporation of a new facility that provides central settlement to major OTC derivatives dealers. In the same vein, why not use central clearing and settlement platforms for reforming the industrial organization of the credit rating industry? Many commentators have indeed accused the credit rating agencies (CRAs) of bearing a strong responsibility in the current credit crisis. They argue that CRAs may have deliberately underestimated the risks of some mortgage backed securities pools or collateralized debt obligations. They criticize the “issuer pays” model as creating the possibility of conflicts of interest. Since the bulk of CRAs’ revenues come from issuers and arrangers, it is not inconceivable that CRAs could have temporarily run the risk of jeopardizing their reputation by

08 Book.indb 479

2/13/09 3:58:56 PM

480

Jean-Charles Rochet

inflating credit ratings in order to earn more structuring fees. Increasing regulatory scrutiny on the ratings process itself would probably be difficult, and in the end, largely inefficient. Returning to the “investors pay” model of the past is likely to be impossible. Brian Clarkson, the president of Moody’s, is pessimistic: “Whoever pays, there will be a conflict” (The Economist, February 7, 2008). I would like to put forward an alternative solution that could solve these conflicts of interest. It is based on the following analogy. People who want to sell valuable paintings often use the services of an auction house like Sotheby’s, who organizes the auctioning of the paintings. Typically the seller requires the assistance of experts, who certify the authenticity of the paintings. For obvious reasons, these experts are almost always hired and remunerated by the auction house and never by the seller itself. The same is true if the seller wants to exhibit his paintings into an art gallery, in order to facilitate the sales. It is the gallery that organizes the certification, not the seller. By analogy, suppose that an arranger wants to issue some asset backed securities and wants to apply for credit ratings by a Nationally Recognized Statistical Rating Organization (NRSRO). The proposal would be that this potential issuer is required to contact a “central platform” that could be a central depository, a clearing house or an exchange. This platform would be completely in control of the rating process and could also provide record keeping services to the different parties in the securitization operation. The idea would be to cut any direct commercial links between issuers and CRAs. The potential issuer would pay a (pre-issue fee) to the central platform, who would then organize the rating of the securities by one or several NRSROs. The rating fees would be paid by the central platform to the NRSROs. These fees would obviously be independent of the outcome of the rating process and of the fact that the issue finally takes place or not. This would eliminate any perverse incentives for a lax behavior by CRAs. This would also solve the conflict of interest between issuers and investors,2 since the central platform’s profit maximization depends on appropriately aggregating the interests of the two sides of the market.

08 Book.indb 480

2/13/09 3:58:56 PM

Commentary

481

Summary and conclusion Let me conclude by briefly summarizing the main points of my comments on this very interesting paper: • Rethinking capital regulation is indeed important: The current crisis has clearly shown how ill-designed regulation could distort incentives in ways that increase systemic risk. In particular, the VaR approach to credit risk has encouraged banks to shift risks towards the upper tail of the loss distributions. I believe it should be reconsidered. Value at Risk may be a good metric for banks, since they are protected by limited liability, but it is certainly not a good risk measure for public authorities, who ultimately bear the costs of large losses. • Other sources of financing for banks, such as the capital insurance contracts suggested by Kashyap, Rajan and Stein, could indeed improve things, but only if regulators make sure that this does not lead to regulatory arbitrage by banks and ultimately increase aggregate risk in the financial sector. • Centralized trading, clearing or depository facilities can also provide a solution to the conflict of interest in the credit rating industry. If the rating process is left entirely to the control of these platforms in such a way that all commercial links between CRAs and issuers are cut, this would reduce perverse incentives for these CRAs to inflate ratings in order to increase their revenues.

08 Book.indb 481

2/13/09 3:58:56 PM

482

Jean-Charles Rochet

Endnotes For example, Tom Hoenig, president and CEO of the Kansas City Fed, has recently argued (in his speech “Perspectives on the Recent Financial Turmoil” for the IIF membership meeting, Rio de Janeiro, March 5, 2008) that “the response to this crisis should be fundamental reform, not Band-Aids and tourniquets” and that “both the private sector and the government will have key roles to play in articulating needed reforms and ensuring that they are implemented.” 1

As rightly pointed out by Charles Calomiris (2008), rating inflation could also be demand driven if there are conflicts of interest between asset managers and investors. Solving the other conflicts of interest would necessitate additional policy measures. 2

08 Book.indb 482

2/13/09 3:58:56 PM

Commentary

483

References Adrian, T. and H.S. Shin (2008). “Financial Intermediaries, Financial Stability and Monetary Policy,” paper presented in the Federal Reserve Bank of Kansas City Symposium, “Maintaining Stability in a Changing Financial System,” Jackson Hole, Wyoming, August 21-23, 2008. Ashcraft, A. and T. Schuermann (2008). “Understanding the Securitization of Subprime Mortgage Credit,” Foundations and Trends in Finance, vol 2 issue 3, 191-309. Brunnermeier, M. (2008). “Deciphering the 2007-08 Liquidity and Credit Crunch,” forthcoming, Journal of Economic Perspectives. Calomiris, C.W. (2008). “The Subprime Turmoil: What’s New, What’s Old, and What’s Next,” paper presented in the Federal Reserve Bank of Kansas City Symposium, “Maintaining Stability in a Changing Financial System” Jackson Hole, Wyoming, August 21-23, 2008. Flannery, M. (2005). “No Pain, No Gain? Effecting Market Discipline via ‘Reverse Convertible Debentures,’” in Hal S. Scott (ed.), Capital Adequacy beyond Basel: Banking, Securities, and Insurance, Oxford: Oxford University Press. Gordy, M. (2003). “A Risk-Factor Model Foundation for Ratings Based Bank Capital Rules,” Journal of Financial Intermediation, 12:199-232. Holmström, B. and J. Tirole (1998). “Private and Public Supply of Liquidity.” Journal of Political Economy, 106, 1-40. Kashyap, A., R. Rajan and J. Stein (2008). “Rethinking Capital Regulations,” paper presented in the Federal Reserve Bank of Kansas City Symposium, “Maintaining Stability in a Changing Financial System,” Jackson Hole, Wyoming, August 21-23, 2008. Kashyap, A. and J. Stein (2003). “Cyclical Implications of the Basel 2 Capital Standards.” Plantin, G. and J.C. Rochet (2008). When Insurers Go Bust, Princeton, Princeton University Press.

08 Book.indb 483

2/13/09 3:58:56 PM

08 Book.indb 484

2/13/09 3:58:56 PM

General Discussion: Rethinking Capital Regulation Chair: Stanley Fischer

Mr. Liikanen: Thank you very much for a very innovative paper. First a comment and then a question. This comment comes actually from Raghu’s paper delivered here three years ago in 2005 titled, “Has Financial Development Made the World Riskier?” His reply was “Yes.” All the critics here said, “No.” So I don’t dare to criticize your paper, Raghu. That’s all. But I want to put a question on the present paper. We are discussing very much in Europe and other areas about multinational banking institutions. How would this apply in the multinational case, where banks operate, let’s say, in four to five countries? Why is it such a concrete question? We have now in Europe banks, for instance, which are not systemically important in the home country, but are systemically important in the host country. The cross-border solution is quite critical to us. Have you had any thoughts on that issue? Mr. Calomiris: I want to applaud the paper and say that I think it may be a good idea. I just want to offer three quick comments to get your reactions about possible refinements or problems you may want to address. First, it is interesting to ask, How is the financial system going to react if this becomes a reality? I can think of two obvious reactions 485

08 Book.indb 485

2/13/09 3:58:56 PM

486

Chair: Stanley Fischer

that are undesirable. The first reaction is a substantial increase in the correlation of risk in the banking system. Why? Because now all banks have a very strong incentive to write deeply out of the money S&P 500 puts. So the amount of systemic risk goes up when you insure systemic risk. That is an important potential problem. The second problem is that we are collateralizing this insurance— mono-line insurance companies might provide this insurance—and we are going to collateralize it with Treasury bonds. That will encourage them to provide other kinds of credit enhancement to the banks more aggressively on an uncollateralized basis, so they can asset strip to get back to where they wanted to get to in the combined exposure they have to the banks. So you collateralize this exposure, but then you create more uncollateralized exposure that basically undoes it. Third, as I read your paper, you seem to be saying that we might as well give up on the ability to enforce traditional capital regulation based on accounting concepts. This seems to require further discussion. Mr. Blinder: I liked this proposal very much … I think. That is what is leading to my question. The first point I want to make takes up right where Charlie left off. In the presentation of the paper, there is a lot of prose that sounds like and says raising capital has a lot of downsides and is not such a good idea. But the proposal does raise capital requirements. That is just a stylistic question. I don’t think you are wrong about that. I think you are right about that, but there is a bit of a tone, as Charlie said, that raising capital is not a good idea. The paper is really about raising capital in a somewhat different way. Now here is the question. A recent year’s piece of Wall Street wisdom you all know is that in a crisis all correlations go to one. So this is what I am wondering about. This is an insurance policy that will pay off only in very bad states of the world when the portfolios of just about everybody are taking a hit. This is part of the question. Can you find a class of people who are actually enjoying these bad times? Because if you can’t, and I don’t think you can, it seems to me the insurance premium on this is going to be extremely high because you are making people pay at times when they don’t want to pay. I

08 Book.indb 486

2/13/09 3:58:56 PM

General Discussion

487

am wondering about that vice as against, for example, the ingenious suggestion we just heard from Mr. Rochet or Flannery’s idea of these convertible bonds, which are basically forcing people to acquire stock when it’s cheap. Mr. Sperling: My question is for the authors, and it is about the insurers. My question is—as you are thinking about your proposal actually succeeding—whether you are at all worried you might create a new class of too-big-to-fail, too-interrelated-to-fail institutions? You talk about will people do this? Now you’re immediate answer will be, “Yes, but we have this lockbox”—but that lockbox is highly essential to your model that you are not creating too-big-to-fail insurers. Even if you do have this lockbox, presumably there would be a couple of institutions—Fannie Capital Reinsurance whatever—and they would become very good at this, and they would rely on being paid to roll over. I wonder whether you think, if this developed, if the companies insuring the capital requirements were to go under, whether you would find yourself in another different too-big-to-fail regulatory issue? Mr. Holmström: Yes, two comments. One going back to the incentive problem that you talked about: It’s fairly easy to blame because incentives are always in a tautological sense the cause. There is some anecdotal evidence on headquarters of institutions being liable rather than their traders. Allegedly the UBS board and top management kicked off a campaign to get traders into lucrative by risky derivatives. In the Scandinavian crisis, it is interesting that banks that were decentralized and let their loan managers decide on the loans largely on their own, those banks actually did pretty well, whereas banks where the center decided on the loan-to-value ratio as a way of controlling lending, those banks really got into trouble. That is true both in Finland and Sweden. So that indicates that the problem is not the rogue traders necessarily. I am not saying there isn’t that problem, too, but I would urge people to look carefully at the incentives before they jump to conclusions about the underlying problem. Then a comment on your insurance scheme: It may be a good scheme if the problem is to redistribute a fixed amount of

08 Book.indb 487

2/13/09 3:58:56 PM

488

Chair: Stanley Fischer

collateral or Treasuries within the private sector. But what if there aren’t enough Treasuries? Then government has a role in supplying additional Treasuries. Government and private sector insurance are not competing schemes; they are complementary. I also want to emphasize that the government’s ability to inject Treasuries ex post saves a lot on the deadweight cost of taxation. The lockbox of Treasuries that you need in your scheme incurs needlessly high deadweight costs of taxation. With private insurance, you have to determine contingencies in advance, but you can’t forecast what contingencies will happen. You may think it’s some crisis of the sort we see now, but if it is a very different crisis, we need ex post judgment and intervention, and only government can provide that. Mr. Carney: I join the others in complimenting the paper and the idea. I want to address this, though, by picking up on Peter Fisher’s point on consolidation in the industry and think about how your proposal might influence consolidation. One of the things, certainly at the margin, is capital is going to flow to the relatively stronger institutions. Or, to put it another way, strong institutions will also get capital with this scheme. That raises a couple of issues that always can be addressed. On the one hand­—for the stronger institutions right now—part of the reason why their share prices are holding up is because they are not going to issue additional capital. So there is this dilution issue with the proposal. Do you have to add to this idea an ability to have the option to flow through the capital proceeds to the shareholders on a tax-efficient basis, so you don’t get an overhang for stronger institutions? Point one. Point two: You mentioned consolidation, but obviously when there is consolidation in the industry, it is almost exclusively done on an equity basis—to achieve a tax-free rollover. Here you would have to do consolidation for cash. The last point: One thing we haven’t had a chance to address is there are some real accounting issues right now that are preventing

08 Book.indb 488

2/13/09 3:58:56 PM

General Discussion

489

consolidation in the sector. As we think going forward, some of these issues may need to be addressed in order to get us out of it. Mr. Lindsey: I would like to inject an ideological heterodox note here. When I hear all your talk about the various agency problems and also the difference between endogenous and exogenous risk, why don’t we default to what we’ve historically always defaulted to, which is some combination of nationalization and monetization? After all, who has better information than the government and regulator? When we think about endogenous versus exogenous risk, let’s face it, all the exogenous risk, as you described, or much of it has to do with public policy. I hate to ask the question I’ve just asked, but you haven’t convinced me yet that we shouldn’t default back to good old Uncle Sam. Mr. Meltzer: Like everyone else, I think this is a very interesting paper. It’s one of several papers here like the paper by Charles Calomiris and many others that at least try to think about the problem of the incentives that are created by the regulation. That is something very different from what lawyers usually do. They don’t think about the incentives, and therefore set up what I call the first law of regulation. They regulate, and the markets circumvent. We have a system where we’re always going to be subject to problems because financial institutions borrow short and lend long—and they’re subject to unforeseen permanent shocks, which are hard or impossible to anticipate in most cases. So we’ve gone from a system which worked very well—I am going to talk about that—to a system which in my opinion cannot survive. We neglect in our discussion, of course, one of the reasons why we’ve shifted from that system. It is called the Congress, or more generally, the members who are much more interested in redistribution than in efficiency. We had a system which was relatively efficient and worked for a hundred years, and that was the British system that became famous as Bagehot’s rule. But Bagehot didn’t criticize the Bank of England for not using his rule. He criticized them for not announcing it in advance. He was an early rational expectationist. He said the Bank of England should announce the rule and, if they did, there would

08 Book.indb 489

2/13/09 3:58:56 PM

490

Chair: Stanley Fischer

be fewer crises. That rule worked very well in Britain for a hundred years. The reason we don’t have it is because today Congress and regulators are much more concerned about redistribution than they are about efficiency. Bagehot’s rule said, “Let them fail.” Failure in the modern world would mean that we wipe out the equity owners and we wipe out the management, as we did on a couple of occasions—for example, Continental Illinois Bank—and lend freely to those people who have a problem. Anna Schwartz showed, as they say, that this worked very well in the U.K. for over a hundred years. That was certainly a system which was a multinational system. They were lending all over the world. They got into the famous Bering crisis because of Argentina’s default and so on. But they managed to survive through those crises without great problems of the kind that we now have. Who will claim the current system is doing better? Not I. I close with this reminder, especially for the authors. In the 1920s, if you went to a bank, the thing that stood out for you most was on the window. It said “capital and surplus” and it listed what those were. By the 1950s, those were gone, and what it said was “member of FDIC.” Franklin Roosevelt recognized that FDIC would create a moral hazard problem. That is why he opposed it. Of course, these rules— like the FDIC and others—may have very good properties, but they have created many of the risks we have. In order to respond to those risks, the best thing we can do is go back to Bagehot’s rule—that is, let them fail, but clean up the secondary consequences. Mr. Redrado: I have been thinking about the implementation capacity of your insurance proposal. In particular, how to make it operational in the international arena, especially when you look at international banks that could suffer losses in one country and shift it to another or move operations from regulated to less-regulated places. What I wonder is: What kind of counterparty have you thought about? When I think about how to implement such a proposal, it seems to me you could give a role to multilateral entities, in particular, the BIS, where most of the central banks have a portion of their own reserves. It could be a conduit to be a counterparty for an

08 Book.indb 490

2/13/09 3:58:56 PM

General Discussion

491

international situation. Moreover, in rethinking the role of the IMF– let me recall that you and I have talked about the possibility of irrelevance of the IMF. I wonder if you have thought that international financial institutions could have a role in being the counterparty of this insurance scheme. Mr. Crockett: I like the insurance proposal quite a lot, but as you recognize, of course, there are lots of details that need to be looked at. One of them that I don’t think has been mentioned yet is the valuation of the claims the insurer would get in the event it was activated. It is very difficult to value a franchise in the circumstances of a financial crisis, if the insurer is constrained automatically to put in the amount, which of course is in the lockbox and then transferred. Mr. Rosengren: The idea of contingent capital is very interesting, and it is a good idea. You framed the proposal in terms of insurance. It would seem like a simpler structure would be using options, so that if you had long-dated, out-of-the-money puts on a portfolio of financial stocks, it would seem to be much easier to implement. It would eliminate some of the counterparty concerns and implementation concerns that people raised. You could set the capital relative to the strike price, rather than the premium you paid at initiation. You could imagine a situation where it would have many of the characteristics that you want, but get around some of the implementation problems that people have discussed. Mr. Bullard: I liked this paper. There are many nice market-oriented ideas. I think the idea of fire-sale asset prices is not so good. The idea of rationalizing government intervention based on the existence of times of large classes of assets trading below fundamentals somehow does not strike me as sound. It links up to this idea of, How is this trigger going to work when, as previous speakers have said, it is very difficult to value the firm in the middle of a crisis? What is the role of marked to market in this scheme? Mr. Stein: Thanks very much. There are a lot of terrific comments and we’ll try to get to a subset of them. First to Jean-Charles, who raised this question of, Do you want to have the private sector do

08 Book.indb 491

2/13/09 3:58:57 PM

492

Chair: Stanley Fischer

this, in which case you rely on the lockbox as opposed to having the government do it? A couple of issues. I don’t think the 100 percent lockboxing is expensive or socially costly, if there is not a general equilibrium shortage of Treasury securities. Now you can earn the interest on them. There is nothing dissipative happening. It is only if there is some kind of a general equilibrium shortage. So that is one reason. If you thought there was an equilibrium shortage, you might be drawn to having the government do it. The other reason, as Bengt alluded to, to have the government do it, would be, Our thing relies on defining a trigger ex ante. We have to specify what the bad state looks like. It is bank losses greater than $200 billion. The government can do it like pornography. They can just recognize it when they see it, which is an advantage. They can condition on more information. The reason we are drawn to the private option is in direct response to this question. We think there is a downside of having the government do it, which is with this discretion comes political economy concerns of all sorts. It’s not that this might not be complementary, but to the extent you can go as far as you can with the private sector, that is a good thing. Charlie Calomiris asked about herding. Will everybody want to take the same kind of risk? There is a logical argument here. It is a little more subtle than maybe people realize. It is not the standard moral hazard problem. You don’t get paid back based on your own losses. So there is no expected return rationale for herding. There is only a second-moment rationale for herding that it might lower the variance of your portfolio. Something like that. Some of this is addressed with the trigger. If this option is sufficiently far out of the money and you do the same stupid thing as everybody—and instead of hitting a crisis—we just hit a very bad state, such that the trigger is not passed. You will bear all the burden of this. One virtue of this option structure is, if you set it sufficiently far out of the money,

08 Book.indb 492

2/13/09 3:58:57 PM

General Discussion

493

there is a big deductible on the herding strategy as well. Just wanted to make that one point. A point that both Charlie and Alan Blinder raised is, Are we giving up on capital regulation? No. If we said that, we’ve misspoke. The way I think about it is, We recognize there is going to be capital regulation. In fact, there is going to be a push to raise capital, and we want to make a form of capital that is cheaper. The specific mechanism is, What makes capital expensive is giving guys discretion in all states of the world raises agency costs. We want to give them cash only in a specific, smaller number of states of the world where the agency costs are lower because the social value of having banks do a lot of investment is higher. To Alan’s second point, the CAPM risk premium. You said, “Well, the insurer is on the wrong side of a contract which pays out in a very adverse state of the world.” That’s going to have not only an unexpected return component to the pricing, but it should have a beta component. That absolutely must be right. Of course, this is true when you give somebody equity as well. There are those bad states. They are going to lose money in those bad states with unconditional capital. You asked about the Flannery proposal. The Flannery proposal is similar to ours—for those of you who don’t know it—but it is insurance that is firm-specific, so it will pay off whenever the individual bank does badly as opposed to the whole system doing badly. That just pays off in more states of world. It pays off in the very bad systemic states of the world and it pays off in the firm-specific states. Since it’s paying off in more states, it has to be more expensive. There is certainly an equilibrium cost to be borne here. I don’t know if you get around it. Last one I’ll speak to—put options. Our thing is a put option. The strike price is, as we have envisioned it, bank earnings, rather than stock prices. There is a potential real problem with striking the option on stock prices because they are endogenous, and you worry about all kinds of feedback effects­—manipulation and things. If

08 Book.indb 493

2/13/09 3:58:57 PM

494

Chair: Stanley Fischer

people think the banking sector is not going to be paid off, that will affect the prices. You like these things to be hooked to something that is hopefully a little bit more exogenous. Mr. Kashyap: Let me start with Allan Meltzer’s point. The Bagehot advice in these illiquid markets is very difficult implement because there are multiple equilibria. Let’s suppose we are not sure what the price is because a security is not trading and everybody is unsure what the price is going to be. If you don’t lend, that removes buyers, pushes down the price, and something that could start out as a liquidity problem quickly becomes a solvency problem. In figuring out whether or not you want to mark to market and just make these decisions is much more difficult in practice than it was a hundred years ago. So as a practical matter, it is very difficult to decide how to apply Bagehot when you are looking at actual entities. On the multinational question that came up several times, we struggled with a way to do that. The way we would imagine this is you will have a principal regulator. You will have to go to your principal regulator and convince them your operations across all markets are substantially insured. Whether the BIS would be the place the reporting goes to, so everybody knows what the operations are across markets, is a detail that is quite important to work out. We are worried about this tradeoff between a bank getting in trouble in Vietnam and then exporting its problems into the United States. We would like to make sure the Vietnam stuff is insured there, the U.S. part is insured in the United States, and there is enough collective insurance. Let me just close with saying two broad things. First of all, we view this as a complement to lots of other good existing proposals. Secondly, even if you don’t buy into the proposal, we hope to change the discussion about capital regulation from simply trying to keep forcing them to hold capital and never thinking about why they don’t want to hold it, to recognizing there are good reasons why they view capital as expensive. Attempting to design regulation so that you address those underlying frictions—whether using our solution or not —we think that is the single biggest point.

08 Book.indb 494

2/13/09 3:58:57 PM

Central Banks and Financial Crises Willem H. Buiter

Introduction In this paper I draw lessons from the experience of the past year for the conduct of central banks in the pursuit of macroeconomic and financial stability. Modern central banks have three main tasks: (1) the pursuit of macroeconomic stability; (2) maintaining financial stability and (3) ensuring the proper functioning of the “plumbing” of a monetary economy, that is, the payment, clearing and settlement systems. I focus on the first two of these, and on the degree to which they can be separated and compartmentalised, conceptually and institutionally. My thesis is that both monetary theory and the practice of central banking have failed to keep up with key developments in the financial systems of advanced market economies, and that as a result of this, many central banks were to varying degrees ill-prepared for the financial crisis that erupted on August 9, 2007. The empirical illustrations will be drawn mainly from the experience of three central banks, the Federal Reserve System (Fed), the Eurosystem (ECB) and the Bank of England (BoE), with occasional digressions into the experience of other central banks. Discussion of mainly Fed-related issues will account for well over one third of the paper, partly in deference to the location of the Jackson Hole Symposium, but mainly because I consider the performance of the Fed to have been 495

08 Book.indb 495

2/13/09 3:58:57 PM

496

Willem H. Buiter

by some significant margin the worst of the three central banks, as regards both macroeconomic stability and financial stability. In many ways, August 2008 is far too early for a post mortem. Both the financial crisis and dysfunctional macroeconomic performance are still with us and are likely to remain with us well into 2009: Inflation and inflation expectations are above-target and rising (see Chart 1 and Charts 2a,b), output is falling further below potential (see Charts 3a,b) and there is a material risk of recession in the US, the UK and the euro area.1,2 Nevertheless, I believe that, although a final verdict may have to wait another couple of generations, there are some lessons that can and should be learnt right now, because they are highly relevant to policy choices the monetary authorities will face in the months and years immediately ahead. Such, in any case, have been the justifications for even earlier crisis post-mortems written by myself and others (see e.g. Buiter, 2007f, 2008b, and Cecchetti, 2008). Possibly because truly systemic financial crises have been few and far between in the advanced industrial countries since the Great Depression (the Nordic financial crisis of 1992/1993 is a notable exception (see Ingves and Lind, (1996), and Bäckström, (1997)), most central banks in the North Atlantic region—the region where the crisis started and has done the most damage—were not prepared for the storm that hit them. It is therefore not surprising that mistakes were made. The incidence and severity of the mistakes were not the same, however, for the three central banks. I find that the Fed performed worst as regards macroeconomic stability and as regards one of the two time dimensions of financial stability—minimising the likelihood and severity of future financial crises. As regards the other time dimension of financial stability, dealing with the immediate crisis, the BoE gets the wooden spoon, because of its failure to act appropriately in the early days of the crisis. I argue that three factors contribute to the Fed’s underachievement as regards macroeconomic stability. The first is institutional: The Fed is the least independent of the three central banks and, unlike the ECB and the BoE, has a regulatory and supervisory role; fear of political encroachment on what limited independence it has and

08 Book.indb 496

2/13/09 3:58:57 PM

Central Banks and Financial Crises

497

cognitive regulatory capture by the financial sector make the Fed prone to over-react to signs of weakness in the real economy and to financial sector concerns. The second is a sextet of technical and analytical errors: (1) misapplication of the “Precautionary Principle”; (2) overestimation of the effect of house prices on economic activity; (3) mistaken focus on “core” inflation; (4) failure to appreciate the magnitude of the macroeconomic and financial correction/adjustment required to achieve a sustainable external equilibrium and adequate national saving rate in the US following past excesses; (5) overestimation of the likely impact on the real economy of deleveraging in the financial sector; and (6) too little attention paid (especially during the asset market and credit boom that preceded the current crisis) to the behaviour of broad monetary and credit aggregates. All three central banks have been too eager to blame repeated and persistent upwards inflation surprises on “external factors beyond their control,” specifically food, fuel and other commodity prices. The third cause of the Fed’s macroeconomic underachievement has been its proclivity to use the main macroeconomic stability instrument, the federal funds target rate, to address financial stability problems. This was an error both because the official policy rate is a rather ineffective tool for addressing liquidity and insolvency issues and because more effective tools were available, or ought to have been. The ECB, and to some extent the BoE, have assigned the official policy rate to their respective price stability objectives and have addressed the financial crisis with the liquidity management tools available to the lender of last resort and market maker of last resort. The BoE made the worst job of handling the immediate financial crisis during the early months (until about November 2007). The ECB, partly as the result of an accident of history, did best as regards putting out fires. The most difficult part of financial stability management is to handle the inherent tension between the two key dimensions of financial stability: The urgent short-term task of “putting out fires,” that is, managing the immediate crisis, and the vital long-run task of

08 Book.indb 497

2/13/09 3:58:57 PM

498

Willem H. Buiter

minimizing the likelihood and severity of future financial crises. Through their pricing of illiquid collateral, all three central banks may have engaged in behaviour that created unnecessary moral hazard, thus laying the foundations for future reckless lending and borrowing. In the case of the Fed this is all but certain, in the case of the ECB quite likely and in the case of the BoE merely possible. As regards the Fed, the nature of the arrangements for pricing illiquid collateral offered by primary dealers invites abuse. In the case of the BoE and the ECB, the secrecy surrounding their pricing methodology and models, and their unwillingness to provide information about the pricing of specific types and items of illiquid collateral make one suspect the worst. These distorted arrangements (in the case of the Fed) and lack of transparency as regards actual pricing (for all three central banks) continue. The reason the Fed did worst in this area also is probably again due to the fact that, unlike the ECB and the BoE, the Fed is a financial regulator and supervisor for the banking sector. Cognitive regulatory capture of the Fed by Wall Street resulted in excess sensitivity of the Fed not just to asset prices (the “Greenspan-Bernanke put”) but also to the concerns and fears of Wall Street more generally. All three central banks have gone well beyond the provision of emergency liquidity to solvent but temporarily illiquid banks. All three have allowed themselves to be used in varying degrees as quasi-fiscal agents of the state, either by providing implicit subsidies to banks and other highly leveraged institutions, and/or by assisting in the recapitalisation of insolvent institutions, while keeping the resulting contingent exposure off the budget and balance sheet of the fiscal authorities. Such subservience to the fiscal authorities undermines the independence of the central banks even in the area of monetary policy. The unwillingness of the three central banks to reveal their valuation models for and actual valuations of illiquid collateral and, more generally, their unwillingness to provide the information required to calculate the magnitude of all their quasi-fiscal interventions, make a mockery of their accountability for the use of public resources. In Section I, I discuss the principles of macroeconomic stability and in Section II the principles of financial stability. Section III reviews the

08 Book.indb 498

2/13/09 3:58:57 PM

Central Banks and Financial Crises

499

records of the three central banks during the past year, first as regards macroeconomic stability and then as regards financial stability. Section IV concludes. I.

Macroeconomic stability

I.1

Objectives

The macroeconomic stability objectives of the three central banks are not the same. Both the ECB and the BoE have a lexicographic or hierarchical preference ordering with price stability in pole position. Only subject to the price stability objective being met (for the BoE) or without prejudice to the price stability objective (for the ECB) can these central banks pursue other objectives, including growth and employment. In the UK, the operationalization of the price stability objective is the responsibility of the Chancellor of the Exchequer. It takes the form of a 2 percent annual target inflation rate for the headline consumer price index or CPI. The ECB sets its own operational inflation target, an annual rate of inflation for the CPI that is below but close to 2 percent in the medium term. The Fed formally has a triple mandate: maximum employment, stable prices and moderate long-term interest rates.3 The third of these is habitually ignored, leaving the Fed in practice with a dual mandate: maximum employment and stable prices. Unlike the lexicographic ordering of ECB and BoE objectives, the Fed’s objective function can be interpreted as symmetric between price stability and real economic activity, in the sense that, in the central bank’s objective function, the one can be traded off for the other. This is captured well by the traditional flexible inflation targeting loss function L shown in equations (1) and (2). Here Et is the conditional expectation operator at time t, p is the rate of inflation, p* the (constant) target rate of inflation, y real GDP (or minus the unemployment rate) and y* the target level of output, which could be potential output (or minus the natural rate of unemployment) or, where this differs from potential output, the efficient level of output (the efficient rate of unemployment).

08 Book.indb 499

2/13/09 3:58:57 PM

500

Willem H. Buiter L t = Et

∞

 1   L 1+ δ 

∑ i=0

δ>0

i

(1)

t +i



Lt+i = (p t+i − p * )2 + w ( yt+i − yt*+i )2 w>0

(2)

With a lexicographic ordering, the central bank can be viewed as first minimizing the loss function in (1) and (2) with the weight on the squared output gap, w , set equal to zero. If there is a unique policy rule that solves this problem, this is the optimal policy rule. If there is more than one solution, the policy authority chooses among these i

the one that minimizes something like

∞  1  * L ty = Et ∑  yt+i − yt+i   i =0 1 + δ

(

2

).

“Maximum employment” is not a well-defined concept. Recent Fed chairmen have interpreted it as something close to the natural rate of unemployment or the NAIRU (the non-accelerating inflation rate of unemployment). In employment space this translates into the maximum sustainable level or rate of employment. In output space it becomes the maximum sustainable output gap (excess of actual over potential GDP) or the maximum sustainable growth rate of GDP. Price stability has not been given explicit numerical content by the Fed, the US Congress or any other authority. Since the Greenspan years, the Fed appears to have targeted a stable, low rate of inflation for the core personal consumption expenditure (PCE) deflator index. It has not always been clear whether the Fed actually targets core inflation or whether it targets headline inflation in the medium term and treats core inflation as the best predictor of medium-term headline inflation. As late as March 2005, the current Chairman of the Fed admitted to a “comfort zone” for the core PCE deflator of 1 to 2 percent (Bernanke, 2005). This is also consistent with the FOMC members’ inflation forecasts at a three-year horizon. In what follows, I will treat the Fed’s implicit inflation target as 1.5 percent for the headline PCE deflator or just below 2.0 percent for the headline CPI, given the usual wedge between PCE and CPI inflation rates.

08 Book.indb 500

2/13/09 3:58:58 PM

Central Banks and Financial Crises

501

The recent performance of the CPI inflation rates, of survey-based measures of 1-year and long-term inflation expectations and of real GDP growth rates for the US, the euro area and the UK are shown in Charts 1, 2a,b and 3a,b. I.2

Instruments

The key instrument of monetary policy for macroeconomic stabilisation policy is the short risk-free nominal rate of interest on nonmonetary financial instruments, henceforth the official policy rate, denoted i. This is the federal funds target rate in the US, the inelegantly named Main Refinancing Operations Minimum Bid Rate of the ECB and Bank Rate in the UK. In principle, the nominal exchange rate (either a bilateral exchange rate or a multilateral index) could be used as the instrument of monetary policy instead of the official policy rate. In practice, all three countries have market-determined exchange rates.4 I don’t consider sterilised foreign exchange market intervention (unilateral or internationally co-ordinated) to be a significant additional instrument of policy, unless foreign exchange markets were to become disorderly and illiquid—something that hasn’t happened yet. Reserve requirements on eligible deposits, when they are unremunerated, are best thought of as a quasi-fiscal tax. When remunerated, they can be viewed as part of a set of capital and liquidity requirements that can be used as financial stability instruments (see Section II below), but not as significant macroeconomic stabilisation instruments. The non-negativity constraint on the official policy rate has not been an issue so far in the current crisis. With the federal funds target rate at 2.00 percent, it is by no means inconceivable that i > 0 could become a binding constraint on the Fed’s interest rate policy before this crisis and cyclical downturn are over.5 In what follows, the official policy rate will be the only macroeconomic stabilisation instrument of the central bank I consider in detail. Because economic behaviour (consumption, portfolio demand, investment, employment, production, price setting) is strongly influenced by expectations of the future, both directly and through the

08 Book.indb 501

2/13/09 3:58:58 PM

502

Willem H. Buiter

Chart 1 Headline CPI inflation rates, 1989M1-2008M7 (percent) 9.0

Percent

9.0 UK euro area US

8.0

8.0 7.0

6.0

6.0

5.0

5.0

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

198901 198907 199001 199007 199101 199107 199201 199207 199301 199307 199401 199407 199501 199507 199601 199607 199701 199701 199801 199807 199901 199907 200001 200007 200101 200107 200201 200207 200301 200307 200401 200407 200501 200507 200601 200607 200701 200707 200801 200807

7.0

Percent change month on same month one year earlier Source: UK: ONS; euro area: Eurostat; US: BEA.

Chart 2a One-year ahead inflation expectations, 2000Q2-2008Q2 (percent) 6.0

6.0 UK US euro area

5.0

5.0

2008 Q2

2008 Q1

2007 Q4

2007 Q3

2007 Q2

2007 Q1

2006 Q4

2006 Q3

2006 Q2

2006 Q1

2005 Q4

2005 Q3

2005 Q2

2005 Q1

2004 Q4

2004 Q3

2004 Q2

2004 Q1

2003 Q4

2003 Q3

2003 Q2

2003 Q1

2002 Q4

2002 Q3

0.0 2002 Q2

0.0

2002 Q1

1.0

2001 Q4

1.0

2001 Q3

2.0

2001 Q2

2.0

2001 Q1

3.0

2000 Q4

3.0

2000 Q3

4.0

2000 Q2

4.0

Source: UK: Bank of England/GfK NOP Inflation Attitudes Survey (median); US: University of Michigan Survey (median); Euro area: ECB survey of professional forecasters (mean).

08 Book.indb 502

2/13/09 3:58:58 PM

Central Banks and Financial Crises

503

Chart 2b Long-term inflation expectations 4.5

Percent

4.5 USA UK euro area

4.0

4.0

Jun-08

Feb-08

Jun-07

Oct-07

Feb-07

Jun-06

Oct-06

Feb-06

Jun-05

Oct-05

Feb-05

Jun-04

0.0 Oct-04

0.0 Feb-04

0.5

Jun-03

0.5

Oct-03

1.0

Feb-03

1.0

Jun-02

1.5

Oct-02

1.5

Feb-02

2.0

Jun-01

2.0

Oct-01

2.5

Feb-01

2.5

Jun-00

3.0

Oct-00

3.0

Feb-00

3.5

Oct-99

3.5

Sources. USA: The University of Michigan 5-10 Yr Expectations: Annual Chg in Prices: Median Increase (%). UK: YouGov\CitiGroup Inflation expectations 5-10 years ahead. Median Increase (%) Euro area: ECB Survey of Professional Forecasters five years ahead forecast. Mean Increase (%)

Chart 3a Real GDP Growth Rates USA, Euro Area and UK 1956Q1 - 2008Q2 12.0

Percent

Percent UK

10.0

12.0 10.0

Euro Area 8.0 6.0

6.0

4.0

4.0

2.0

2.0

0.0

0.0

20064

20051

20032

20013

19994

19981

19962

19943

19924

19911

19892

19873

19854

19841

19822

19803

19784

19771

19752

19733

19714

19701

19682

19663

-6.0 19644

-6.0

19631

-4.0

19612

-4.0

19593

-2.0

19574

-2.0

19561

08 Book.indb 503

8.0

USA

2/13/09 3:58:59 PM

504

Willem H. Buiter

Chart 3b Real GDP growth rates US, Euro Area and UK 1996Q1 - 2008Q2* 6.0

6.0

UK Euro Area USA

5.0

5.0

20081

20073

20071

20063

20061

20053

20051

20043

20041

20033

20031

20023

20021

20013

20011

20003

0.0 20001

0.0

19993

1.0

19991

1.0

19983

2.0

19981

2.0

19973

3.0

19971

3.0

19963

4.0

19961

4.0

Source: UK and euro area: Eurostat; US: Bureau of Economic Analysis. Notes: quarter on same quarter one year earlier; * for euro area, data end in 2008Q1.

effect of these expectations on long-duration asset prices, it is not just past and current realisations of the official policy rate that drive outcomes, but the entire distribution of the contingent future sequence of official policy rates. The effect of a change in the current official policy rate is therefore the sum of the direct effect (holding constant expectations of future rates) and the indirect effect of a change in the current official policy rate on the distribution of the sequence of future contingent official policy rates. This leveraging of future expectations effectively permits future interest rates to be used as instruments multiple times (provided announcements are credible): Once at the date the actual official policy rate is set, i(t1), say, and through announcements or expectations of that official policy rate at dates before t1. By abuse of certainty equivalence, I will summarise this announcement effect as {At (it );j > 1}, where At (it ) is the 1-j

1

1-j

1

announcement of the period t1 policy rate in period t1-j. “Announcement” should be interpreted broadly to include all the hints, nudges,

08 Book.indb 504

2/13/09 3:58:59 PM

Central Banks and Financial Crises

505

winks and other forms of verbal and non-verbal communication engaged in by the authorities. This means that an opportunistic policy authority (one incapable of credible commitment to a specific contingent future policy rule) will be tempted, if it has any credibility at all, to use announcements of future policy rates as independent instruments of policy, unconstrained by the commitment or consistency constraint that the announcement of the future official policy rate, or of the future rule for setting the official policy rate, be equal to the best available current guess about what the authorities will actually do at that future date, which can be expressed as At (it )=Et (it ). 1-j

II.

1

1-j

1

Financial stability

I adopt a narrow view of financial stability. Sometimes financial instability is defined so broadly that it encompasses any inefficiency or imbalance in the financial system. In what follows, financial stability means (1) the absence of asset price bubbles; (2) the absence of illiquidity of financial institutions and financial markets that may threaten systemic stability; and (3) the absence of insolvency of financial institutions that may threaten systemic stability. I deal with the three in turn. II.1 Should central banks use the official policy rate to try to influence asset price bubbles? The original Greenspan-Bernanke position that the official policy rate should not be used to tackle asset booms/bubbles is convincing (Greenspan, 2002; Bernanke, 2002; Bernanke and Gertler, 2001). To the extent that asset booms influence or help predict the distribution of future outcomes for the macroeconomic stability objectives (price stability or price stability and sustainable economic growth), they will, of course, already have been allowed for under the existing approaches to maintaining macroeconomic stability in the US, the euro area and the UK. But the official policy rate should not be used to “lean against the wind” of asset booms and bubbles beyond addressing their effect on

08 Book.indb 505

2/13/09 3:58:59 PM

506

Willem H. Buiter

or informational content about the objectives of macroeconomic stabilisation policy, that is, asset prices should not be targeted with the official policy rate “in their own right.” First, this would “overburden” the official policy rate, which is already fully engaged in the pursuit of price stability and, in the case of the US, in the pursuit of price stability and sustainable growth. Second, asset price bubbles are, by definition, driven by non-fundamental factors. Going after an asset bubble with the official policy rate—a fundamental determinant of asset prices—may well turn out to be like going after a rogue elephant with a pea shooter. It could require a very large peashooter (a very large increase in the official policy rate) to have a material effect on an asset price bubble. The collateral damage to the macroeconomic stability objectives caused by interest rate increases capable of subduing asset price bubbles would make hunting bubbles with the official policy rate an unattractive policy choice. Mundell’s principle of effective market classification (Mundell, 1962) suggests that the official policy rate not be assigned to asset bubbles in their own right. That, however, leaves a major asymmetry in the macroeconomic policy and financial stability framework. This asymmetry is not that the official policy rate responds more sharply to asset market price declines than to asset market price increases. Even if there were no “Greenspan-Bernanke put,” such asymmetry should be expected because asset price booms and busts are not symmetric. Asset price busts are sudden and involve sharp, extremely rapid asset price falls. Even the most extravagant asset price boom tends to be gradual in comparison. So an asymmetric response to an asymmetric phenomenon is justified. This does not mean that there has been no evidence of a “Greenspan-Bernanke” put during the current crisis. I believe that phenomenon—excess sensitivity of the federal funds target rate to sudden declines in asset prices, and especially US stock prices—to be real, and will address the issue in Section III.2a below. Operationally, the asymmetry is that there exists a panoply of liquidity-enhancing, credit-enhancing and capital-enhancing measures

08 Book.indb 506

2/13/09 3:58:59 PM

Central Banks and Financial Crises

507

that can be activated during an asset market bust or a credit crunch, to enhance the availability of credit and capital and to lower its cost, but no corresponding liquidity- and credit-restraining and capital-diminishing instruments during a boom. When financial markets are disorderly, illiquid or have seized up completely, the lender of last resort and market maker of last resort (discussed in Section II.3) can spring into action. Examples abound. Sensible proposals from the SEC in the US that require putting a range of off-balance sheet vehicles back on the balance sheets of commercial banks are waived or postponed for the duration of the financial crisis because implementation now would further squeeze the available capital of the banks. Given where we are, this makes sense, but where was the matching regulatory insistence on increasing capital and liquidity ratios during the good times? We even have proposals now that mark-to-market accounting rules be suspended during periods of market illiquidity (see e.g. IIF, 2008). The argument is that illiquid asset markets undervalue assets compared to their fundamental value in orderly markets, and that because of this fair value accounting and reporting rules are procyclical. The observation that mark-to-market behaviour is procyclical is correct, but suspending mark-to-market when markets are disorderly would introduce a further asymmetry, because orderly and technically efficient asset markets can produce valuations that depart from the fundamental valuation because of the presence of a bubble. There have been no calls for mark-to-market accounting and reporting standards to be suspended during asset price booms and bubbles. Fundamentally, what drives this operational asymmetry is the fact that the authorities are unable or unwilling to let large highly leveraged financial institutions collapse. There is no matching inclination to expropriate, to subject to windfall taxes, to penalise financially or to restrain in other ways extraordinarily profitable financial institutions. This asymmetry creates incentives for excessive risk taking by the financial institutions concerned and has undesirable distributional consequences. It needs to be corrected. I believe a regulatory response is the only sensible one.

08 Book.indb 507

2/13/09 3:58:59 PM

508

Willem H. Buiter

II.2 Regulatory measures for restraining asset booms I propose that any large and highly leveraged financial institution (commercial bank, investment bank, hedge fund, private equity fund, SIV, conduit, other SPV or off-balance sheet entity, currently in existence or yet to be created—whatever it calls itself, whatever it does and whatever its legal form—be regulated according to the same set of principles aimed at restraining excessive credit growth and leverage during financial booms. Again, this regulation should apply to all institutions deemed too systemically important (too large or too interconnected) to fail. Therefore, while I agree with the traditional Greenspan-Bernanke view that the official policy rate not be used to target asset market bubbles, or even to lean against the wind of asset booms, I do not agree that the best that can be done is for the authorities to clean up the mess after the bubble bursts. II.2a Leverage is the key The asymmetries have to be corrected through regulatory measures, effectively by across–the–board credit (growth) controls, probably in the form of enhanced capital and liquidity requirements. Every asset and credit boom in history has been characterised by rising, and ultimately excessive leverage, and by rising and ultimately excessive mismatch. Mismatch here means asset-liability mismatch or resources-exposure mismatch as regards maturity, liquidity, currency denomination, credit risk and other risk characteristics. The crisis we are now suffering the consequences of is no exception. Because mismatch only becomes a systemic issue if there is excessive leverage, and because increased leverage is largely motivated by the desire of the leveraged entity for increased mismatch, I will focus on leverage in what follows. Leverage is a simple concept which may be very difficult to measure, as those struggling to quantify the concept of embedded leverage will know. In the words of the Counterparty Risk Management

08 Book.indb 508

2/13/09 3:58:59 PM

Central Banks and Financial Crises

509

Group II (2005), “...leverage exists whenever an entity is exposed to changes in the value of an asset over time without having first disbursed cash equal to the value of that asset at the beginning of the period.” And: “...the impact of leverage can only be understood by relating the underlying risk in a portfolio to the economic and funding structure of the portfolio as a whole.” Traditional sources of leverage include borrowing, initial margin (some money up front—used in futures contracts) and no initial margin (no money up front—when exposure is achieved through derivatives). I propose using simple measures of leverage, say a measure of gross exposure to book equity, as a metric for constraining capital insolvency risk (liabilities exceeding assets) of all large, highly leveraged institutions. Common risk-adjusted Basel II-type capital adequacy requirements and reporting requirements would be imposed on all large institutions whose leverage, according to this simple metric, exceeds a given value. These capital adequacy requirements would be varied (or vary automatically) in countercyclical fashion. To address the second way financial entities can fail, what the CRMG calls liquidity insolvency (meaning they cannot meet their obligations as they become due because they run out of cash and are unable to raise new funds), I propose that minimal funding liquidity and market or asset liquidity requirements be imposed on, respectively, the liability side and the asset side of the balance sheets of all large highly leveraged financial institutions. These liquidity requirements would also be tightened and loosened in countercyclical fashion. The regular Basel II capital requirements would provide a floor for the capital requirements imposed on all highly leveraged financial institutions above a certain threshold size. It is possible that Basel II will be revised soon to include minimum funding liquidity and asset liquidity requirements for banks and other highly leveraged financial institutions. If not, national regulators should impose such minimum funding liquidity and asset liquidity requirements on all highly leveraged financial institutions above a threshold size.

08 Book.indb 509

2/13/09 3:59:00 PM

510

Willem H. Buiter

Countercyclical variations in capital and liquidity requirements could either be imposed in a discretionary manner by the central bank or be built into the rule defining the capital or liquidity requirement itself. An example of such an automatic financial stabiliser is the proposal by Charles Goodhart and Avinash Persaud (Goodhart and Persaud, 2008a,b), to make the supplementary capital requirement for any given institution (over and above the Basel II requirement, which would set a common floor) an increasing function of the growth rate of that institution’s balance sheet. My wrinkle on this proposal (which Goodhart and Persaud propose for banks only) is that the same formula would apply to all highly leveraged financial institutions above a given threshold size. The Goodhart-Persaud proposal makes the supplementary-capitalrequirement-defining growth rate a weighted average (with declining weights) of the growth rate of the institution’s assets over the past three years. The details don’t matter much, however, as long as the criterion is easily monitored and penalises rapid expansion of balance sheets. A similar Goodhart-Persaud approach could be taken to liquidity requirements for highly leveraged institutions. If the assets whose growth rate is taxed or penalised under this proposal are valued at their fair value (that is, marked-to-market where possible), its stabilising properties would be enhanced. Finally, I would propose that all large leveraged institutions that are deemed too large, too interconnected, or simply too well-connected to fail, be made subject to a Special Resolution Regime along the lines that exist today for federally insured deposit-taking institutions through the FDIC. A concept of regulatory insolvency, which could bite before either capital insolvency or liquidity insolvency kick in, must be developed that allows an official administrator to take control of any large, leveraged financial institution and/or to engage in Prompt Corrective Action. The intervention of the administrator would be expected to impose serious penalties on existing shareholders, incumbent board and management and possibly on the creditors as well. The intervention should aim to save the institution, not its owners, managers or board, nor should it aim to “make whole,” that is, compensate in full, its creditors.

08 Book.indb 510

2/13/09 3:59:00 PM

Central Banks and Financial Crises

511

II.3 Liquidity management: From lender of last resort to market maker of last resort Liquidity management is central to the financial stability role of the central bank. Liquidity can be a property of economic agents and institutions or of financial instruments. Funding liquidity is the capacity of an economic agent or institution to attract external finance at short notice, subject to low transaction costs and at a financial cost that reflects the fundamental solvency of the agent or institution. It concerns the liability side of the balance sheet. Market liquidity is the capacity to sell a financial instrument at short notice, subject to low transaction costs and at a price close to its fundamental value. It concerns the asset side of the balance sheet. Both funding liquidity and market liquidity are continuous rather than binary concepts, that is, there can be varying degrees of liquidity. Funding liquidity (a property of institutions) and market liquidity (a property of financial instruments or the markets they are traded in) are distinct but interdependent. This is immediately apparent when one recognises that access to external funds often requires collateral (secured lending); the cost of external funds certainly depends on the availability and quality of the collateral offered. The value of the assets offered as collateral depends on the market liquidity of the assets. The central bank is unique because it can never encounter domestic-currency liquidity problems (domestic-currency funding illiquidity). This is because the monetary liabilities it issues, as agent of the state—the sovereign—provide unquestioned, ultimate domestic-currency liquidity. Often this finds legal expression through legal tender status for the central bank’s monetary liabilities. Central banks can, of course, encounter foreign-currency liquidity problems. The recent experience of Iceland is an example. There is no such thing as a perfectly liquid private financial instrument or a private entity with perfect funding liquidity, since the liquidity of private entities and instruments is ultimately dependent on confidence and trust. Liquidity, both funding liquidity and market liquidity, is very much a fair weather friend: It is there when you don’t need it, absent when you urgently need it. Although private

08 Book.indb 511

2/13/09 3:59:00 PM

512

Willem H. Buiter

agents may also lose confidence in the real value of the financial obligations of the state, including those of the central bank, the state is in the unique position of having the legitimate use of force at its disposal to back up its promises. The power to declare certain of your liabilities to be legal tender, the power to tax and the power to regulate (that is, to prescribe and proscribe behaviour) are unique to the state and its agents. The quality of private sector liquidity therefore cannot exceed that of central bank liquidity. Funding illiquidity and market illiquidity interact in ways that can create a vicious downward spiral, well described in Adrian and Shin (2007a,b) and Spaventa (2008). Faced with the disappearance of normal sources of funding, banks or other financial institutions sell assets to raise liquidity to meet their maturing obligations. With illiquid asset markets, these assets sales can trigger a sharp decline in asset prices. Mark-to-market valuation, accounting and reporting requirements can cause capital ratios to fall below critical levels in other institutions, or may prompt margin calls. This prompts further asset sales that can turn the asset price decline into a collapse. Although these vicious circles can occur even in the absence of mark-to-market or fair value accounting and reporting, the adoption of such rules undoubtedly exacerbates the problem. The procyclicality of the Basel requirements (and especially of Basel II) (which began to be introduced just around the time the crisis erupted) had, of course, been noted before (see e.g. Borio, Furfine and Lowe, 2001; Goodhart, 2004; Kashyap and Stein, 2004; and Gordy and Howells, 2004). II.3a Funding liquidity, the relationships-oriented model of intermediation and the lender of last resort Funding liquidity is central to the traditional “relationships-oriented” model (ROM) of financial capitalism and the traditional lender of last resort (LLR) role of the central bank. In the traditional banking model, banks fund themselves through deposits (fixed market value claims withdrawable on demand and subject to a sequential service constraint—first come, first served). On the asset side of the balance sheet the traditional bank holds a small amount of liquid reserves, but mainly illiquid assets—loans to households or to businesses, partly

08 Book.indb 512

2/13/09 3:59:00 PM

Central Banks and Financial Crises

513

secured (mortgages), partly unsecured. In the ideal-type ROM bank, loans are held to maturity (e.g. the “originate to hold model” of mortgage finance). Even when loans mature, the borrowers tend to stay with the same bank for their future financial needs. Although deposits can be withdrawn on demand, depositors too tend to stick with the same bank, with which they often have a variety of other financial relations. The long-term relationships mitigate asymmetric information problems and permit the parties to invest in reputations and to build on trust. It inhibits risk-trading and makes entry difficult. This combination of very short-maturity liabilities and long-maturity, illiquid assets is vulnerable to speculative attacks—bank runs. Such runs can occur, and be individually rational, even though the bank is solvent, in the sense that the value of the assets, if held to maturity, would be sufficient to pay off the depositors (and any other creditors). If the assets have to be liquidated prior to maturity, they would, however, be worthless (in milder versions the assets would be sold at a hefty discount on their fair value) and not all depositors would be made whole. This has been known since deposit-taking banks were first created. It has been formalised for instance in Diamond and Dybvig’s famous paper (Diamond and Dybvig, 1983, see also Diamond, 2007). There are typically two equilibria. One equilibrium has no run on the bank. No depositor withdraws his deposits; this is because he believes that total withdrawals will not exceed the liquid reserves of the banks. This is confirmed in equilibrium. The other equilibrium has a run on the bank. Each depositor tries to withdraw his deposit because he believes that the withdrawals by other depositors will exceed the bank’s liquid reserves. The bank fails. Solutions to this problem take the form of deposit insurance, standstills (mandatory bank holidays until the run subsides) and lender of last resort (LLR) intervention. All three require state intervention. Private deposit insurance can only cope with runs on individual banks or on a subset of the banks. It cannot handle a run on all banks. A creditor (depositor) standstill—making it impossible to withdraw deposits—could be part of the deposit contract, to be invoked at the discretion of the bank. This would, however, create

08 Book.indb 513

2/13/09 3:59:00 PM

514

Willem H. Buiter

rather serious moral hazard and adverse selection problems, so a bank regulator/supervisor would be a more plausible party to which to delegate the authority to suspend the right to withdraw deposits. Lending to a single troubled bank can be and has been provided by other banks. Again this cannot work if a sufficiently large number of banks are faced with a run. Individually rational bank runs don’t require that bank’s liabilities be deposits. They are possible whenever funding sources are shortterm and assets are of longer maturity and illiquid. When creditors to a bank refuse to renew maturing loans or credit lines, this is economically equivalent to a withdrawal of deposits. This applies to credit obtained in the interbank market or funds obtained by issuing debt instruments in the capital markets. Lending to a solvent but illiquid bank to prevent a socially costly bank failure should satisfy Bagehot’s dictum, which can be paraphrased as: Lend freely, against collateral that will be good in the long run (even if it is not good today), and at a penalty rate (Bagehot, 1873). Taking collateral and charging a penalty rate is part of the LLR rule book to avoid skewing incentives towards future excessive risk taking in lending and funding by the banks, that is, to avoid moral hazard. The discount window is an example of a LLR facility (in the case of the Fed I will mean by this the primary discount window, in the case of the ECB the marginal lending facility and in the case of the BoE the standing lending facility). The effective operation of LLR facility requires that the central bank determine all of the following: 1. The maturities of the loans and the total quantity of liquidity to be made available at each maturity. 2. The nature of the liquidity provided (e.g. central bank reserves or Treasury bills). 3. The interest rates charged on the loans and the other financial terms of the loan contract.

08 Book.indb 514

2/13/09 3:59:00 PM

Central Banks and Financial Crises

515

4. The set of eligible counterparties (who has access to the LLR facility?). 5. The regulatory requirements imposed on the eligible counterparties. 6. Whether the loan is collateralised or unsecured. 7. The set of financial instruments eligible as collateral. 8. The valuation of the collateral when there is no appropriate market price (when the collateral is illiquid). 9. Any further haircut (discount) applied to the valuation of the collateral and any other fees or financial charges imposed on the collateral. Items (3), (5), (8) and (9) jointly determine the cost to the borrower of access to the LLR facility, and thus the moral hazard created by the arrangement. In the case of the discount window (which can be described as an LLR facility “lite”), once points (1) to (9) have been determined, access to the facility is at the discretion of the borrower, that is, discount window borrowing is demand-driven. Strangely, and rather unfortunately, use of discount window facilities has become stigmatised in both the US and the UK. I assume the same applies to use of the discount window facilities of the Eurosystem, but I have less directly relevant information for this case. This stigmatisation of the use of the discount window may be individually rational, because a would-be discount window borrower could reasonably fear that future access to private sources of funding might be compromised if use of the discount window were seen as a signal that the borrower is in trouble. While this would be an unfortunate equilibrium, it is unlikely to be a fatal problem for a fearful discount window borrower: As long as the illiquid institution has a sufficient quantity of good collateral to be able to survive by using discount window funding (or through access to market-maker-of-last-resort facilities, discussed below in Section II.3b), discount window stigmatisation should not be a matter of corporate life or death. LLR facilities other than the discount window tend not to be “on demand.” They often involve borrowers whose solvency the central

08 Book.indb 515

2/13/09 3:59:00 PM

516

Willem H. Buiter

bank is not fully confident of. Such ad-hoc LLR facilities typically accept a wider range of collateral than the discount window, and the use of the facility is subject to bilateral negotiation between the wouldbe borrower(s) and the central bank. The Treasury and the regulator, if this is not the central bank, may also be involved (this was the case with the LLR facility arranged by the BoE for Northern Rock in September 2007—the Liquidity Support Facility). Such ad-hoc LLR arrangements are often arranged in secret and kept confidential as long as possible. Even after the fact, when commercial confidentiality concerns no longer apply, the information needed to determine whether the LLR (and the Treasury) made proper use of public funds in rescue operations are often not made public. The terms on which deposit insurance was made available to Northern Rock by the UK Treasury and the terms on which Northern Rock could access the Liquidity Support Facility created by the BoE are still not in the public domain. There is no justification for such secrecy. The LLR facilities (including the discount window) are only there to address liquidity issues, not solvency problems. Of course, future solvency is a probabilistic concept, not a binary one. When continued solvency is in question (discussed below in Section II.6), the central bank may be a party to a public-sector rescue and recapitalisation. The arrangement through which public resources are made available may well look like an LLR facility “on steroids.” The key difference with the regular LLR facility is that the resources made available through a normal LLR facility are not meant to be provided on terms that involve a subsidy to the borrower, its owners or its creditors. The risk-adjusted rate of return to the central bank on its LLR loans should cover its funding cost, essentially the interest rate on sovereign debt instruments of the relevant maturity. In a funding liquidity crisis, there is likely to be a wedge between the risk-adjusted cost of funds to the central bank and the (prohibitive) cost of obtaining funding from private sources. Under these conditions the central bank can provide liquidity to a borrower on terms that make it both subsidy-free (or even profitable ex-ante for the central bank) and cheaper than what the liquidity-constrained borrower could obtain elsewhere. Such actions correct a market failure.

08 Book.indb 516

2/13/09 3:59:00 PM

Central Banks and Financial Crises

517

In the case of the UK, the discount window (the standing lending facility) is highly restrictive in the maturity of its loans (overnight only) and in the collateral it accepts (only sovereign and supranational securities, issued by an issuer rated Aa3 [on Moody’s scale] or higher by two or more of the ratings agencies [Moody’s, Standard and Poor’s, and Fitch].6 The UK discount window therefore does not provide liquidity in any meaningful sense. It provides overnight liquidity in exchange for longer-term liquidity. It is of use only to banks that are caught short at the end of the trading day because of some technical glitch. Because the BoE has no discount window in the normal sense of the word, it had to create one when Northern Rock, a private commercial bank engaged mainly in home lending, found itself faced with both market liquidity and funding liquidity problems in September 2007. The resulting construct, the Liquidity Support Facility, is just what a normal discount window ought to have been, and is in the US and the euro area. Most central banks make, under special circumstances, unsecured loans to eligible counterparties as part of their LLR role, but these tend to be separate from the discount window. Also, as regards (2), discount window loans tend to be in exchange for central bank liquidity (reserves) rather than some other highly liquid instrument like Treasury bills. With the longer-maturity (up to 3 months) discount window loans that are now available in the US (for eligible deposit-taking banks), there is, in principle, no reason why the Fed should not make TBs or Federal Reserve Bills (non-monetary liabilities of the Fed) available at the discount window. It certainly could make such non-reserve liquidity available at LLR facilities other than the discount window. If a central bank engages in LLR loans to a solvent but illiquid bank, the central bank should expect to end up making a profit. It can extract this rent because the central bank is the only entity that is never illiquid (as regards domestic-currency obligations). It can always afford to hold good but illiquid assets till maturity. If the collateral offered is risky (specifically, subject to credit or default risk), the central bank can ex post make a loss even if it ex ante prices risky assets

08 Book.indb 517

2/13/09 3:59:00 PM

518

Willem H. Buiter

to properly reflect the risk of both the borrowing bank defaulting and the issuer of the collateral defaulting. I believe it is essential for a clear division of responsibilities between the central bank and the Treasury, and for proper public accountability for the use of public funds (to Congress/Parliament and to the electorate), that any such losses be made good immediately by the Treasury. Ideally, all collateral offered to the central bank other than sovereign instruments should be exchanged immediately with the Treasury for sovereign debt instruments, at the valuation put on that collateral in the LLR transaction. This removes the risk that the central bank is (ab)used as a quasi-fiscal agent of the government. To avoid regulatory arbitrage, any institutions eligible to access the discount window or any of the other LLR facilities of the central bank should be subject to a uniform regulatory regime. A special and key feature of such a common regulatory regime ought to be that access to LLR facilities only be granted to financial institutions for which there is a Special Resolution Regime which provides for Prompt Corrective Action and which establishes criteria under which the central bank, or a public agency working closely with the central bank like the FDIC, can declare a financial institution to be regulatorily insolvent before balance sheet insolvency or funding/liquidity insolvency can be established. The SRR managed by the FDIC for federally insured deposit-taking banks is a model. The SRR would allow a public administrator to be appointed who can take over the management of the institution, dismiss the board and the management, suspend the voting rights of the shareholders, place the shareholders at the back of the queue of claimants to the value that can be realised by the administrator, transfer (part of ) its assets or liabilities to other parties etc. Outright nationalisation would also have to be an option. The need for such an SRR for all institutions eligible to access LLR facilities follows from the fact that it is impossible for the central bank to determine whether a would-be user of the LLR facility is merely illiquid or both illiquid and insolvent. Without the SRR, the existence of the LLR facility would encourage quasi-fiscal abuse of

08 Book.indb 518

2/13/09 3:59:00 PM

Central Banks and Financial Crises

519

the central bank and would become a source of adverse selection and moral hazard. II.3b Market liquidity, the transactions-oriented model of intermediation and the market maker of last resort The defining feature of the financial crisis that started on August 9, 2007 was not runs on banks or other financial institutions. A few of these did occur. Ignoring smaller regional and local banks, a classic depositors’ bank run brought down Northern Rock in the UK (a mortgage lending bank that funded itself 75 percent in the wholesale markets), and non-deposit creditor runs were instrumental in killing off Bearn Stearns, a US investment bank and primary dealer, and IndyMac, a large US mortgage lending bank. These, however, were exceptional events. The new and defining feature of the crisis was the sudden and comprehensive closure of a whole range of financial wholesale markets, including the asset-backed commercial paper (ABCP) markets, the auction-rate securities (ARS) market, other asset–backed securities (ABS) markets, including the markets for residential mortgage-backed securities (RMBS), and many other collateralised debt obligations (CDO) and collateralised loan obligations (CLO) markets (see Buiter, 2007b, 2008b). The unsecured interbank market became illiquid to the point that Libor now is the rate at which banks won’t engage in unsecured lending to each other. The sudden increase in Libor rates at the beginning of August 2007 and the continuation of spreads over the overnight indexed swap (OIS) rate is shown for three-month Libor, historically an important benchmark, in Chart 4.7 The fact that the Libor-OIS spreads look rather similar for the three monetary authorities (with the obvious exception of a few idiosyncratic early spikes upwards in the sterling spread, reflecting the BoE’s late and belated conversion to the market-maker-of-last-resort cause) does not mean that all three did equally well in addressing the liquidity crunch in their jurisdiction. First, the magnitude of the challenge faced by each of the three may not have been the same. Second, the spreads are rather less interesting than the volumes of lending and borrowing that actually take place at these spreads. A 90-basis points

08 Book.indb 519

2/13/09 3:59:00 PM

520

1.40

Willem H. Buiter

Chart 4 Three-Month Libor-OIS spreads 11/02/2006-08/02/2008

Percent

1.40

1.20

1.20 UK Euro Area USA

1.00

1.00

8/2/08

7/2/08

6/2/08

5/2/08

4/2/08

3/2/08

2/2/08

1/2/08

12/2/07

11/2/07

9/2/07

10/2/07

-0.20

8/2/07

0.00 7/2/07

0.00

6/2/07

0.20

5/2/07

0.20

4/2/07

0.40

3/2/07

0.40

2/2/07

0.60

1/2/07

0.60

12/2/06

0.80

11/2/06

0.80

-0.20

spread with an active market is much less of a problem than a 90-basis points spread at which noone transacts. Unfortunately, turnover data for the interbank markets are not in the public domain. Third and most important, international financial integration ensures that liquidity can leak on a large scale between the jurisdictions of the national central banks, as long as the foreign exchange markets remain liquid, as they did for the major currencies. Unlike foreign branches, foreign subsidiaries of internationally active banks tend to have full access to the discount windows of their host central banks and they often also are eligible counterparties in the repos and other open market operations of their host central banks. Subsidiaries of UK banks made use of Eurosystem and Fed liquidity facilities. Indeed UK parents used their euro area subsidiaries to obtain liquidity for themselves. At least one subsidiary of a Swiss bank accessed the Fed’s discount window. Icelandic banks used their euro area subsidiaries to obtain euro liquidity, etc. In August 2008, Nationwide, a UK mortgage lender, announced it was setting up an Irish subsidiary. Gaining access to Eurosystem liquidity, both at the discount window and as a counterparty in repos, was a key motivating factor in this decision.

08 Book.indb 520

2/13/09 3:59:00 PM

Central Banks and Financial Crises

521

The de facto closure of many systemically important wholesale markets continues even now, a year since the start of the crisis. Overthe-counter credit default swap (CDS) markets and exchange-traded CDS derivatives markets became disorderly, with spreads far exceeding any reasonable estimate of default risk; key players in the insurance of credit risk, the so-called Monolines, lost their triple-A ratings and became irrelevant to the functioning of these markets. The rating agencies, which had moved aggressively from rating sovereigns and large corporates into the much more lucrative business of rating complex structured products (as well as advising on the design of such instruments), lost all credibility in these new product lines. This underlines the fact that the minimum shared understanding and information required for organised markets to function no longer existed for many structured products. One example: In the year since August 2007 there have been just two new issues of RMBS in the UK (one by HBOS for £500 million in May 2008, one by Alliance & Leicester for £400 million in August 2008).8 Central banks (outside the UK), in principle had the tools to address failing systemically important institutions—the LLR facilities. They did not have the tools to address failing, disorderly and illiquid markets. Central banks had developed and honed their skills during the era of traditional relationships-oriented financial intermediation centred on deposit-taking banks. Most were not prepared, institutionally and in mindset, to deal with the increasingly transactionsoriented financial intermediation that characterises modern financial sectors, especially in the US and the UK. Fortunately, all that was required to meet the new reality were a number of extensions to and developments of existing open market operations, specifically in relation to the sale and repurchase operations (repos) used by central banks to engage in collateralised lending. The main extensions were: larger transactions volumes, longer maturities, a broader range of counterparties and a wider set of eligible collateral, including illiquid private securities. Increased scale and scope for outright purchases of securities by central banks, which could have been part of the new model, have not (yet) been used.

08 Book.indb 521

2/13/09 3:59:00 PM

522

Willem H. Buiter

Central banks learnt fast to increase the scale and scope of their market-supporting operations. Unfortunately, the Fed did not sufficiently heed Bagehot’s admonition to provide liquidity only at a penalty rate. The ECB is also likely to have created, through its acceptance of illiquid collateral at excessively generous valuations, adverse incentives for excessive future risk-taking. The ECB has not provided the information required to confirm or deny the suspicions about its collateral facilities. The BoE, on the basis of the limited available information, is the least likely of the three central banks to have over-priced the illiquid collateral it has been offered. Even here, however, the hard information required for proper accountability has not been provided. Not designing the financial incentives faced by their counterparties in these new facilities to minimize moral hazard has turned out to be the central banks’ Achilles heel in the current crisis. It will come back to haunt us in the next crisis. Modern financial systems tend to be a convex combination of the traditional ROM and the transactions-oriented model of financial capitalism (TOM). The TOM (also called arms-length model or capital markets model) commoditises financial interactions and relationships and trades the resulting financial instruments in OTC markets or in organised exchanges. Securitisation of mortgages is an example. This makes the illiquid liquid and the non-tradable tradable. Scope for risk-trading is greatly enhanced. This is, potentially, good news. It also destroys information. In the “originate-and-hold” model, the originator of the illiquid individual loan works for the Principal; he works as Agent of the Principal in the “originate-to-distribute” model. This reduces the incentive to collect information on the creditworthiness of the ultimate borrower and to monitor the performance of the borrower over the life of the loan. Securitisation and resale then misplace whatever information is collected: After a couple of transactions in [RMBS], neither the buyer nor the seller has any idea about the creditworthiness of the underlying assets. This is the bad news. Inappropriate securitisation permitted, indeed encouraged, the subversion of ordinary bank lending standards that was an essential input in the subprime disaster in the US.

08 Book.indb 522

2/13/09 3:59:00 PM

Central Banks and Financial Crises

523

The TOM affects banks in two ways. First, it provides competition for banks as intermediaries, since non-financial corporates can issue securities in the capital markets instead of borrowing from the banks, thus potentially bypassing banks completely. Savers can buy these securities as alternatives to deposits or other forms of credit to banks. Second, banks turn their illiquid assets into liquid assets which they either sell on (to special purpose vehicles [SPVs] set up to warehouse RMBS, or to investors) or hold on their balance sheet in the expectation that they can be sold at short notice and at a predictable price close to fair value, i.e. that they are liquid. It may seem that this commoditisation and marketisation of financial relationships that are the essence of the TOM would solve the banks’ liquidity problem and would make even bank runs non-threatening. If the bank’s assets can be sold in liquid markets, the cost of a deposit run or a “strike” by other creditors need not be a fatal blow. Unfortunately, the liquidity of markets is not a deep, structural characteristic, but the endogenous outcome of the interaction of many partially and poorly informed would-be buyers and sellers. Market liquidity can vanish at short notice, just like funding liquidity. Bank runs have their analogue in the TOM world in the form of a market freeze, run, strike, seizure or paralysis (the terminology is not settled yet). A potential buyer of a security who has liquid resources available today, may refuse to buy the security (or accept it as collateral), even though he believes that the security has been issued by a solvent entity and will earn an appropriate risk-adjusted rate of return if held to maturity. This socially excessive hoarding of scarce liquid assets can be individually rational because the potential buyer believes that he may be illiquid in the next trading period (and may therefore have to sell the security next period), and that other potential buyers of the security may likewise be illiquid in the future or may strategically refuse to buy the security, to gain a competitive advantage or even to put him out of business. If the transaction is a repo, he would have to believe also that the party trying to sell the security to him today, may be illiquid in the future and unable to make good on his commitment.

08 Book.indb 523

2/13/09 3:59:00 PM

524

Willem H. Buiter

It remains an open question whether this approach to market and funding illiquidity today as a result of fear of market and funding illiquidity tomorrow either needs to be iterated ad infinitum or requires a fear of insolvency at some future date to support a full-fledged individually rational but socially inefficient equilibrium. Charles Goodhart (2002) believes that without the threat of insolvency there can be no illiquidity (see also the excellent collection of readings in Goodhart and Illing, 2002). Strategic behaviour, Knightian uncertainty, bounded rationality and other behavioural economics approaches to modelling the transactions flows in financial markets, including the rules-of-thumb that lead to information cascades and herding behaviour, may offer a better chance of understanding, predicting and correcting the market pathologies that lead to socially destructive hoarding of liquidity than relentlessly optimising models. The jury is still out on this one.9 Market illiquidity addresses the phenomenon that a financial instrument that is traded abundantly one day suddenly finds no buyers the next day at any price, or only at a price that represents a massive discount relative to its fundamental or fair value. That is, illiquidity is an endogenous outcome, a dysfunctional equilibrium in a market or game for which alternative liquid equilibria also exist, but have not materialised (or have not been coordinated on). Market illiquidity is a form of market failure. Liquidity can be provided privately, by banks and other economic agents holding large amounts of inherently liquid assets (like central bank reserves or TBs). That would, however, be socially and privately inefficient. Maturity transformation and liquidity transformation are essential functions of financial intermediaries. A private financial entity should hold (or have access to, through credit lines, swaps, etc.) enough liquidity to manage its business during normal times, that is, when markets are liquid and orderly. It should not be expected to hoard enough liquid assets (or arrange liquid stand-by funding) during normal times to be able to survive on its own during abnormal times, when markets are disorderly and illiquid. That is what central banks are for. Central banks can create any amount of domestic currency liquidity at little or no notice and at effectively zero marginal cost.

08 Book.indb 524

2/13/09 3:59:00 PM

Central Banks and Financial Crises

525

It would be inefficient to privatise and decentralise the provision of emergency liquidity when there is an abundant source of free liquidity readily available. Anne Sibert and I (Buiter and Sibert, 2007a,b, see also Buiter, 2007a,b,c,d) have called the role of the central bank as provider of market liquidity during times when systemically important financial markets have become disorderly and illiquid, that of the market maker of last resort (MMLR). The central bank as market maker of last resort either buys outright (through open market purchases) or accepts as collateral in repos and similar secured transactions, systemically important financial instruments that have become illiquid.10 If no market price exists to value the illiquid securities, the central bank organises reverse auctions that act as value discovery mechanisms. There is no need for the central bank to know more about the value of the securities than the sellers, or indeed for the central bank to know anything at all. The central bank should organise the auction because it has the liquid “deep pockets.” A reverse Dutch auction, for instance, would be likely to be particularly punitive for the sellers of the illiquid securities. A second-lowest price (sealed bid) reverse auction would have other attractive properties. With so many Nobel-prizes and Nobel-prize calibre economists specialised in mechanism design, I don’t think the expertise to design and run these auctions would be hard to find. The auctions to value the illiquid securities could be organised jointly by the central bank and the Treasury if, as I advocate, the Treasury would immediately take onto its balance sheet any illiquid assets acquired in the auctions, either outright or as part of a repo or swap. For the MMLR to function effectively, the central bank has to clarify all of the following: 1. The list of eligible instruments for outright purchase or for use in collateralised transactions like repos. 2. The nature of the liquidity provided (e.g. central bank reserves or Treasury bills). 3. The set of eligible counterparties.

08 Book.indb 525

2/13/09 3:59:01 PM

526

Willem H. Buiter

4. The regulatory requirements imposed on the eligible counterparties. 5. The valuation of the securities offered for outright purchase or as collateral, when there is no appropriate market price (when the collateral is illiquid). 6. Any haircut (discount) applied to the valuation of the securities and any other fees or financial charges imposed. Items (4), (5) and (6) determine the effective penalty imposed by the MMLR for use of its facilities, and thus the severity of the moral hazard created by its existence. Unlike discount window access, which is at the initiative of the borrower, MMLR finance is not available on demand, even if (1) through (6) above have been determined. The policy authority (in practice the central bank) decides when to inject liquidity, on what scale and at what maturity. Injecting large amounts of liquidity against illiquid collateral is easy. The key challenge for the central bank as market maker of last resort is the same as that faced by the central bank as lender of last resort. It is to make the effective performance of the MMLR function during abnormal times, that is, when markets are disorderly and illiquid, compatible with providing the right incentives for risk taking when markets are orderly and liquid. This requires liquidity to be made available only on terms that are punitive. It is here that all three central banks appear to have failed so far, albeit in varying degrees. II.4 The lender of last resort and market maker of last resort when foreign currency liquidity is the problem So far, the argument has proceeded on the assumption that the central bank can provide the necessary liquidity effectively costlessly and at little or no notice. That, however, is true only for domestic-currency liquidity. For countries that have banks and other financial institutions that are internationally active and have significant amounts of foreign-currency-denominated exposure, a domestic-currency LLR and MMLR may not be sufficient. This is especially likely to be an

08 Book.indb 526

2/13/09 3:59:01 PM

Central Banks and Financial Crises

527

issue if the country’s banks or other systemically significant financial businesses have large short-maturity foreign currency liabilities and illiquid foreign currency assets. The example of Iceland comes to mind as do, to a lesser extent, Switzerland and the UK. If the country in question has a domestic currency that is also a serious global reserve currency, the central bank is likely to be able to arrange swaps or credit lines with other central banks on a scale sufficient to enable it to act as a foreign-currency LLR and MMLR for its banking sector. At the moment there are only two serious global reserve currencies, the US dollar, with 63.3 percent of estimated global official foreign exchange reserves at the end of 2007, and the euro, with 26.5 percent (see Table 1). Sterling is a minor-league legacy global reserve currency with 4.7 percent, the yen is fading fast at 2.9 percent and Switzerland is a minute 0.2 percent.11 The Fed, the ECB and the Swiss National Bank have created swap lines of US dollars for euro and Swiss francs respectively since the crisis started. These swap arrangements have recently been extended to cover the 2008 year-end period. The Central Bank of Iceland arranged, in May 2008, swap lines for €500mn each with the central banks of Norway, Denmark and Sweden. In the case of Iceland, one can see how such currency swaps could be useful in the discharge of the Central Bank of Iceland’s LLR and MMLR function vis-à-vis a banking system with a large stock of short-maturity foreign currency liabilities and illiquid foreign currency assets. The swaps between the Fed, the ECB and the SNB are less easily rationalised. Both the euro area- and the Switzerland-domiciled banks experienced a shortfall of liquidity of any and all kinds, not a specific shortage of US dollar liquidity. The foreign exchange markets had not seized up and become illiquid. Certainly, it was expensive for euro-area resident banks with maturing US dollar obligations to obtain US dollar liquidity through the swap markets, but that is no reason for official intervention (or ought not to be): Expensive is not the same as illiquid. I therefore interpret these currency swap

08 Book.indb 527

2/13/09 3:59:01 PM

08 Book.indb 528

6.80%

2.40%

0.30%

Japanese yen

French franc

Swiss franc

11.70%

0.20%

1.80%

6.70%

2.70%

14.70%

62.10%

’96

10.20%

0.40%

1.40%

5.80%

2.60%

14.50%

65.20%

’97

6.10%

0.30%

1.60%

6.20%

2.70%

13.80%

69.30%

’98

1.60%

0.20%

6.40%

2.90%

17.90%

70.90%

’99

1.40%

0.30%

6.30%

2.80%

18.80%

70.50%

’00

1.20%

0.30%

5.20%

2.70%

19.80%

70.70%

’01

1.40%

0.40%

4.50%

2.90%

24.20%

66.50%

’02

1.90%

0.20%

4.10%

2.60%

25.30%

65.80%

’03

1.80%

0.20%

3.90%

3.30%

24.90%

65.90%

’04

1.90%

0.10%

3.70%

3.60%

24.30%

66.40%

’05

1.50%

0.20%

3.20%

4.20%

25.20%

65.70%

’06

1.80%

0.20%

2.90%

4.70%

26.50%

63.30%

’07

Source: Wikipedia

Sources:1995-1999, 2006-2007 IMF: Currency Composition of Official Foreign Exchange Reserves; 1999-2005, ECB: The Accumulation of Foreign Reserves

13.60%

2.10%

Pound sterling

Other

15.80%

59.00%

German mark

Euro

US dollar

’95

Table 1 Currency composition of official foreign exchange reserves

528 Willem H. Buiter

2/13/09 3:59:01 PM

Central Banks and Financial Crises

529

arrangements (unlike the swap arrangements put in place following 9/11) either as symbolic tokens of international cooperation (and more motion than action) or as unwarranted subsidies to euro areaand Switzerland-based banks needing US dollar liquidity. II.5 Macroeconomic stabilisation and liquidity management: Interdependence and institutional arrangements Macroeconomic stabilisation policy and liquidity management (including the LLR and MMLR arrangements and policies) cannot be logically or analytically separated or disentangled completely. Changes in the official policy rate affect output, employment and inflation, but also have an effect on funding liquidity and market liquidity. An artificially low official policy rate can boost bank profitability and help banks to recapitalise themselves. The current level of the federal funds target rate certainly has this effect. Discount window operations, repos, other open market purchases and indeed the whole panoply of LLR and MMLR arrangements and interventions strengthen the financial system, even for a given contingent sequence of current and future official policy rates. This boosts aggregate demand and thus influences growth and inflation. Nevertheless, I believe that the official policy rate has a clear comparative advantage as a macroeconomic stabilisation tool while liquidity management has a corresponding comparative advantage as a financial stabilisation tool. Mundell’s principle of effective market classification (policies should be paired with the objectives on which they have the most influence) therefore suggests that, should we wish to assign each of these instruments to a particular target, the official policy rate be assigned to macroeconomic stability and liquidity management to financial stability (see Mundell, 1962). Both the ECB and the BoE advocate the view that the official policy rate be assigned to the macroeconomic stability objective (for both central banks this is the price stability objective) and that it not be used to pursue financial stability objectives. Any impact of the official policy rate on financial stability will, in that view, have to be reflected in an appropriate adjustment in the scale and scope of

08 Book.indb 529

2/13/09 3:59:01 PM

530

Willem H. Buiter

liquidity management policies. Likewise, liquidity management policies (that is, LLR and MMLR actions) should be targeted at financial stability without undue concern for the impact they may have on price stability and economic activity. If these effects (which are highly uncertain) turn out to be material, there will have to be an appropriate response in the contingent sequence of official policy rates. Undoubtedly, to the unbridled dynamic stochastic optimiser, the joint pursuit of all objectives with all instruments has to dominate the assignment of the official policy rate to macroeconomic stability and of liquidity management to financial stability. I am with Mundell on this issue, partly because it makes both communication with the markets and accountability to Parliament/Congress and the electorate easier. A case can even be made for taking the setting of the official policy rate out of the central bank completely. Obviously, as the source of ultimate domestic-currency liquidity, the central bank is the only agency that can manage liquidity. It will also have to implement the official policy rate decision, through appropriate money market actions. But it does not have to make the official policy rate decision. The knowledge, skills and personal qualities for setting the official policy rate would seem to be sufficiently different from those required for effective liquidity management, that assigning both tasks to the same body or housing them in the same institution is not at all self-evident. In the UK, the institutional setting is ready-made for taking the Monetary Policy Committee out of the BoE. The Governor of the BoE could be a member, or even the chair of the MPC, but need not be either. The existing institutional arrangements in the US and the euro area would have to be modified significantly if the official policy rate decision were to be moved outside the central bank. Through its liquidity management role and more generally through its LLR and MMLR functions, the central bank will inevitably play something of a de facto supervisory and regulatory role vis-à-vis banks and other counterparties. Regulatory capture is therefore a constant threat and a frequent reality, as the case of the Fed, discussed

08 Book.indb 530

2/13/09 3:59:01 PM

Central Banks and Financial Crises

531

below in Section III.2a(xii) makes clear. Moving the official policy rate decision out of the central bank would make it less likely that the official policy rate would display the kind of excess sensitivity to financial sector concerns displayed by the federal funds target rate since Chairman Greenspan.12 Regardless of whether the official policy rate-setting decision is taken out of the central bank, I consider it desirable that all three central banks change their procedures for setting the overnight rate. Chart 5 shows the spread between overnight Libor (an unsecured rate) and the official policy rate for the three central banks. Similar pictures could be shown for the spread between the effective federal funds rate and the federal funds target rate and for spreads between the sterling and euro secured overnight rates and official policy rates. The fact that the central banks are incapable of keeping the overnight rate close to the official policy rate is a direct result of the operating procedures in the overnight money markets (see Bank of England, 2008a, and Clews, 2005; European Central Bank, 2006; and Federal Reserve System, 2002). Setting the official policy rate (like fixing any price or rate) ought to mean that the central bank is willing to lend reserves (against suitable collateral) on demand in any amount and at any time at that rate, and that it is willing to accept deposits in any amount and at any time at that rate. This would effectively peg the secured overnight lending and borrowing rate at the official policy rate. The overnight interbank rate could still depart from the official policy rate because of bank default risk on overnight unsecured loans, but that spread should be trivial almost always. Ideally, there would be a 24/7 fixed rate tender at the official policy rate during a maintenance period, and a 24/7 unlimited deposit facility at the official policy rate. The deviations between the official policy rate and the overnight interbank rate that we observe for the Fed, the ECB and the BoE are the result of bizarre operating procedures—the vain pursuit by the central bank of the pipe dream of setting the price (the official policy rate)

08 Book.indb 531

2/13/09 3:59:01 PM

532

Willem H. Buiter

Chart 5 Spreads between effective overnight money market rates and official policy rates (percent) 01/02/2006 - 07/14/2008 1.50

Percent

1.50 U.K. Eonia U.S.

20080714

20080612

20080513

20080411

20080312

20080211

20080110

20071211

20071109

20071010

20070910

20070809

20070710

20070608

20070509

20070409

20070308

20070206

20070105

20061206

20061106

20061005

20060905

20060804

20060705

-0.50

20060605

0.00 20060504

0.00 20060404

0.50

20060303

0.50

20060201

1.00

20060102

1.00

-0.50

-1.00

-1.00

-1.50

-1.50

while imposing certain restrictions on the quantity (the reserves of the banking system and/or the amount of overnight liquidity provided).13 In the case of the UK, for instance, the commercial banks and other deposit-taking institutions that are eligible counterparties in repos specify their planned reserve holdings just prior to a new reserve maintenance period (roughly the period between two successive scheduled MPC meetings). Those reserves earn the official policy rate. If actual reserves (averaged over the maintenance period) exceed the planned amount, the interest rate received by the banks on the excess is at the standing deposit facility rate, 100 basis points below the official policy rate. If banks’ estimated reserves turn out to be insufficient and the banks have to borrow from the BoE to meet their liquidity needs, they have to do so at the standing lending facility rate, 100 basis points above the official policy rate, except on the last day of the maintenance period, when the penalty falls to 25 basis points. Compared to simply pegging the rate, the BoE’s operating procedure is an example of making complicated something that really is very simple: Setting a rate means supplying any amount demanded at that rate and accepting any amount offered at that rate. The Bank of Canada’s

08 Book.indb 532

2/13/09 3:59:01 PM

Central Banks and Financial Crises

533

operating procedures for setting the overnight rate are closer to my ideal rate-setting mechanism (Bank of Canada, 2008). If the central banks were to fix the overnight rate in the way I suggest, this would probably kill off the secured overnight interbank market, although not necessarily the unsecured overnight interbank market (overnight Libor), and certainly not the longer-maturity interbank markets, secured and unsecured. The loss of the secured overnight market would not represent a social loss: It is redundant. Those who used to operate in it, now can engage in more socially productive labour. There is no right to life for redundant markets. If the prospect of killing the secured overnight market is too frightening, central banks could adjust the proposed procedure by lending any amount overnight (against good collateral) at the official policy rate plus a small margin, and accepting overnight deposits in any amount at the official policy rate minus a small margin; twice the margin would just exceed the normal bid-ask spread in the secured private overnight interbank markets. It does not help communication with the markets, or the division of a labour between interest rate policy and liquidity policy, if the monetary authority sets an official policy rate but there is no actual market rate, that is, no rate at which transactions actually take place, that corresponds to the official policy rate. Fortunately, the remedy is simple. II.6 Central banks as quasi-fiscal agents: Recapitalising insolvent banks Whatever its legal or de facto degree of operational and goal independence, the central bank is part of the state and subject to the authority of the sovereign. Specifically, the state (through the Treasury) can tax the central bank, even if these taxes may have unusual names. In many countries, the Treasury owns the central bank. This is the case, for instance, in the UK, but not in the US or the euro area. As an agent and agency of the state, the central bank can engage in quasi-fiscal actions, that is, actions that are economically equivalent to levying taxes, paying subsidies or engaging in redistribution. Examples are non-remunerated reserve requirements (a quasi-fiscal tax

08 Book.indb 533

2/13/09 3:59:01 PM

534

Willem H. Buiter

on banks), loans to the private sector at an interest rate that does not at least cover the central bank’s risk-adjusted cost of non-monetary borrowing (a quasi-fiscal subsidy), accepting overvalued collateral (a quasi-fiscal subsidy) or outright purchases of securities at prices above fair value (a quasi-fiscal subsidy). To determine how the use of the central bank as a quasi-fiscal agent of the state affects its ability to pursue its macroeconomic stability objectives, a little accounting is in order. In what follows, I disaggregate the familiar “government budget constraint” into separate budget constraints for the central bank and the Treasury. I then derive the intertemporal budget constraints for the central bank and the Treasury, or their “comprehensive balance sheets.” I then contrast the familiar conventional balance sheet of the central bank with its comprehensive balance sheet. My stylised central bank has two financial liabilities: the non-interestbearing and irredeemable monetary base M > 0 and its interest-bearing non-monetary liabilities (central bank Bills), N > 0, paying the risk-free one-period domestic nominal interest rate i .14 On the asset side it has the stock of international foreign exchange reserves, R f, earning a risk-free nominal interest rate in terms of foreign currency, i f, and the stock of domestic credit, which consists of central bank holdings of nominal, interest-bearing Treasury bills, D > 0, earning a risk-free domestic-currency nominal interest rate i, and central bank claims on the private sector, L > 0 , with domestic-currency nominal interest rate iL. The stock of Treasury debt (all assumed to be denominated in domestic currency) held outside the central bank is B; it pays the risk-free nominal interest rate i; Tp is the real value of the tax payments by the domestic private sector to the Treasury; it is a choice variable of the Treasury and can be positive or negative; Tb is the real value of taxes paid by the central bank to the Treasury; it is a choice variable of the Treasury and can be positive or negative; a negative value for Tb is a transfer from the Treasury to the central bank: The Treasury recapitalises the central bank; T=Tp+Tb is the real value of total Treasury tax receipts; P is the domestic general price level; e is the value of the spot nominal exchange rate (the domestic currency price of foreign exchange); Cg > 0 is the real value of Treasury

08 Book.indb 534

2/13/09 3:59:01 PM

Central Banks and Financial Crises

535

spending on goods and services and Cb > 0 the real value of central bank spending on goods and services. Public spending on goods and services is assumed to be for consumption only. Equation (3) is the period budget identity of the Treasury and equation (4) that of the central bank.  B + Dt −1  Bt + Dt ≡ Ctg − Tt p − Tt b + (1 + it )  t −1  Pt Pt  

(3)

M t + N t − Dt − Lt − et Rtf ≡ Ctb + Tt b Pt +

M t −1 − (1 + it )( Dt −1 − N t −1 ) − (1 + itL ) Lt −1 − (1 + itf )et Rtf−1 Pt (4)



The solvency constraints of, respectively, the Treasury and central bank are given in equations (5) and (6): lim Et I N ,t −1 ( BN + DN ) ≤ 0

N →∞



(5)



f lim Et I N ,t −1 ( DN + LN + eN RN − N N ) ≥ 0

N →∞

(6)

where It ,t is the appropriate nominal stochastic discount factor 1 0 between periods t0 and t1. These solvency constraints, which rule out Ponzi finance by both the Treasury and the central bank, imply the following intertemporal budget constraints for the Treasury (equation 7) and for the central bank (equation 8). ∞

Bt −1 + Dt −1 ≤ Et ∑ I j ,t −1Pj (T jp + T jb − C gj j =t

(7)15

∞

Dt −1 + Lt −1 + et −1 Rtf−1 − N t −1 ≤ Et ∑ I j ,t −1 ( Pj (C Jb + T jb + Q j ) − ∆M j ) j =t

(8)

where

08 Book.indb 535

2/13/09 3:59:03 PM

536

Willem H. Buiter

 e  Pj Q j  (i j − i jL ) L j−1 + 1 + i j − (1 + i jf ) j  e j−1 R jf−1 e j−1  

(9)

The expression Q in equation (9) stands for the real value of the quasi-fiscal implicit interest subsidies paid by the central bank. If the rate of return on government debt exceeds that on loans to the private sector, there is an implicit subsidy to the private sector equal in period t to (it − itL ) Lt −1. If the rate of return on foreign exchange reserves is less than what would be implied by Uncovered Interest Parity (UIP), there is an implicit subsidy to the issuers of these reserves,  et  f f  et −1 Rt −1 . given in period t by 1 + it − (1 + it ) e  t −1  When comparing the conventional balance sheet of the central bank to its comprehensive balance sheet or intertemporal budget constraint, it is helpful to rewrite (8) in the following equivalent form: M t −1 − ( Dt −1 + Lt −1 + et −1 Rtf−1 − N t −1 ) 1 + it ∞     i ≤ Et ∑ I j ,t −1 Pj (−C bj − T jb − Q j ) +  j+1  M j    1 + i j+1   j =t

(10)

Summing (3) and (4) gives the period budget identity of the government (the consolidated Treasury and central bank), in equation (11); summing (5) and (6) gives the solvency constraint of the government in equation (12) and summing (7) and (8) gives the intertemporal budget constraint of the government in equation (13). M t + N t + Bt − Lt − et Rtf ≡ Pt (Ctg + Ctb − Tt )

+ M t −1 + (1 + it )( Bt −1 + N t −1 ) − (1 + itL ) Lt −1 − et (1 + itf ) Rtf−1

lim Et I N ,t −1 ( BN + N N − LN − eN RNf ) ≤ 0

N →∞

∞

j =t

(

08 Book.indb 536

(12)



Bt −1 + N t −1 − Lt −1 − et −1 Rtf−1 ≤ Et ∑ I j ,t −1 Pj (T j − Q j − (C gj + C bj )) + ∆M j

(11)

)

(13)

2/13/09 3:59:05 PM

Central Banks and Financial Crises

537

Table 2 Central bank conventional financial balance sheet Assets

Liabilities

D

M 1+ i

L

N

eR f Wb

Consider the conventional financial balance sheet of the Central Bank in Table 2. The Central Bank’s conventional financial net worth or equity, W b  D + L + eR f − N −

M 1+ i ,

is the excess of the value of its financial assets (Treasury debt, D, loans to the private sector, L and foreign exchange reserves, eR f ) over its non-monetary liabilities N and its monetary liabilities M / (1+i ). On the left-hand side of (10) we have (minus) the conventionally measured equity of the central bank. On the right-hand side of (10) we can distinguish two terms. The first is ∞

−Et ∑ I j , t −1 Pj (C bj + T jb + Q j ) j =t

—the present discounted value of current and future primary (noninterest) surpluses of the central bank. Important for what follows, this contains both the present value of the sequence of current and future transfer payments made by the Treasury to the central bank ( {−T jb ; j ≥ t } and (with a negative sign) the present value of the sequence of quasi-fiscal subsidies paid by the central bank {Q j ;j>t }. ∞   i Et ∑ I j ,t −1  j+1  M j  1 + i j+1  ,one of the measures of cenj =t

The second terms is tral bank “seigniorage”—the present discounted value of the future interest payments saved by the central bank through its ability to

08 Book.indb 537

2/13/09 3:59:07 PM

538

Willem H. Buiter

issue non-interest-bearing monetary liabilities. The other conventional measure of seigniorage, motivated by equation (8), is the ∞

present discounted value of future base money issuance: Et ∑ I j ,t −1∆M j . j =t

Even if the conventionally defined net worth or equity of the central bank is negative, that is, if Wt b−1  Dt −1 + Lt −1 + et −1 Rtf−1 − N t −1 − the central bank can be solvent provided



∞ ∞   i Wt b−1 + Et ∑ I j ,t −1  j+1  M j ≥ Et ∑ I j .t −1 Pj (C bj + T jb + Q j )  1 + i j+1  j =t j =t

M t −1 < 0, 1 + it

(14)

Conventionally defined financial net worth or equity excludes the present value of anticipated or planned future non-contractual outlays and revenues (the right-hand side of equation 10). It is therefore perfectly possible for the central bank to survive and thrive with negative financial net worth. If there is a seigniorage Laffer curve, however, there always exists a sufficient negative value for central bank conventional net worth, that would require the central bank to raise  ∆M j  so much seigniorage in real terms,  P ; j ≥ t , or  j 

 i   j +1  M j ; j ≥ t    1 + i j+1  

through current and future nominal base money issuance, that, given the demand function for real base money, unacceptable rates of inflation would result (see Buiter, 2007e, 2008a). While the central bank can never go broke (that is, equation 14 will not be violated) as long as the financial obligations imposed on the central bank are domestic-currency denominated and not index-linked, it could go broke if either foreign currency obligations or index-linked obligations were excessive. I will ignore the possibility of central bank default in what follows, but not the risk of excessive inflation being necessary to secure solvency without recapitalisation by the Treasury, if the central bank’s conventional balance sheet were to take a sufficiently large hit. This situation can arise, for instance, if the central bank is used as a quasi-fiscal agent to such an extent that the present discounted value

08 Book.indb 538

2/13/09 3:59:08 PM

Central Banks and Financial Crises

539 ∞

of the quasi-fiscal subsidies it

provides, Et ∑ I j ,t −1 Pj Q j , j =t

is so large, that

its ability to achieve its inflation objectives is impaired. In that case (if we rule out default of the central bank on its own non-monetary obligations, Nt-1), the only way to reconcile central bank solvency and the achievement of the inflation objectives would be a recapitalisation of the central bank by the Treasury, that is, a sufficient large ∞

increase in

−Et ∑ I j ,t −1 Pj T jb j =t

.16

There are therefore in my view two reasons why the Fed, or any other central bank, should not act as a quasi-fiscal agent of the government, other than paying to the Treasury in taxes,Tb, the profits it makes in the pursuit of its macroeconomic stability objectives and its appropriate financial stability objectives. The appropriate financial stability objectives are those that involve providing liquidity, at a cost covering the central bank’s opportunity cost of non-monetary financing, to illiquid but solvent financial institutions. The two reasons are, first, that acting as a quasi-fiscal agent may impair the central bank’s ability to fulfil its macroeconomic stability mandate and, second, that it obscures responsibility and impedes accountability for what are in substance fiscal transfers. If the central bank allows itself to be used as an off-budget and off-balance-sheet special purpose vehicle of the Treasury, to hide contingent commitments and to disguise de facto fiscal subsidies, it undermines its independence and legitimacy and impairs political accountability for the use of public funds—“tax payers’ money.” II.6a Some interesting central bank balance sheets What do the conventional balance sheets look like in the case of the Fed, the BoE and the ECB/Eurosystem? The data for the Fed are summarised in Table 3, those for the BoE in Table 4, for the ECB in Table 5 and for the Eurosystem in Table 6. The data for the Fed are updated weekly in the Consolidated Statement of Condition of All Federal Reserve Banks. In Table 3, I have

08 Book.indb 539

2/13/09 3:59:09 PM

540

Willem H. Buiter

Table 3 Conventional financial balance sheet of the Federal Reserve System March12, 2008, US$ billion Assets

Liabilities

D: 703.4

M: 811.9

L: 182.2

N: 47.4

R: 13.0 W: 39.7

Table 4 Conventional balance sheet of the Bank of England (£ billion) Jun 1, 2006

Dec 24, 2007

Mar 12, 2008

82

102

97

Notes in circulation

38

45

41

Reserves balances

22

26

21

N:

Other

20

30

33

W b:

Equity

2

2

2

82

102

97

Liabilities M:

Assets D:

Advances to HM Government

13

13

7

L&D:

Securities acquired via market transactions

8

7

9

L:

Short-term market operations & reverse repos with BoE counterparties

12

44

43

Other assets

33

38

38

Source: Financial Statistics

08 Book.indb 540

2/13/09 3:59:09 PM

Central Banks and Financial Crises

541

Table 5 Conventional balance sheet of the European Central Bank (€ billion) Liabilities

December 31, 2006

December 31, 2007

106

126

Notes in circulation

50

54

N:

Other

56

72

Wb:

Equity

4

4

106

126

54

71

10

11

3

4

40

39

M:

Assets D: L:

Other Assets Claims on euro-area residents in forex

R:

Gold and forex reserves

Source: European Central Bank (2008a)

Table 6 Conventional balance sheet of the Eurosystem (€ billion) Liabilities M:

December 22, 2006

February 29, 2008

1142

1379

805

887

N:

Other

273

421

Wb:

Equity

64

71

1142

1379

40

39

Assets D:

Euro-denominated government debt

L:

Euro-denominated claims on euro-area credit institutions

452

519

Other Assets

330

480

Gold and forex reserves

321

340

R:

Source: European Central Bank (2008b)

08 Book.indb 541

2/13/09 3:59:09 PM

542

Willem H. Buiter

for simplicity lumped $2.1 billion worth of buildings and $40 billion worth of other assets together with claims on the private sector, L. The Federal Reserve System holds but small amounts of assets in the gold certificate account and SDR account as foreign exchange reserves, R. The foreign exchange reserves of the US are on the balance sheet of the Treasury rather than the Fed. As of February 2008, US Official Reserve Assets stood at $73.5 billion.17 US gold reserves (8133.8 tonnes) were valued at around $261.5 billion in March 2008. Table 3 shows that, as regards the size of its balance sheet, the Fed would be a medium-sized bank in the universe of internationally active US commercial banks, with assets of around $900 billion and capital (which corresponds roughly to financial net worth or conventional equity) of about $40 billion. By comparison, at the time of the run on the investment bank Bear Stearns (March 2008), that bank’s assets were around $340 billion. Citigroup’s assets as of 31 December 2007 were just under $2,188 billion (Citigroup is a universal bank, combining commercial banking and investment banking activities). With 2007 US GDP at around $14 trillion, the assets of the Fed are about 6.4% of annual US GDP. At the end of January 2008, seasonally adjusted assets of domestically chartered commercial banks in the US stood at $9.6 trillion (more than ten times the assets of the Fed). Of that total, credit market assets were around $7.5 trillion. Equity (assets minus all other liabilities) was reported as $1.1 trillion.18 Commercial banks exclude investment banks and other non-deposit taking banking institutions. The example of Bear Stearns has demonstrated that all the primary dealers in the US are now considered by the Fed and the Treasury to be too systemically important (that is too big, and/or too interconnected, to fail). The 1998 rescue of LTCM—admittedly without the use of any Fed financial resources or indeed of any public financial resources, but with the active “good offices” of the Fed—suggests that large hedge funds too may fall in the “too big or too interconnected to fail” category. We appear to have arrived at the point where any highly leveraged financial institution above a certain size is a candidate for direct or indirect Fed financial support, should it, for whatever reason, be at risk of failing.19

08 Book.indb 542

2/13/09 3:59:09 PM

Central Banks and Financial Crises

543

Like its private sector fellow-banks, the Fed is quite highly leveraged, with assets just under 22 times capital. The vast majority of its liabilities are currency in circulation ($781 billion out of a total monetary base of $812 billion). Currency is not just non-interest-bearing but also irredeemable: having a $10 Federal Reserve note gives me a claim on the Fed for $10 worth of Federal Reserve notes, possibly in different denominations, but nothing else. Leverage is therefore not an issue for this highly unusual inherently liquid domestic-currency borrower, as long as the liabilities are denominated in US dollars and not index-linked. The BoE, whose balance sheet is shown in Table 4, also has negligible foreign exchange reserves of its own. The bulk of the UK’s foreign exchange reserves are owned directly by the Treasury. The shareholders’ equity in the BoE is puny, just under £2 billion. The size of its balance sheet grew a lot between early 2007 and March 2008, reflecting the loans made to Northern Rock as part of the government’s rescue programme for that bank. The size of the balance sheet is around £100 billion, about 20 percent smaller than Northern Rock at its acme. Leverage is just under 50. The size of the equity and the size of the balance sheet appear small in comparison to the possible exposure of the BoE to credit risk through its LLR and MMLR operations. Its total exposure to Northern Rock was, at its peak, around £25 billion. This exposure was, of course, secured against Northern Rock’s prime mortgage assets. More important for the solvency of the BoE than this credit risk mitigation through collateral is the fact that the central bank’s monopoly of the issuance of irredeemable, non-interest-bearing legal tender means that leverage is not a constraint on solvency as long as most of the rest of the liabilities on its balance sheet are denominated in sterling and consist of nominal, that is, non-index-linked, securities, as is indeed the case for the BoE. The balance sheet of the ECB for end-year 2006 and 2007 is given in Table 5, that for the consolidated Eurosystem (the ECB and the 15 national central banks [NCBs] of the Eurosystem) as of 29 February 2008 in Table 6. The consolidated balance sheet of the Eurosystem is about 10 times the size of the balance sheet of the ECB, but the equity of the Eurosystem is about 17 times that of the ECB. Gearing

08 Book.indb 543

2/13/09 3:59:09 PM

544

Willem H. Buiter

of the Eurosystem is therefore quite low by central bank standards, with total assets just over 19 times capital. Between the end of 2006 and end-February 2008, the Eurosystem expanded its balance sheet by €237 billion. On the asset side, most of this increase was accounted for by a €67 billion increase in claims on the euro area banking sector and a €150 billion increase in other assets. Both items no doubt reflect the actions taken by the Eurosystem to relieve financial stress in the interbank markets and elsewhere in the euro area banking sector. II.6b How will the central banks finance future LLR- and MMLR- related expansions of their portfolios? Both the Fed and the BoE have tiny balance sheets and minuscule equity or capital relative to the size of the likely financial calls that may be made on these institutions. For instance, the exposure of the Fed to the Delaware SPV used to house $30 billion (face value) worth of Bear Stearns’ most toxic assets is $29 billion. The Fed’s total equity is around $40 billion. Despite my earlier contention that there is nothing to prevent a central bank from living happily ever after with negative equity, I doubt whether the Fed would want to operate with its financial liabilities larger than its financial assets. It just doesn’t look right. It is clear that the exercise of the LLR and MMLR functions may require a further rapid and large increase in L, central bank holdings of private sector securities. The central bank can always finance this increase in its exposure to the private financial sector by increasing the stock of base money, M (presumably through an increase in bank reserves with the central bank). If the economy is in a liquidity crunch, there is likely to be a large increase in liquidity preference which will cause this increase in reserves with the central bank to be hoarded rather than loaned out and spent. This increase in liquidity will therefore not be inflationary, as long as it is reversed promptly when the liquidity squeeze comes to an end. Alternatively, the central bank could finance an expansion in its holdings of private securities by reducing its holdings of government securities. Once these get down to zero, the only option left is for the

08 Book.indb 544

2/13/09 3:59:09 PM

Central Banks and Financial Crises

545

central bank to increase its non-monetary, interest-bearing liabilities, that is, an issuance of Fed Bills, BoE Bills or ECB Bills (or even Fed Bonds, BoE Bonds or ECB Bonds). As long as the central bank’s claims on the private sector earn the central bank an appropriate riskadjusted rate of return, issuing central bank bills or bonds to finance the acquisition of private securities will not weaken the solvency of the central bank ex ante. But if a significant amount of its exposure to the private sector were to default, the central bank would have to be recapitalised by the Treasury or have recourse to monetary financing. In the conventional balance sheet of the central bank, the result of a recapitalisation would be an increase in D, that is, it would look like a Treasury Bill or Treasury Bond “drop” on the central bank. It may well come to that in the US and the UK. III.

How did the three central banks perform since August 2007?

III.1 Macroeconomic stability At the time the financial crisis erupted, in August 2007, all three central banks faced rising inflationary pressures and at least the prospect of weakening domestic activity. The evidence for weakening activity was clearest in the US. In the UK, real GDP growth in the third quarter of 2007 was still robust, although some of the survey data had begun to indicate future weakness. In the euro area also, GDP growth was healthy. As late as August, the ECB was verbally signalling an increase in the policy rate for September or soon after. Since then, inflationary pressures have risen in all three currency areas, and so have inflationary expectations. There has been a marked slowdown in GDP growth, first in the US, then in UK and most recently in the euro area. While it is not clear yet whether any of the three economies are in technical recession (using the arbitrary definition of two consecutive quarters of negative real GDP growth), there can be little doubt that all three are growing below capacity, with unemployment rising in the US and in the UK and, one expects, soon also in the euro area.

08 Book.indb 545

2/13/09 3:59:09 PM

546

Willem H. Buiter

Table 7 Monetary policy actions since August 2007 by the Fed, ECB and BoE •

•

•

•

Official policy rate – Fed: -325 bps (current level: 2.00%) – ECB: +25 bps (current level: 4.25%) – BoE: -75 bps (current level: 5.00%) Unscheduled meetings, out-of-hours announcements – Fed: one for OPR (21/22 Jan.) – ECB: none – BoE: none Discount rate penalty – Fed: -75 bps (current level: 25 bps) – ECB: ±0 bps (current level: 100 bps) – BoE: ±0 bps (current level: 100 bps) Open mouth operations – ECB: repeated hints at/threats of OPR increases that did not materialise until July 2008 (“talk loudly & carry a little stick”)

The monetary response to rather similar circumstances has, however, been very different in the three economies, as is clear from the summary in Table 7. The Fed cut its official policy rate aggressively—by 325 basis points cumulatively so far. On September 18, 2007, the Fed cut the federal funds target by 50 basis points to 4.75 percent, with a further reduction of 25 basis points following on October 31. On December 11 there was a further 25 basis points cut, on January 21, 2008 a 75 basis points cut, on January 30 a 50 basis points cut, on March 18 a 75 basis points cut and on April 30 another 25 basis points cut. This brought the federal funds target to 2.00 percent, where it remains at the time of writing (August 10, 2008). The Fed also reduced the “discount window penalty,” that is, the excess of the rate charged on overnight borrowing at the primary discount window over the federal funds target rate, from 100 bps to 50 bps on August 17, 2007 and to 25 bps on March 18, 2008. This cut in the discount rate penalty can be viewed as a liquidity management measure as well as a (second-order) macroeconomic policy measure. Finally, one of the Fed’s rate cuts (the 75 basis points reduction on January 21, 2008), was at an “unscheduled” meeting and was announced out of normal

08 Book.indb 546

2/13/09 3:59:09 PM

Central Banks and Financial Crises

547

working hours, thus signalling a sense of urgency in one interpretation, a sense of panic in another. The BoE kept its official policy rate at 5.75 percent until December 6, 2007 when it made a 25 basis points cut. Further 25 bps cuts followed on February 7, 2008 and April 10, 2008, so Bank Rate now stands at 5.00 percent. The discount rate (standing lending facility) penalty over Bank Rate remained constant at 100 bps. There were no meetings or policy announcements on unscheduled dates or at unusual times. The ECB kept its official policy rate unchanged at 4.00 percent until July 3, 2008 when it was raised to 4.25 percent, where it still stands. There has also been no change in the discount rate penalty: The marginal lending facility continues to stand at 100 basis points above the official policy rate. There were no meetings on unscheduled dates or announcements at non-standard hours. Unlike the other two central banks, the ECB repeatedly, between June 2007 and July 2008, talked tough about inflation and hinted at possible rate increases. This talk was matched by official policy rate action only on July 3, 2008. The markedly different monetary policy actions of the Fed compared to the other two central banks can, in my view, not be explained satisfactorily with differences in objective functions (the Fed’s dual mandate versus the ECB’s and the BoE’s lexicographic price stability mandate) or in economic circumstances. The slowdown in the US did come earlier than in the UK and in the euro area, but the inflationary pressures in the US were, if anything, stronger than in the UK and the euro area. I conclude that the Fed overreacted to the slowdown in economic activity. It cut the official policy rate too fast and too far and risked its reputation for being serious about inflation. I believe that part of the reason for these policy errors is a remarkable collection of analytical flaws that have become embedded in the Fed’s view of the transmission mechanism. These errors are shared by many FOMC members and by senior staff. They are worth outlining here, because they serve as a warning as to what can happen when the research and

08 Book.indb 547

2/13/09 3:59:09 PM

548

Willem H. Buiter

economic analysis underlying monetary policy making become too insular and inward-looking, and is motivated more by the excessively self-referential internal dynamics of academic research programmes than by the problems and challenges likely to face the policy-making institution in the real world. III.1a The macroeconomic foibles of the Fed There are some key flaws in the model of the transmission mechanism of monetary policy that shapes the thinking of a number of influential members of the FOMC. These relate to the application of the Precautionary Principle to monetary policy making, the wealth effect of a change in the price of housing, the role of core inflation as a guide to future underlying inflation, the possibility of achieving a sustainable external balance for the US economy without going through a deep and/or prolonged recession, the effect of financial sector deleveraging on aggregate demand, and the usefulness of the monetary aggregates as a source of information about macroeconomic and financial stability. III.1a(i) Risk management and the “Precautionary Principle” Consider the following example of optimal decision making under uncertainty. I stand before an 11-foot-wide ravine that is 2,000 feet deep. I have to jump across. A safe jump is one foot longer than the width of the ravine. I can jump any distance, but a longer jump requires more effort, something I dislike moderately. I also am strongly averse to falling to my death. Without uncertainty, I make the shortest leap that will get me safely over the precipice—12 feet. Now assume that I cannot see how wide the ravine is. All I know is that its width is equally likely to be anywhere on the interval 1 foot to 21 feet. So the expected width of the ravine is 11 feet. There continues to be certainty about the depth of the ravine—2,000 feet, that is, certain death if I were to fall in. It clearly would not be rational for me to adopt the certainty-equivalent strategy and make a 12-foot jump. I would be cautious and make a much larger jump, of 23 feet. Caution and prudence here dictate more radical action—a longer jump. A dramatic departure from symmetry in the payoff function

08 Book.indb 548

2/13/09 3:59:09 PM

Central Banks and Financial Crises

549

accounts for the difference between rational behaviour and certainty-equivalent behaviour. The Fed justifies its radical interest cuts in part by asserting that these large cuts minimize the risk of a truly catastrophic outcome. I want to question whether the Fed’s official policy rate actions can indeed be justified on the grounds that the US economy was tottering at the edge of a precipice, and that aggressive rate cuts were necessary to stop it from tumbling in. Under Chairman Greenspan, so-called risk-based “decision theory” approaches became part of the common mindset of the FOMC (see Greenspan 2005). They continue to be influential in the Bernanke Fed. A clear articulation can be found in Mishkin (2008b). At last year’s Jackson Hole Symposium, Martin Feldstein (2008) also made an appeal to a risk-based decision theory approach to justify looking after the real economy first, through aggressive interest rate cuts, despite the obvious risk this posed to inflation and moral hazard. Mishkin (2008b) argued that the combination of non-linearities in the economy with both a higher degree of uncertainty and a high probability of extreme (including extremely bad) outcomes (so-called “fat tails”) justified the Fed’s focus on extreme risks. In addition, he asserts that the extreme risk faced by the US economy is a financial instability/collapse-led sharp contraction in economic activity. This is the “Precautionary Principle” (PP) applied to monetary policy. At times of high uncertainty, policy should be timely, decisive, and flexible and focused on the main risk. Even where it is applied correctly, I don’t think much of the PP. Except under very restrictive conditions, unlikely to be satisfied ever in the realm of economic policy making, I consider the behaviour it prescribes to be pathologically risk-averse. In its purest incarnation —under complete Knightian uncertainty—it amounts to a minimax strategy: You focus all your policy instruments on doing as well as you can in the worst possible outcome. Despite its axiomatic foundations, the minimax principle has never appealed to me either as a normative or a positive theory of decision under uncertainty. But I don’t have to fight the PP, or minimax, here. The application of the PP to the monetary policy choices made by the Fed in 2007

08 Book.indb 549

2/13/09 3:59:09 PM

550

Willem H. Buiter

and 2008 is bogus. The PP came to the social sciences from the application of decision theory to regulatory decisions involving environmental risk (global warming, species extinction) or technological risk (genetically modified crops, nanotechnology). Its basic premise in these areas is “... that one should not wait for conclusive evidence of a risk before putting control measures in place designed to protect the environment or consumers.”(Gollier and Treich, 2003). For instance, Principle 15 of the 1992 Rio Declaration states, “Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.” Attempts to make sense of the PP in a setting of sequential decision making under uncertainty lead to the conclusion that, for something like the Rio Declaration version of the PP to emerge as a normative guide to behaviour, all of the following must be present (see Collier and Treich, 2003, from which the following sentence is paraphrased): a long time horizon, stock externalities, irreversibilities (physical and socio-economic), large uncertainties and the possibility of future scientific progress (learning). Short-term policy should keep the option value of future learning alive. When the long-term effects of certain contingencies are unknown (but may be uncovered later on), it may be optimal to be more cautious in the early stages of the sequential management of risk. I believe the analysis of Collier and Treich to be essentially correct. The question then becomes: What does this imply for whether the Fed, in the circumstances of the second half of 2007 and the first half of 2008, did the right thing when it cut the official policy rate from 5.25 percent to 2.00 percent rather than cutting it by less, keeping it constant or raising it? The Fed decided to give priority to minimising the risk of a sharp contraction in real economic activity. It accepted the risk of higher inflation. How does this square with the PP? The answer is: Not very well at all. The extreme risk faced by the US economy during the past year has not been a sharp contraction in real economic activity caused by a financial collapse. There is no irreversibility involved in a sharp contraction in economic activity. Mishkin’s rather vague “non-linearities” are no substitute for the irreversibility

08 Book.indb 550

2/13/09 3:59:09 PM

Central Banks and Financial Crises

551

required for the PP to apply. This is not like a catastrophic species extinction or a sudden melting of the polar ice caps. The crash of 1929 became the Great Depression of the 1930s because the authorities permitted the banking system to collapse and did not engage in sustained aggressive expansionary fiscal and monetary policy even when the unemployment rate reached almost 25 percent in 1933. In addition, the international trading system collapsed. The Fed as LLR and MMLR has effectively underwritten the balance sheet of all systemically important US banks (investment banks as well as commercial banks) with the rescue of Bear Stearns in March 2008. Current worries about the international trading system concern the absence of progress rather than the risk of a major outbreak of protectionism. Most of all, should economic activity fall sharply and remain depressed for longer than is necessary to correct the fundamental imbalances in the US economy (the external trade deficit, excessive household indebtedness and the low national saving rate), monetary and fiscal policy can be used aggressively at that point in time to remedy the problem. There is no need to act now to prevent some irreversible or even just costly-to-reverse catastrophy from occurring. Boosting demand through expansionary monetary and fiscal policy is not hard. It is indeed far too easy. We are also not buying time to uncover some new scientific fact that will allow us to improve the short-run inflationunemployment trade–off or to boost the resilience of the economy to future disinflationary policies. Cutting rates to support demand does not create or preserve option value. Even when there is a zero lower bound constraint on the short nominal interest rate and even if there is a non-negligible probability that this constraint will become binding, aggregate demand management continues to be effective. Indeed, it is precisely when the zero lower bound constraint on the nominal interest is binding that fiscal policy is at its most effective. If anything, the (weak) logic of the PP points to giving priority to fighting inflation rather than to preventing a sharp contraction of demand and output. Output contractions can be reversed easily through expansionary monetary and fiscal policies. High inflation, once it becomes embedded in inflationary expectations, may take

08 Book.indb 551

2/13/09 3:59:09 PM

552

Willem H. Buiter

a long time to squeeze out of the system again. If the sacrifice ratio is at all unfriendly, the cumulative unemployment or output cost of achieving a sustained reduction in inflation could be large. The irreversibility argument (strictly, the costly reversal argument) supports erring on the side of caution by not letting inflation and inflationary expectations rise.20 “Fat tails”, the Precautionary Principle and other decision theory jargon should only be arbitraged into the area of monetary policy if the substantive conditions are satisfied. Today, in the US, they are not.21 With existing policy tools, we can address a disastrous collapse in activity if, as and when it occurs. There is no need for preventive or precautionary drastic action. I agree that dynamic stochastic optimisation based on the LQG (linear-quadratic-Gaussian) assumptions, and the certainty-equivalent decision rules they imply are inappropriate for monetary policy design. This is because (1) the objectives of most central banks cannot be approximated well with a quadratic functional form (especially in the case of the BoE and the ECB with their lexicographic preferences), (2) no relevant economic model is linear and (3) the random shocks perturbing the economic system are not Gaussian.22 I was fortunate in having Gregory Chow as a colleague during my first academic job (at Princeton University). The periodic rediscoveries, in the discussion of macroeconomic policy design, of aspects of his work (Chow, 1975, 1981, 1997) are encouraging, but they also demonstrate that progress in economic science is not monotonic. Mishkin (2008b) admits that “Formal models of how monetary policy should respond to financial disruptions are unfortunately not yet available....” This, however, does not stop him from giving, in that same speech, confident and quite detailed prescriptions for the response of monetary policy to financial disruptions. “Monetary policy cannot—and should not—aim at minimizing valuation risk, but policy should aim at reducing macroeconomic risk…. Monetary policy needs to be timely, decisive, and flexible.... Monetary policy must be at least as preemptive in responding to financial shocks as in responding to other types of disturbances to the economy.” Possibly, but not based on any rigorous analysis of a coherent, quantitative model of the US economy or any

08 Book.indb 552

2/13/09 3:59:09 PM

Central Banks and Financial Crises

553

other economy. Emphatic statements do not amount to a new science of monetary policy. Repeated assertion is not a third mode of scientific reasoning, on a par with induction and deduction. III.1a(ii)

Housing wealth isn’t wealth

This bold statement was put to me about ten years ago by Mervyn King, now Governor of the BoE, then Chief Economist of the BoE, shortly after I joined the Monetary Policy Committee of the BoE as an External Member in June 1997. Like most bold statements, the assertion is not quite correct; the correct statement is that a decline in house prices does not make you worse off, that is, it does not create a pure wealth effect on consumer demand. The argument is elementary and applies to coconuts as well as to houses. When does a fall in the price of coconuts make you worse off? Answer: when you are a net exporter of coconuts, that is, when your endowment of coconuts exceeds your consumption of coconuts. A net importer of coconuts is better off when the price of coconuts falls. Someone who is just self-sufficient in coconuts is neither worse off nor better off. Houses are no different from durable coconuts in this regard. The fundamental value of a house is the present discounted value of its current and future rentals, actual or imputed. Anyone who is “long” housing, that is, anyone for whom the value of his home exceeds the present discounted value of the housing services he plans to consume over his remaining lifetime will be made worse off by a decline in house prices. Anyone “short” housing will be better off. So the young and all those planning to trade up in the housing market are made better off by a decline in house prices. The old and all those planning to trade down in the housing market will be worse off. Another way to put this is that landlords are worse off as a result of a decline in house prices, while current and future tenants are better off. On average, the inhabitants of a country own the houses they live in; on average, every tenant is his own landlord and vice versa. So there is no net housing wealth effect. You have to make a distributional argument to get an aggregate pure net wealth effect from a change

08 Book.indb 553

2/13/09 3:59:09 PM

554

Willem H. Buiter

in house prices. A formal statement of the proposition that a change in house prices has no wealth effect on private consumption demand can be found in Buiter (2008b,c). Informal statements abound (see e.g. Buchanan and Fiotakis, 2004, or Muellbauer, 2008). Most econometric or calibrated numerical models I am familiar with treat housing wealth like the value of stocks and shares as a determinant of household consumption. They forget that households consume housing services (for which they pay or impute rent) but not stock services. An example is the FRB/US model. It is used frequently by participants in the debate on the implication of developments in the US housing market for US consumer demand. A recent example is Frederic S. Mishkin’s (2008a) paper “Housing and the Monetary Transmission Mechanism.” The version of the FRB/US model Mishkin uses a priori constrains the wealth effects of housing wealth and other financial wealth to be the same. The long-run marginal propensity to consume out of non-human wealth (including housing wealth) is 0.038, that is, 3.8 percent. In several simulations, Mishkin increases the value of the long-run marginal propensity to consume out of housing wealth to 0.076, that is, 7.6 percent, while keeping the long-run marginal propensity to consume out of nonhousing financial wealth at 0.038. The argument for an effect of housing wealth on consumption other than the pure wealth effect is that housing wealth is collateralisable. Households-consumers can borrow against the equity in their homes and use this to finance consumption. It is much more costly and indeed often impossible, to borrow against your expected future labour income. If households are credit-constrained, a boost to housing wealth would relax the credit constraint and temporarily boost consumption spending. The argument makes sense and is empirically supported (see e.g. Edelstein and Lum, 2004, or Muellbauer, 2008). Of course, the increased debt will have to be serviced, and eventually consumption will have to be brought down below the level it would have been at in the absence of the mortgage equity withdrawal. At market interest rates, the present value of current and future consumption will not be affected by the MEW channel.23

08 Book.indb 554

2/13/09 3:59:10 PM

Central Banks and Financial Crises

555

Ben Bernanke (2008a), Don Kohn (2006), Fredric Mishkin (2008a), Randall Kroszner (2007) and Charles Plosser (2007) all have made statements to the effect that there is a pure wealth effect through which changes in house prices affect consumer demand, separate from the credit, MEW or collateral channel.24 The total effect of a change in house prices on consumer demand adds the credit or collateral effect to the standard (pure) wealth effect. This is incorrect. The benchmark should be that the credit, MEW or collateral effect is instead of the normal (pure) wealth effect. By overestimating the contractionary effect on consumer demand of the decline in house prices, the Fed may have been induced to cut rates too fast and too far. There are channels other than private consumption through which a change in house prices affects aggregate demand. One obvious and empirically important one is household investment, including residential construction. A reduction in house prices that reflects the bursting of a bubble rather than a lower fundamental value of the property also produces a pure wealth effect (Buiter, 2008b,c). My criticism of the Fed’s overestimation of the effect of house price changes on aggregate demand relates only to the pure wealth effect on consumption demand, not to the “Tobin’s q” effect of house prices on residential construction. III.1a(iii)

The will-o’-the-wisp of “core” inflation

The only measure of core inflation I shall discuss is the one used by the Fed, that is, the inflation rate of the standard headline CPI or PCE deflator excluding food and energy prices. Other approaches to measuring core inflation, including the vast literature that attempts to extract trend inflation or some other measure of “underlying” inflation using statistical methods in the time or frequency domains, including “trimmed mean” measures and “approximate band pass filters” will not be considered (see e.g. Bryan and Cecchetti 1994; Quah and Vahey, 1995; Baxter and King 1999; Cogley 2002; Cogley and Sargent, 2001, 2005; Dolmas, 2005; Rich and Steindel, 2007; Kiley, 2008). I assume that the price stability leg of the Fed’s mandate refers to price stability, now and in the future, defined in terms of a representative

08 Book.indb 555

2/13/09 3:59:10 PM

556

Willem H. Buiter

basket of consumer goods and services that tries to approximate the cost of living of some mythical representative American. It is well-known that price stability, even in terms of an ideal cost of living index, cannot be derived as an implication of standard microeconomic efficiency arguments. The Friedman rule gives you a zero pecuniary opportunity cost of holding cash balances as (one of) the optimality criteria, that is, i=i M. When cash bears a zero rate of interest, this gives us a zero risk-free nominal interest rate as (part of) the optimal monetary rule. With a positive real interest rate, this gives us a negative optimal rate of inflation for consumer prices, something even the ECB is not contemplating. Menu costs imply the desirability of minimising price changes for those goods and services for which menu costs are highest. Presumably this would call for stabilisation of money wages, since the cost of wage negotiations is likely to exceed that of most other forms of price setting. With positive labour productivity growth, a zero money wage inflation target would give us a negative optimal rate of producer price inflation. New-Keynesian sticky-price models of the Calvo-Woodford variety yield (in their simplest form) two distinct optimal inflation criteria, one for consumer prices and one for producer prices. Neither implies that stability of the sticky-price sub-index is optimal. Equations (15) and (16) below show the log-linear approximation at the deterministic steady state of the (negative of the) social welfare function (which equals the utility function of the representative household) and of the New-Keynesian Phillips curve in the simple sticky-price Woodford-Calvo model, when the potential level of output (minus the natural rate of unemployment),ŷ, is efficient (see Calvo, 1983 Woodford, 2003; and Buiter, 2004). i

∞  1  L t = Et ∑     i =o 1 + δ

(( p

t +i

− p t +i ) + w ( yt +i − yˆt +i ) + f (it + j − itM+ j ) 2

2

)

(15)

δ > 0, w > 0, f ≥ 0

p t − p t = β Et ( p t +1 − p t +1 ) + γ ( yt − yˆt ) + j (it − itM ) 0 < β < 1; γ > 0

08 Book.indb 556

2

(16)

2/13/09 3:59:10 PM

Central Banks and Financial Crises

557

In the Calvo model of staggered overlapping price setting, in each period, a randomly selected constant fraction of the population of monopolistically competitive firms sets prices optimally. The remainder follows a simple rule of thumb or heuristic for its price. The inflation rate chosen by the constrained price setters in period t isp t . Optimality in this model requires it = itM



(17)

pt=p t

(18)

Equations (17) and (18) then imply that yt=ŷt. The requirement in (17) that the pecuniary opportunity cost of holding cash be zero is Friedman’s misnamed Optimal Quantity of Money rule. The second optimality condition, given in (18), requires that the headline producer price inflation rate, p, be the same as the inflation rate of the constrained price setters,p . If in any given period the inflation rate of the constrained price setters is predetermined, then the second optimality requirement becomes the requirement that overall producer price inflation accommodates the inflation rate set by the constrained price setters, whatever this happens to be. Even if one identifies the inflation rate set by the constrained price setters with “core” inflation (which would be a stretch), this New-Keynesian framework does not generate an optimal rate of inflation either for core inflation or for headline inflation. All it prescribes is a constant relative price of core to non-core goods and services. Without luck or additional instruments (such as indirect taxes and subsidies driving a wedge between consumer and producer prices) it will not in general be possible to satisfy both the Friedman rule and the constant relative price rule (of free and constrained price setters). How then can this framework be used to rationalise (a) targeting Woodford-Calvo “core” inflation and (b) aiming for stability of the Woodford-Calvo “core” producer price level? Two steps are required. First, the Friedman rule is finessed or ignored. This requires either the counterfactual assumption that the interest rate on cash is not constrained to equal zero but can instead be set equal at all times to the interest rate on non-cash financial instruments (that is, equation 17

08 Book.indb 557

2/13/09 3:59:11 PM

558

Willem H. Buiter

always holds, but i remains free), or the assumption that the technology and preferences in this economy take the rarefied form required to make the demand for cash independent of its opportunity cost, in which case f = j = 0. Second, the Woodford-Calvo “core” inflation rate, which plays the role of the target inflation rate in the social welfare function (15) is zero:p =0. This is the assumption Calvo made in his original paper (Calvo, 1983). Clearly, the assumption that the constrained price setters will always keep their prices constant, regardless of the behaviour of prices and inflation in the rest of the economy is unreasonable. It assumes the absence of any kind of learning, no matter how partial and unsophisticated. It has strange implications, including the existence of a stable, exploitable inflation-unemployment trade-off or inflationoutput gap trade-off across deterministic steady states. Calvo recognised the unpalatable properties of his unreasonable original price setting function in Calvo, Celasun and Kumhof (2007). An attractive alternative, in the spirit of John Flemming’s (1976) theory of the “gearing” of inflation expectations, would be to impose as a minimal rationality requirement the assumption that the inflation rate set by the constrained price setters is cointegrated with that of the unconstrained price setters or the headline inflation rate. Because price stability cannot be rationalised as an objective of monetary policy using standard microeconomic efficiency arguments, I fall back on legal mandate/popular consensus justifications for price stability as an objective of monetary policy. In the US, the euro area and UK, stable prices or price stability is a legally mandated objective of monetary policy. In the UK, the Chancellor defines the price index. It is the CPI (the harmonised version). In the euro area the ECB’s Governing Council itself chooses the index used to measure price stability. Again, it is the CPI. In the US there is no such verifiable source of legitimacy for a particular index. I therefore appeal to what I believe the public at large understands by price stability, which is a constant cost of living. I take it as given that the Fed’s definition of price stability is to be operationalized through a representative cost of living index. This means that the Fed does not care intrinsically about core inflation (in the

08 Book.indb 558

2/13/09 3:59:11 PM

Central Banks and Financial Crises

559

sense of the rate of inflation of a price index that excludes food and energy). Americans do eat, drink, drive cars, heat their homes and use air conditioning. The proper operational target implied by the price stability leg of the Fed’s dual mandate is therefore headline inflation. Core inflation is relevant to the price stability leg of the Fed’s mandate to the extent that it is a superior predictor of future headline inflation, over the horizon that the Fed can influence headline inflation—a better predictor not only than headline inflation itself, but than any readily available set of predictors. After all, the monetary authority should not restrict itself to univariate or bivariate predictor sets, let alone univariate or bivariate predictor sets consisting of the price series itself and its components.25 Non-core prices tend to be set in auction-type markets for commodities. They are flexible. Core goods and services tend to have prices that are subject to short-run Keynesian nominal rigidities. They are sticky. The core price index and its rate of inflation tend to be both less volatile and more persistent than the index of non-core prices and its rate of inflation, and also than the headline price index and its rate of inflation. However, the ratio of core to non-core prices and of the core price index to the headline price index is predictable, and so are the relative rates of inflation of the core and headline inflation indices. This is clear from Charts 6a and 6b. The phenomenon driving the increase in the ratio of headline to core prices in recent years is well-understood. Newly emerging market economies like China, India and Vietnam have entered the global economy as demanders of non-core commodities and as suppliers of core goods and services. This phenomenon is systematic, persistent and ongoing. When core goods and services are subject to nominal price rigidities but non-core goods prices are flexible, a relative demand or supply shock that causes a permanent increase (decrease) in the relative price of non-core to core goods will, for a given path of nominal official policy rates, cause a temporary increase in the rate of headline inflation, and possibly a temporary reduction in the rate of core inflation as well.

08 Book.indb 559

2/13/09 3:59:11 PM

560

Willem H. Buiter

Chart 6a US CPI headline-to-core ratio 01/1957 - 04/2008; SA, 1982-84=100 104

104 CPI Headline-to-Core Price Ratio

200607

200310

200101

199804

199507

199210

199001

198704

92

198407

92

198110

94

197901

94

197604

96

197307

96

197010

98

196801

98

196504

100

196207

100

195910

102

195701

102

Source: Bureau of Labor Statistics

Chart 6b US PCE deflator headline-to-core ratio 01/1959 - 03/2008; SA, 2000=100 106

106 PCE deflator headline-to-core ratio

200604

200401

200110

199907

199704

199501

199210

199007

198804

198601

198310

94

198107

94

197904

96

197701

96

197410

98

197207

98

197004

100

196801

100

196510

102

196307

102

196104

104

195901

104

Source: Bureau of Economic Analysis

08 Book.indb 560

2/13/09 3:59:11 PM

Central Banks and Financial Crises

561

This pattern is clear from Charts 7a, b, c and d, which plot the difference between the headline inflation rate and the core inflation rate on the horizontal axis against the rate of headline inflation on the vertical axis. This is done, in Charts 7a and 7b, for the CPI over, respectively, the 1958-2008 period and the 1987-2008 period. It is repeated in Charts 7c and 7d for the PCE deflator over, respectively, the 1960-2008 and the 1987-2008 periods. Therefore, when there is a continuing upward movement in the relative price of non-core goods to core goods, core inflation will be a poor predictor of future headline inflation for two reasons. First, even if headline inflation were unchanged and independent of the factors that drive the change in relative prices, core inflation would, for as long as the upward movement in the relative price of non-core goods continued, be systematically below both non-core inflation and headline inflation. Second, for a given path of nominal interest rates, the increase in the relative price of non-core goods will temporarily raise headline inflation above the level it would have been if there had been no increase in the relative price of non-core goods to core goods and services. The implication is that for many years now (starting around the turn of the century), the Fed has missed the boat on the implications of the global increase in the relative price of non-core goods for the usefulness of core inflation as a predictor of future headline inflation. Medium-term inflationary pressures have been systematically higher than the Fed thought they were. I am not arguing that the Fed has focused on core rather than on headline inflation because this permits it to take a more relaxed view of inflationary pressures. My argument is that because the Fed, for whatever reason(s), decided to focus on core rather than on headline inflation, and because for most of this decade there has been a persistent increase in the relative price of non-core goods to core goods and services, the Fed has, for most of this decade, underestimated the underlying inflationary pressures in the US. Should the recent upward trend in non-core to core prices go into reverse, the opposite bias would result. With a global economic slowdown in the works, a cyclical decline in real commodity prices is quite likely for the next couple of years or so. Following the end

08 Book.indb 561

2/13/09 3:59:11 PM

562

Willem H. Buiter

Chart 7a US CPI headline inflation vs. headline minus core inflation 1958/01 - 2008/04 16

Headline inflation (percent)

14

12

10

8

6

4

2

-3

-2

-1

0

1

2

3

4

5

6

Headline minus core inflation (percent)

Source: Bureau of Labor Statistics

Chart 7b US CPI headline inflation vs. headline minus core inflation 1987/01-2008/04 7

Headline inflation (percent)

6

5

4

3

2

1

0 -3

-2

-1

0

1

2

3

4

Headline minus core inflation (percent)

Source: Bureau of Labor Statistics

08 Book.indb 562

2/13/09 3:59:12 PM

Central Banks and Financial Crises

563

Chart 7c US PCE headline inflation vs. headline minus core inflation 1960/01-2008/03 PCE headline inflation (percent)

14

12

10

8

6

4

2

0 -2

-1

0

1

2

3

4

5

PCE headline minus core inlflation (percent)

Source: Bureau of Economic Analysis

Chart 7d US PCE headline inflation vs. headline minus core inflation 1987/01-2008/03 6

PCE headline inflation (percent)

5

4

3

2

1

0 -2

-1.5

-1

Source: Bureau of Economic Analysis

08 Book.indb 563

-0.5

0

0.5

1

1.5

2

PCE headline minus core inflation (percent)

2/13/09 3:59:12 PM

564

Willem H. Buiter

of this global cyclical correction, however, I expect that a full-speed resumption of commodity-biased demand growth and of core goods and services-biased supply growth in key emerging markets will in all likelihood lead to a further trend increase in the relative price of non-core goods to core goods and services. The other main lesson from the core inflation debacle is that those engaged in applied statistics should not leave their ears and eyes at home. Specifically, it pays to get up from the keyboard and monitor occasionally to open the window and look out to see whether a structural break might be in the works that is not foreshadowed in any of the sample data at the statistician’s disposal. Two-and-a-half billion Chinese and Indian consumers and producers entering the global economy might qualify as an epochal event capable of upsetting established historical statistical regularities. Finally, a brief remark on the Fed’s fondness for the PCE deflator. Communication with the wider public (all those not studying index numbers for a living) is made more complicated when the index in terms of which inflation and price stability are measured bears no obvious relationship to a reasonably intuitive concept like the cost of living. I believe the PCE deflator falls into this obscure category. Furthermore, being a price deflator (current-weighted), the PCE deflator (headline or core) will tend to produce inflation rates lower than the corresponding CPI index (which is base-weighted). Since 01/1987, the difference between the headline CPI and PCE deflator inflation rates has been 0.44 percent at an annual rate. The difference between the core CPI and PCE deflator inflation rates has been 0.45 percent. Over the longer period 01/1960-03/2008 the difference between the headline CPI and PCE inflation rates has been 0.47 percent, that between core CPI and PCE inflation rates 0.55 percent. This further reinforces the inflationary bias of the Fed’s procedures. III.1a(iv) Is the external position of the US sustainable? If not, can it be corrected without a recession? The argument of this subsection is in two parts. First, the external positions of the US and the UK are unsustainable. Second, it is all but unavoidable that the US and the UK will have to go through prolonged

08 Book.indb 564

2/13/09 3:59:12 PM

Central Banks and Financial Crises

565

and/or deep slowdowns in economic activity to achieve sustainable external balances and desirable national saving rates. Attempts to stimulate demand, whether through interest rate cuts or through tax stimuli like the £100 billion fiscal package implemented in the US during the second quarter of 2008, are therefore counterproductive, as they delay a necessary adjustment. The additional employment and growth achieved through such monetary and fiscal stimuli are unsustainable because they make an already unsustainable imbalance worse. If the Fed’s real economic activity leg of its dual mandate refers to sustainable growth and sustainable employment, the interest rate cut stimuli provided since August 2007 are therefore in conflict with that mandate. Almost the same conclusion is reached even if one is either not convinced or not bothered by the argument that the external position of the US economy is unsustainable. It is possible to reach pretty much the same conclusion as long as one subscribes to the argument that the US national saving rate is dangerously low for purely domestic reasons (providing for the comfortable retirement of an ageing population), and needs to be raised materially. Policies or shocks that raise the US national saving rate are highly unlikely to produce a matching increase in the US domestic investment rate, given the growing array of more profitable investment opportunities abroad, especially in emerging markets. The unsustainability of the US and UK external balances Around the middle of 2007, when the financial crisis started, the US had an external primary deficit of about six percent of GDP (see Chart 8b).26 The US is also a net external debtor (see Chart 8a). Its net international investment position is not easily or accurately marked to market, but something close to a negative 20 percent of GDP is probably a reasonable estimate. Let ft be the ratio of end-of-period t net external liabilities as a share of period t GDP, rt the real rate of return paid during period t on the beginning-of-period net foreign investment position, gt the growth rate of real GDP between periods t-1 and t and xt the external primary balance as a share of GDP. It follows that:

08 Book.indb 565

2/13/09 3:59:12 PM

566

Willem H. Buiter

Chart 8a US external assets and liabilities, 1980-2007 (percent of GDP) 160.0

Percent

160.0 US Net International Investment Position US External Assets US External Liabilities

140.0 120.0

140.0 120.0

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

-20.0

1991

0.0 1990

0.0 1989

20.0

1988

20.0

1987

40.0

1986

40.0

1985

60.0

1984

60.0

1983

80.0

1982

80.0

1981

100.0

1980

100.0

-20.0 -40.0

-40.0

Source: Bureau of Economic Analysis

Chart 8b US investment income and primary surplus 1980QI–2007QI (percent of GDP) 8

6

Percent

Percent US Foreign Income Credits US Foreign Income Debits

US Net Foreign Income US Primary Surplus

8

6

4

2

2

0

0

-2

1980-I 1980-IV 1981-III 1982-II 1983-I 1983-IV 1984-III 1985-II 1986-I 1986-IV 1987-III 1988-II 1989-I 1989-IV 1990-III 1991-II 1992-I 1992-IV 1993-III 1994-II 1995-I 1995-IV 1996-III 1997-II 1998-I 1998-IV 1999-III 2000-II 2001-I 2001-IV 2002-III 2003-II 2004-I 2004-IV 2005-III 2006-II 2007-IV 2007-I

4

-2

-4

-4

-6

-6

-8

-8

Source: Bureau of Economic Analysis

08 Book.indb 566

2/13/09 3:59:13 PM

Central Banks and Financial Crises

 1 + rt  ft ≡   ft −1 − xt  1 + gt 

567

(19)

The primary surplus that keeps constant net foreign liabilities as a share of GDP, xt , is given by:  r − gt  xt =  t  ft −1.  1 + gt 

I assume that the long-run growth rate of the net external liabilities is less than the long-run rate of return on the net external liabilities or, equivalently, that the present discounted value of the net external liabilities is non-positive in the long run (the usual national solvency constraint). The nation’s intertemporal budget constraint then becomes the requirement that the existing net external liabilities should not exceed the present discounted value of current and future primary external surpluses. This can be written more compactly as follows:  r p − gtp  xt p ≥  t f p  t −1  1 + gt 

(20)

Here xtp is the permanent primary surplus as a share of GDP and

rt p and gtp are

the permanent real rate of return paid on the net external liabilities and the permanent growth rate of real GDP respectively. “Permanent” here is used in the sense of permanent income. Its approximate meaning is “expected long-run average” (see Buiter and Grafe, 2004). All I need to make my point is that the US is a net external debtor and that the permanent real rate of return paid on US net external liabilities in the future will indeed in the future exceed the permanent growth rate of US real GDP. If this second assumption is not satisfied, the US can engage in external Ponzi finance forever. Possible, but not likely, especially following the ongoing crisis. Given rt p > gtp and ft−1 > 0, it follows that the US will have to generate, henceforth, a permanent external primary surplus: xtp > 0. Unless the US expects to be a permanent net recipient of foreign aid, this means that the US has to run a permanent trade surplus. From the position the US was in immediately prior to the crisis, this means

08 Book.indb 567

2/13/09 3:59:15 PM

568

Willem H. Buiter

that a permanent increase in the trade balance surplus as a share of GDP of at least six percentage points is required. The UK is in a similar position, with a Net International Investment Position of around minus 27 percent of GDP in 2007 and a primary deficit of almost 5 percent of GDP. This can be seen in Charts 9a and 9b. Note that, unlike the US and the euro area, where gross external assets and liabilities are just over 100 percent of annual GDP, in the UK both external assets and external liabilities are close to 500 percent of annual GDP. The characterisation of the UK as a hedge fund is only a mild exaggeration. The euro area, like the US and the UK, has a small negative Net International Investment Position. Unlike the US and the UK, its primary balance has averaged close to zero since the creation of the euro. Charts 10a and 10b show the behaviour of the external assets, liabilities and investment income for the euro area. The mid-2007 6 percent of GDP US primary deficit was probably an overstatement of the structural trade deficit, because the US economy was operating above capacity. Since the middle of 2007, the US primary deficit has shrunk to about 5 percent of GDP. With the economy now operating with some excess capacity, this probably understates the structural external deficit. I will assume that the US economy has to achieve at least a five percent of GDP permanent increase in the primary balance to achieve external solvency. The corresponding figure for the UK is probably about at least four percent of GDP. The euro area has been in rough structural balance for a number of years. To say that the US needs a permanent 5 percent of GDP reduction in the external primary deficit is to say that the US needs a 5 percent fall in domestic absorption (the sum of private consumption, private investment and government spending on goods and services, or “exhaustive” public spending) relative to GDP. This reduction in domestic absorption is also necessary to support a lasting depreciation of the US real exchange rate (an increase in the relative price of

08 Book.indb 568

2/13/09 3:59:16 PM

Central Banks and Financial Crises

569

Chart 9a UK external assets and liabilities, 1980-2007 (percent of GDP) 600

Percent

30 UK External Assets (LH axis) UK External Liabilities (LH axis) UK Net International Investment Position (RH axis)

500

20

2007

2006

2004

2005

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

-30 1988

0

1987

-20

1986

100

1985

-10

1984

200

1983

0

1982

300

1981

10

1980

400

Source: Office of National Statistics

Chart 9b UK investment income and primary surplus, 1980-2007 (percent of GDP) 25

Percent

25 UK Investment Income Credits UK Investment Income Debits

20

UK Investment Income Balance UK Primary Surplus

20

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

0 1987

0 1986

5

1985

5

1984

10

1983

10

1982

15

1980 1981

15

-5

-5

-10

-10

Source: Office of National Statistics

08 Book.indb 569

2/13/09 3:59:16 PM

570

Willem H. Buiter

Chart 10a Euro area external assets and liabilities, 1999Q1–2008Q1 (percent of GDP) 180.00

Percent

Percent

0.00

160.00

-2.00

140.00

-4.00

120.00 -6.00 100.00 -8.00 80.00 -10.00 60.00 -12.00

40.00

200712

200707

200702

200609

200604

200511

200506

200501

200408

-14.00

200403

200310

200305

200212

200207

200202

200109

200104

200011

200006

199908

0.00

199903

20.00

200001

euro area Foreign Assets (LH axis) euro area Foreign Liabilities (LH axis) euro area Net International Investment Position (RH axis)

-16.00

Source: Eurostat and ECB

Chart 10b Euro area investment income and primary surplus, 1999Q1-2008Q1 (percent of GDP) 8

Percent

Percent 8 euro area Investment Income Credits (% of GDP) euro area Investment Income Debits (% of GDP)

7

euro area Net Investment Income (% of GDP)

7

euro area Primary Surplus (% of GDP)

20081

20073

20071

-2

20063

-1

20061

-1

20053

0

20051

0

20043

1

20041

1

20033

2

20031

2

20023

3

20021

3

20013

4

20011

4

20003

5

20001

5

19993

6

19991

6

-2

Source: Eurostat and ECB

08 Book.indb 570

2/13/09 3:59:17 PM

Central Banks and Financial Crises

571

traded to non-traded goods). Such a depreciation of the real exchange rate is an essential part of the mechanism for shifting resources from the non-traded sectors (construction, domestic banking and financial services) to the tradable sectors (manufacturing, tourism, international banking and financial services, and other tradable services). The end of Ponzi finance for the US and the UK My view that the US and the UK will have to achieve a large external primary balance correction to maintain external solvency is based on the assumption that, in the future, rt p > gtp , i.e. that permanent Ponzi finance (a growth rate of the debt permanently greater than the interest rate on the debt) will not be possible for the US or the UK. I am therefore asserting that the future will, in this regard, be quite unlike the past. In the past couple of decades, as is clear from Charts 8b, 9b and 10b, both the US and the UK have been net debtor nations that received a steady stream of net payments from their creditors. As regards the net foreign asset income payments recorded in the balance of payments accounts, it looks therefore as though the US and the UK have not only been able, in the past, to engage in (temporary) Ponzi finance, they appear to have paid an effective negative nominal rate of return on their net external liabilities: Net Foreign investment Income is positive for the US and the UK (zero for the euro area) even though the Net International Investment Position is negative for all three. If this could be sustained, it would be a form of “über-Ponzi finance.” The reliability of the data summarized in Charts 8a,b, 9a,b and 10a,b is much debated, and the interpretation of the anomaly of a net debtor getting paid by his creditors is disputed (see e.g. Buiter, 2006; Gourinchas and Rey, 2007; and Hausmann and Sturzenegger, 2007). Part of the reason the US, the UK and (to a lesser extent) the euro area have been able to earn a much higher rate of return on their external assets than the rate of return earned by foreigners on their investments in the US and the UK, is that the US and the UK (Wall Street and the City of London) have, first, been acting as bankers to the world, providing unique liquidity and security for investments made in or channelled through these countries and, second, (may)

08 Book.indb 571

2/13/09 3:59:17 PM

572

Willem H. Buiter

have been acting as venture capitalists to the world (Gourinchas and Rey, 2007), earning a much higher return on US FDI abroad than foreigners earned on FDI in the US. I have my doubts about the reliability of the data on which this second mechanism is based, but not on the historical accuracy of the first. It is my belief that the North Atlantic region financial crisis will do great and lasting damage to the ability of the US and the UK to borrow cheaply and invest in assets yielding superior rates of return. Wall Street and the City of London have traded on the liquidity of their institutions and markets. Their leading banks and other financial institutions have benefited from huge liquidity premia and favourable risk spreads. These spreads reflected in part the perceived security of the investments that Wall Street and the City of London managed for clients or for their proprietary accounts.27 More fundamentally, it reflected global confidence and trust in the absence of malfeasance and gross incompetence. These valuable virtues and talents could be found only among the professionals in the heartland of financial capitalism. These unique assets, including trust and confidence, have been damaged badly. Key markets and institutions became illiquid and continue to be so. Incompetence, unethical practices and, not infrequently, outright illegal behaviour are now associated in the minds of the global investing community with many of the former giants of global finance in Wall Street and the City of London. That is why I have no serious reservations about assuming that, even for the US and the UK, we will have rt p > gtp in the future: For the first time in a long time, the external intertemporal budget constraint will bite. The rest of the world is unlikely to continue to provide the US and UK consumer (private or public) with credit on the terms of the past. The current financial crisis was made in the heartland of financial capitalism—on Wall Street, in the City of London, in Zurich and Frankfurt. It has revealed fundamental flaws in the heart of the financial system of the North Atlantic region. For many investors, the old, lingering suspicion that self-regulation meant no regulation has been confirmed. Those who sold or tried to sell this defective financial system to the rest of the world have been exposed as frauds or fools.

08 Book.indb 572

2/13/09 3:59:17 PM

Central Banks and Financial Crises

573

The rest of the world will not see the US (and the US dollar) or the UK (and sterling) or even the euro area and the euro as uniquely safe havens and as providers of uniquely safe and secure financial instruments. Risk premia for lending to the US and the UK are bound to increase significantly, even if there is no US dollar or sterling crisis. The position of New York and London as bankers to the world, and especially to the emerging markets, will be permanently impaired. How and when to boost the external balance If a large permanent decline in the ratio of domestic absorption to GDP is necessary, why wait, even if you could? Postponing the necessary adjustment will just raise the magnitude of the permanent correction that is eventually required. Five percentage points of GDP (a likely underestimate of the correction that is required) is already a very large permanent correction. Escalating that number further through inaction or, worse, through actions aimed at boosting consumption demand in the short run, risks destroying the credibility of an eventual adjustment. In addition, the terms of access to external finance can be expected to worsen rapidly for the US and the UK if durable adjustment measures are not implemented soon. I believe that the required permanent reduction in domestic absorption relative to GDP in the US ought to come mainly through a reduction in private consumption. Public spending on goods and services in the US is already low by international standards. Underfunded public services and substandard infrastructure also support the view that exhaustive public spending should not be cut significantly. US private investment rates are not particularly high, either by historical or by international standards. There is also the need to invest on a large scale in energy security, energy efficiency and other green ventures. While a cyclical weakening of energy prices can be expected, the trend is likely to be upwards. The US is far less energy-efficient in production and consumption than Europe or Japan, and much of the US stocks of productive equipment and consumer durables (including housing) will have to be scrapped or adapted to make them economically viable at the new high real energy prices. US investment rates, private and public, should therefore not fall.

08 Book.indb 573

2/13/09 3:59:17 PM

574

Willem H. Buiter

That leaves private consumption as the domestic spending or absorption component to be lowered permanently by at least five percentage points of GDP. The argument that the US will have to go through a protracted and/or deep slowdown to achieve a sustainable external balance is not dependent on whether it is private or public consumption that needs to be cut. The US national saving rate is astonishingly low, both by international and by historical standards, as is apparent from Table 8. Of the G7 countries, only the UK comes close to saving as little as the US. The belief that saving is unnecessary because capital gains will provide the desired increase in real financial wealth has been undermined by the successive implosions of all recent asset booms/ bubbles, including the tech bubble (which burst in late 2000) and the housing bubble (which burst at the end of 2006). It is logically possible that a country like the US can reduce consumption as a share of GDP by five percentage points or more without this causing a temporary slowdown in economic activity. Asset markets (including the real interest rate and the real exchange rate) could adjust promptly and by the right amount to provide the correct signals for a reallocation of resources from consumption to domestic and foreign investment and from the non-traded to the traded sectors. Prices of goods and services and factor prices could respond promptly to re-enforce these asset market signals. Real resource mobility between the traded and non-traded sectors could be high enough to permit a sizable intersectoral reallocation of labour and capital without the need for periods of idleness or inactivity. Absent a supply-side miracle, however, I believe that the US economy is too Keynesian in the short run to produce such a seamless and painless change in the composition of domestic production and in source of demand for domestically produced goods and services unless the right enabling macroeconomic policies are implemented. Although most policies and events that raise the national saving rate will result in a temporary decline in effective demand, in slowing or negative growth and in rising unemployment, in principle, the right combination of fiscal tightening and monetary loosening could boost

08 Book.indb 574

2/13/09 3:59:17 PM

Central Banks and Financial Crises

575

Table 8 Gross national saving rates for the G7 Percent of nominal GDP 1990

2000

2001

2002

2003

2004

2005

2006

2007

Canada

17.3

23.6

22.2

21.2

21.4

22.8

23.7

24.3

..

France

20.8

21.6

21.3

19.8

19.1

19.0

18.5

19.1

19.3

Germany

25.3

20.2

19.5

19.4

19.5

21.5

21.8

23.0

25.2

Italy

20.8

20.6

20.9

20.8

19.8

20.3

19.6

19.6

19.7

Japan

33.2

27.5

25.8

25.2

25.4

25.8

26.8

26.6

..

United Kingdom

16.5

15.4

15.6

15.8

15.7

15.9

15.1

14.9

..

United States

15.3

17.7

16.1

13.9

12.9

13.4

13.5

13.7

..

Note: Based on SNA93 or ESA95 except Turkey that reports on SNA68 basis. Sources: OECD, National accounts of OECD countries database.



the external primary deficit without changing aggregate demand for domestic output.28 Unfortunately, instead of fiscal tightening we have had discretionary fiscal loosening in the US worth about $150 billion since the crisis began. With these perverse fiscal policies in the US (from the perspective of restoring external balance), the re-orientation of domestic production towards tradables and the switch of global demand towards domestic goods is delayed and will ultimately be made more painful. It is therefore ironic, and to me incomprehensible, that leading economists who have argued for decades that US households need to save more would, as soon as the US consumer is at long last showing signs of wanting to save more (that is, consume less), propose fiscal and monetary measures aimed at stopping the US consumer from doing what (s)he ought to have been doing all along. Martin Feldstein (2008) is a notable example; Larry Summers (2008) is another. This is a vivid example of St. Augustine’s: “Lord, give me chastity and virtue, but do not give it yet.” The fall in private consumption growth, and indeed in private consumption, should be welcomed, not fought.

08 Book.indb 575

2/13/09 3:59:17 PM

576

Willem H. Buiter

The Chairman of the Fed also appears to dropped the qualifier “sustainable” from the objectives of growth and employment. Statements by Chairman Bernanke like the following abound: “...we stand ready to take substantive additional action as needed to support growth and to provide additional insurance against downside risks”(Bernanke, 2008a). The omission of the word “sustainable” in front of growth is no accident. The Fed has chosen to do all it can to maintain output and employment at the highest possible levels, with no regard to their sustainability. III.1a(v)

How dangerous to the real economy is financial sector deleveraging?

Consider the following stylized description of the financial system in the North Atlantic region in the 1920s and 1930s. Banks intermediate between households and non-financial corporations. There is a reasonable-size stock market, a bond market and a foreign exchange market. Banks are the only significant financial institutions—the financial sector is but one layer deep. When the financial sector is but one layer deep, the collapse of the net worth of financial sector institutions and the contraction of the gross balance sheet of the financial sector can seriously impair the entire intermediation process. The spillovers into the real economy—household spending and investment spending by non-financial corporates—are immediate and direct. This was the picture in the Great Depression of the 1930s. This is the world studied in depth by the current Fed Chairman, Ben Bernanke, but it is not the world we live in today. Today, the financial sector is many layers deep. Most financial institutions interact mainly with other financial institutions rather than with households or non-financial enterprises. They lend and borrow from each other and invest in each others’ contingent claims. Part of this financial activity is socially productive and efficiency-enhancing. Part of it is privately profitable but socially wasteful churning, driven by regulatory arbitrage and tax efficiency considerations. During periods of financial boom and bubble, useless financial products and pointless financial enterprises proliferate, often achieving enormous

08 Book.indb 576

2/13/09 3:59:17 PM

Central Banks and Financial Crises

577

scale. Finance is, after all, trade in promises, and can be scaled almost costlessly, given optimism, confidence, trust and gullibility. Interestingly, during the most recent leverage boom, many of the non-bank financial businesses that accounted for much of the increase in leverage, chose to hold a non-negligible part of their assets as bank deposits and also borrowed from banks on a sizable scale. So the growth of bank credit to non-bank financial entities and the growth of the broad monetary aggregates tracked the financial, credit and leverage boom quite well. We don’t know whether this is a stable structural relationship or just a fragile co-movement between jointly endogenous variables. Still, it suggests that central banks that take their financial stability role seriously should pay attention to the broad monetary aggregates and to the behaviour of bank credit, even if these aggregates are useless in predicting inflation or real economic activity in real time (see e.g. Adalid and Detken, 2007, and Greiber and Setzer, 2007). The visible sign of this growth of intra-financial sector intermediation/churning is the growth of the gross balance sheets of the financial sector and the growth of leverage, both in the strict sense of, say, assets to equity ratios and in the looser sense of the ratio of gross financial sector assets or liabilities to GDP. During the five years preceding the credit crunch, this financial leverage was rising steadily, without much apparent impact on actual or potential GDP. If it had to be brought back to its 2002 level over, say, a five-year period, it is likely that no one would notice much of an impact on real or potential GDP. The orderly, gradual destruction of “inside” assets and liabilities need not have a material impact on the value of the “outside” assets and on the rest of the real economy. But financial sector deleveraging and leveraging are not symmetric processes, in the same way that assets price booms and busts are not symmetric. Compared to the deleveraging phase, the increasing leverage phase is gradual. Rapid deleveraging creates positive, dysfunctional feedback between falling funding liquidity, distress sales of assets, low market liquidity, falling asset prices and further tightening of funding liquidity.

08 Book.indb 577

2/13/09 3:59:17 PM

578

Willem H. Buiter

At some point, the deleveraging, even though it still involves almost exclusively the destruction of inside assets (and the matching inside liabilities), will impair the ability of the financial sector as a whole to supply finance to financial deficit units in the household sector and the non-financial corporate sector. Among the outside assets whose value collapses is the equity of the banks and other financial intermediaries. Given external (regulatory) and internal prudential lower limits on permissible or desirable capital ratios, these intermediaries are faced with the choice of reducing or suspending dividends, initiating rights issues or restricting lending to new or existing customers. Inevitably, lending is cut back and the financial crunch is transmitted to households and non-financial enterprises. The LLR and MMLR roles of the central bank, backed by the Treasury, are designed to prevent excessively speedy, destructive deleveraging. If it does that, there can be massive gradual deleveraging in the financial sector, without commensurate impact on households and nonfinancial corporates. Inside and outside assets I believe that the Fed has consistently overestimated the effect of the overdue sharp contraction in the size of the financial sector balance sheet on the real economy. Much of this can, I believe, be attributed to a failure to distinguish carefully between inside and outside assets. All financial instruments are inside assets. If an inside asset loses value, there is a matching decline in an inside liability. Both should always be considered together. This has not been common practice. Just one example. Even before August 9, 2007, Chairman Bernanke provided estimates of the loss the US banking sector was likely to suffer on its holdings of subprime mortgages due to write-downs and write-offs on the underlying mortgages. For instance, on July 20, 2007, in testimony to Congress, Chairman Bernanke stated subprime-related losses could be up to $100 billion out of a total subprime mortgage stock of around $2 trillion; there have been a number of higher estimates since then. Not once have I heard a member of the FOMC reflect on the corresponding gain on the balance sheets

08 Book.indb 578

2/13/09 3:59:17 PM

Central Banks and Financial Crises

579

of the mortgage borrowers. Mortgages are inside assets/liabilities. So are securities backed by mortgages. Consider a household that purchases for investment purposes a second home worth $400,000 with $100,000 of its own money and a non-recourse mortgage of $300,000 secured against the property.29 Assume the price of the new home halves as soon as the purchase is completed. With negative equity of $100,000 the homeowner chooses to default. The mortgage now is worth nothing. The bank forecloses, repossesses the house and sells it for $200,000, spending $50,000 in the process. The loss of net wealth as a result of the price collapse and the subsequent default and repossession is $250,000: the $200,000 reduction in the value of the house and the $50,000 repossession costs (lawyers, bailiffs, etc.) The homeowner loses $100,000: his original, pre-price collapse equity in the house—the difference between what he paid for the house and the value of the mortgage he took out. The bank loses $150,000: the sum of the $100,000 excess of the value of the mortgage over the post-collapse price of the house and the $50,000 real foreclosure costs. The $300,000 mortgage is an inside asset—an asset to the bank and a liability to the homeowner-borrower. When it gets wiped out, the borrower gains (by no longer having to service the debt) what the lender loses. The legal event of default and foreclosure, however, is certainly not neutral. In this case it triggers a repossession procedure that uses up $50,000 of real resources. This waste of real resources would, however, constitute aggregate demand in a Keynesian-digging-holesand-filling-them-again sense, a form of private provision of pointless public works. Continuing the example, how does the redistribution, following the default, of $100,000 from the bank to the defaulting borrower— the write-off of the excess of the face value of the mortgage over the new low value of the house—affect aggregate demand? There is one transmission channel that suggests it is likely that demand would have been weaker if, following the default, the lender had continued recourse to the borrower (say, through a lien on the borrower’s future

08 Book.indb 579

2/13/09 3:59:17 PM

580

Willem H. Buiter

income or assets). The homeowner-borrower is likely to have a higher marginal propensity to spend out of current resources than the owners of the bank—residential mortgage borrowers are more likely to be liquidity-constrained than the shareholders of the mortgage lender. This transmission channel has, as far as I can determine, never been mentioned by any FOMC member. Finally, we have to allow for the effect of the mortgage default on the willingness and ability of the bank to make new loans and to roll over existing loans. Clearly, the write-off or write-down of the mortgage will put pressure on the bank’s capital. The bank can respond by reducing its dividends, by issuing additional equity or by curtailing lending. The greatest threat to economic activity undoubtedly comes from curtailing new lending and the refusal to renew maturing loans. The magnitude of the effect on demand of a cut in bank lending depends on whom the banks are lending to and what the borrower uses the funds for. If the banks are lending to other financial intermediaries that are, directly or indirectly, lending back to our banks, then there can be a graceful contraction of the credit pyramid, a multilayered deleveraging without much effect on the real economy. If bank A lends $1 trillion to bank B, which then uses that $1 trillion to buy bonds issued by bank A, there could be a lot of gross deleveraging without any substantive impact on anything that matters. With a few more near-bank or non-bank intermediaries interposed between banks A and B, such intra-financial sector lending and borrowing (often involving complex structured products) has represented a growing share of bank and financial sector business this past decade. In our non-Modigliani-Miller world, financial structure matters. We cannot just “net out” inside financial assets and liabilities—they are an essential part of the transmission mechanism. But there also is no excuse for ignoring half of the distributional effects inherent in changing valuations of inside assets and liabilities. If their public statements are anything to go by, the Fed and the FOMC may have systematically overestimated the effects of declining inside financial asset valuations on aggregate demand.

08 Book.indb 580

2/13/09 3:59:17 PM

Central Banks and Financial Crises

III.1a(vi)

581

Disdain for the monetary aggregates

Monetary targeting for macroeconomic stability died because the velocity of circulation of any monetary aggregate turned out to be unpredictable and unstable. Even so, the decision to cease publishing M3 statistics effective 23 March 2006 was extraordinary. The reason given was: “M3 does not appear to convey any additional information about economic activity that is not already embodied in the M2 aggregate. The role of M3 in the policy process has diminished greatly over time. Consequently, the costs of collecting the data and publishing M3 now appear to outweigh the benefits.” Information is probably the purest of all pure public goods. The cost-benefit analysis argument against its continued publication, free of charge to the ultimate user, by a public entity like the Fed, is completely unconvincing. Broad monetary aggregates, including M3 and their counterparts on the asset side of the banking sector’s balance sheet are in any case informative for those interested in banking sector leverage and other financial stability issues, including asset market booms and bubbles (see e.g. Ferguson, 2005; Adalid and Detken, 2007; and Greiber and Setzer, 2007). The decision to discontinue the collection and publication of M3 data supports the view that the Fed took its eye off the credit boom ball just as it was assuming epic proportions. The decision to discontinue publication of the M3 series also smacks of intellectual hubris; effectively, the Fed is saying: We don’t find these data useful. Therefore you shall not have them free of charge any longer. III.1b The world imports inflation All three central banks have tried to absolve themselves of blame for the recent bouts of inflation in their jurisdictions by attributing much or most of it to factors beyond their control—global relative price shocks, global supply shocks, global inflation or global commodity price inflation. A prominent use of this fig-leaf can be found in the open letter to the Chancellor of the Exchequer by Mervyn King, Governor of the BoE, in May 2008.30 The gist of the Governor’s analysis was: it’s all global commodity prices—something beyond our control.

08 Book.indb 581

2/13/09 3:59:18 PM

582

Willem H. Buiter

I will quote him at length, so there is no risk of distortion: “Inflation has risen sharply this year, from 2.1% in December to 3.3% in May. That rise can be accounted for by large and, until recently, unanticipated increases in the prices of food, fuel, gas and electricity. These components alone account for 1.1 percentage points of the 1.2 percentage points increase in the CPI inflation rate since last December. Those sharp price changes reflect developments in the global balance of demand and supply for foods and energy. In the year to May: • world agricultural prices increased by 60% and UK retail food prices by 8%. • oil prices rose by more than 80% to average $123 a barrel and UK retail fuel prices increased by 20% • wholesale gas prices increased by 160% and UK household electricity and gas bills by around 10% The global nature of these price changes is evident in inflation rates not only in the UK but also overseas, although the timing of their impact on consumer prices differs across countries. In May, HICP inflation in the euro area was 3.7% and US CPI inflation was 4.2%.” Later on in the open letter the Governor amplifies the argument that this increase in inflation has nothing to do with the BoE: “There are good reasons to expect the period of above-target inflation we are experiencing now to be temporary. We are seeing a change in commodity, energy and import prices relative to the prices of other goods and services. Although this clearly raises the price level, it is not the same as continuing inflation. There is not a generalised rise in prices and wages caused by rapid growth in the amount of money spent in the economy. In contrast to past episodes of rising inflation, money spending is increasing at a normal rate. In the year to 2008 Q1, it rose by 5½%, in line with the average rate of increase since 1997—a period in which inflation has been low and stable. Moreover, in recent months the growth rate of the broad money supply has eased and credit conditions have tightened. This will restrain the growth of money spending in the future.” (emphasis in the original).

08 Book.indb 582

2/13/09 3:59:18 PM

Central Banks and Financial Crises

583

Very similar statements have been made by President Jean-Claude Trichet of the ECB and Chairman Ben Bernanke. Here is a quote from the August 7, 2008 Introductory statement before the press conference by President Trichet: “...Annual HICP inflation has remained considerably above the level consistent with price stability since last autumn, reaching 4.0% in June 2008 and, according to Eurostat’s flash estimate, 4.1% in July. This worrying level of inflation rates results largely from both direct and indirect effects of past sharp increases in energy and food prices at the global level” (Trichet, 2008). Ditto for Chairman Bernanke (2008): “Inflation has remained high, largely reflecting sharp increases in the prices of globally traded commodities.” And, in the same speech : “Rapidly rising prices for globally traded commodities have been the major source of the relatively high rates of inflation we have experienced in recent years, underscoring the importance for policy of both forecasting commodity price changes and understanding the factors that drive those changes.” This analysis makes no sense. Except at high frequencies, headline inflation can be effectively targeted and controlled by the monetary authority and is therefore the responsibility of the monetary authority. Supply shocks or demand shocks make the volatility of actual headline inflation around the target higher, but should not create a bias. The only obvious caveat is that the economy in question have a floating effective exchange rate. This is the case for the UK and the euro area. The US is hampered somewhat in its monetary autonomy by the fact that the Gulf Cooperation Council members and some other countries continue to peg to the US dollar, and by the fact that the exchange rate with the US dollar of the Chinese Yuan continues to be managed in a rather unhelpful manner by the Chinese authorities. Although the Yuan appreciated vis-à-vis the US dollar by more than 10 percent in 2007 and by more than 7 percent so far this year, it is clearly not a market-determined exchange rate.

08 Book.indb 583

2/13/09 3:59:18 PM

584

Willem H. Buiter

If we add together the statements by the world’s central bank heads (from the industrial countries, from the commodity-importing emerging markets and from the commodity exporting emerging markets) on the origins of their countries’ inflation during the past couple of years, we must conclude that interplanetary trade is now a fact: The world is importing inflation from somewhere else (Wolf, 2008). Consider the following stylised view of the inflation process in an open economy. The consumer price level, as measured by the CPI, say, is a weighted average of a price index for core goods and services and a price index for non-core goods. Core goods and services have sticky prices—these are the prices that account for Keynesian nominal rigidities (money wages and prices that are inflexible in the short run) and make monetary policy interesting. Non-core goods are commodities traded in technically efficient auction markets. It includes oil, gas and coal, metals and agricultural commodities, both those that are used for food production and those that provide raw materials for industrial processing, including bio fuels. The prices of non-core goods are flexible. I will treat the long-run equilibrium relative price of core and noncore goods and services as determined by the rest of the world. In the short run, nominal rigidities can, however, drive the domestic relative price away from the global relative price. I also make domestic potential output of core goods and services a decreasing function of the relative price of non-core goods to that of core goods and services. The effect of an increase in real commodity prices on productive potential in the industrial countries is empirically well-established. A recent study by the OECD (2008) suggests that the steady-state effect of a $120 per barrel oil price could be to lower the steady-state path of US potential output by about four percentage points, and that of the euro area by about half that (reflecting the lower euro area energy-intensity of GDP).31 The short-and medium-term effect on the growth rate of potential output in the US of the real energy price increase would be about 0.2 percent per annum, and half that in the euro area. Negative effects on potential output of the higher cost of capital since the summer of 2007 could magnify the negative potential growth rate effects, according to the

08 Book.indb 584

2/13/09 3:59:18 PM

Central Banks and Financial Crises

585

OECD study, to minus 0.3 percent per annum for both the US and the euro area. I also treat the world (foreign currency) price of non-core goods as exogenous. It simplifies the analysis, but is not necessary for the conclusions, if we assume that the country produces only core goods and services and imports all non-core goods. Non-core goods are both consumed directly and used as imported raw materials and intermediate inputs in the production of core goods and services. The weight of non-core goods in the CPI, which I will denote μ, represents both the direct weight of non-core goods in the consumption basket and the indirect influence of core goods prices as a variable cost component in the production of core goods and services. I haven’t seen any up-to-date input-output matrices for the US, the euro area and the UK, so I will have to punt on μ. For illustrative purposes, assume that μ = 0.25 for the UK, 0.10 for the US and 0.15 for the euro area. The inflation rate is the proportional rate of change of the CPI. If p is the CPI inflation rate, pc the core inflation rate and pn the noncore inflation rate, then: p =(1-μ ) pc + μ pn

(21)

The inflation rate of non-core goods measured in domestic currency prices is the sum of the world rate of inflation of non-core goods pf and the proportional rate of depreciation of the currency’s nominal exchange rate, e. That is, pn= pf+ e

(22)

By assumption, the central bank has no influence on the world rate of inflation of non-core goods, pf. The same cannot be said, however, for the value of the nominal exchange rate. High global inflation need not be imported if the currency is permitted to appreciate. In the UK, between end of the summer of 2007 and the time of Governor King’s open letter in May 2008, sterling’s effective exchange rate depreciated by 12 percent, reinforcing rather than offsetting the domestic inflationary effect of global price increases. The heads of our three central banks appear to treat the nominal exchange rate as exogenous—independent of monetary policy.32

08 Book.indb 585

2/13/09 3:59:18 PM

586

Willem H. Buiter

The values of μ are probably quite reasonable, but the one-for-one instantaneous structural pass-through assumed in equation (22) for exchange rate depreciation on the domestic currency prices of noncore goods is somewhat over the top, at any rate in the short run. But it is a reasonable benchmark for medium- and long-term analysis. In the short run, one can, for descriptive realism, add a little distributed lag or error-correction mechanism to (22), reflecting pricing-to-market behaviour etc. Core inflation, which can be identified with domestically generated inflation in the simplest version of this approach, depends on such things as the inflation rate of unit labour costs and of unit rental costs plus the growth rate of the mark-up. For simplicity, I will assume that core inflation depends on the domestic output gap,y- ŷ, on expected future headline inflation, Etpt+1 and on past core inflation, so core inflation is driven by the following process:

p tc = γ ( yt − yˆt ) + β Et p t +1 + (1 − β )p tc−1 γ > 0, 0 < β ≤ 1

(23)

Monetary policy influences core inflation through two channels: by raising interest rates and expectations of future policy rates, it can lower output and thus the output gap. And if past, current and anticipated future actions influence expectations of future CPI inflation, that too will reduce inflation today, through the (headline) expectations channel. It is true that an increase in the relative price of non-core goods to core goods and services means, given a sticky nominal price of core goods and services, an increase in the general price level but not, in and of itself, ongoing inflation. That is arithmetic. With the domestic currency price of core goods and services given in the short run, the only way to have an increase in the relative price of non-core goods is to have an increase in the domestic currency price of non-core goods. The level of the CPI therefore increases. This one-off increase in the general price level will show up in real time as a temporary increase in CPI inflation. If there is a sequence of such relative price increases, there will be a sequence of such temporary increases in CPI inflation, which will rather look like, but is not, ongoing inflation.

08 Book.indb 586

2/13/09 3:59:18 PM

Central Banks and Financial Crises

587

Of course, as time passes even sticky Keynesian prices become unstuck. The nominal price of core goods and services can and does adjust. It can even adjust in a downward direction, as the spectacular declines in IT-related product prices illustrate on a daily basis. Whether the medium-term and longer-term increase in the relative price of non-core goods and services will continue to be reflected in a higher future path for the CPI, an unchanged CPI path or even an ultimately lower CPI path, is determined by domestic monetary policy. Furthermore, an increase in the relative price of non-core goods to core goods and services does more than cause a one-off increase in the price level. As argued above, and as supported by many empirical studies, including the recent OECD (2008) study cited above, it reduces potential output or productive capacity by making an input that is complementary with labour and capital more expensive.33 Letpn pc

ting denote the relative price of non-core and core goods, I write this as: yˆt = yˆt − η

η>0

ptn ptc

(24)

In addition, if labour supply is responsive to the real consumption wage, then the adverse change in the terms of trade that is the other side of the increase in the relative price of non-core goods to core goods and services will reduce the full-employment supply of labour, and this too will reduce productive capacity. Thus, unless actual output (aggregate demand) falls by more than potential output as a result of the adverse terms of trade change, the output gap will increase and the increase in the relative price of non-core goods will raise domestic inflationary pressures for core goods and services. Clearly, the adverse terms of trade change will lower the real value of consumption demand, measured in terms of the consumption basket, if claims on domestic GDP (capital and labour income) are owned mainly by domestic consumers. It lowers the purchasing power of domestic output over the domestic consumption bundle. Real income measured in consumer goods falls, so real consumption

08 Book.indb 587

2/13/09 3:59:19 PM

588

Willem H. Buiter

measured in consumer goods should fall. But even if the increase in the relative price of non-core goods is expected to be permanent, real consumption measured in terms of the consumption bundle is unlikely to fall by a greater percentage than the decline in the real consumption value of domestic production. With homothetic preferences, a permanent deterioration in the terms of trade will not change consumption measured in terms of GDP units. If the period utility function is Cobb-Douglas between domestic output and imports, the adverse terms of trade shock lowers potential output but does not reduce domestic consumption demand for domestic output. Unless the sum of investment demand for domestic output, public spending on domestic output and export demand falls in terms of domestic output, aggregate demand (actual GDP) will not fall. The output gap therefore increases as a result of an increase in the relative price of non-core goods to goods and services. Domestic inflationary pressures rise. Interest rates have to rise to achieve the same inflation trajectory. This inflationary impact of the increase in the relative price of commodities appears to be ignored by the Governor, the President and the Chairman. III.1c False comfort from limited “pass-through” of inflation expectations into earnings growth? Both the Fed and the BoE (less so the ECB) take comfort from the fact that earnings growth has remained moderate despite the increase in inflation expectations, based on both break-even inflation calculations (or the inflation swap markets) and on survey-based expectations. For instance, in the exchange of letters between the Governor of the BoE and the Chancellor in May 2008, it was noted by the Chancellor that, although median inflation expectations for the coming year had risen to 4.3 percent in the Bank’s own survey, earnings growth (including bonuses) is running at only 3.9 percent. However, this observation does not mean that inflation expectations are not translated, ceteris paribus, one-for-one into higher wage settlements or into higher actual inflation. Time series analysis (earning growth is not rising) is not the same as counterfactual analysis

08 Book.indb 588

2/13/09 3:59:19 PM

Central Banks and Financial Crises

589

(earnings growth would have been the same if inflation expectations had not risen). It is certainly possible that the global processes that have depressed the share of labour income in GDP in most industrial countries during the past 10 years (labour-saving technical change, China and India entering the global markets as producers of goods and services that are frequently competitive with those produced by the labour force in the advanced industrial countries, increased cross-border labour mobility, legal constraints weakening labour unions etc.) have not yet run their course and that labour’s share will continue to decline. Arithmetically, a decrease in labour’s share in GDP is an increase in the mark-up of the GDP deflator on unit labour costs. So if an increase in the expected rate of (consumer price) inflation coincided with a reduction in labour’s share of GDP because of structural factors (and if no other determinant of earnings growth changed), unit labour cost growth could well rise (in a time-series sense) by less than the increase in expected inflation or might even decline. The price inflation process (on the GDP deflator definition) would, however, include the growth rate of the mark-up on unit labour costs, and would show the full impact of the increase in expected inflation (even in a time-series sense). Clearly, the GDP deflator is not quite the same as the core price index, but qualitatively, the point remains valid, that a declining equilibrium share of labour will be offset, in the price inflation process, by a rising equilibrium mark-up on unit labour cost and that this can distort the interpretation of simple correlations between inflation expectations and earnings growth. III.2 Financial stability: LLR, MMLR and Quasi-fiscal actions III.2a The Fed The Fed, as soon as the crisis hit, injected liquidity into the markets at maturities from overnight to three-months. The amounts injected

08 Book.indb 589

2/13/09 3:59:19 PM

590

Willem H. Buiter

were somewhere between those of the BoE (allowing for differences in the size of the US and UK economies) and those of the ECB. III.2a(i)

Extending the maturity of discount window loans

On August 17, 2007 the Fed extended the maturity of loans at the discount window from overnight to up to one month. On March 16, 2008, it further extended the maximum term for discount window lending to 90 days. These were helpful measures, permitting the provision of liquidity at the maturities it was actually needed. III.2a(ii)

The TAF

On December 12, 2007, the Fed announced the creation of a temporary term auction facility (TAF). This allows a depository institution to place a bid for a one-month advance from its local Federal Reserve Bank at an interest rate that is determined as the result of an auction. The TAF allows the Fed to inject term funds through a broader range of counterparties and against a broader range of collateral than open market operations. When the normal open market operations counterparties are hoarding funds, and the unsecured interbank market is not disseminating liquidity provisions efficiently throughout the banking sector, this facility is clearly helpful. III.2a(iii)

International currency swaps

Also on December 12, the Fed announced swap lines with the European Central Bank and the Swiss National Bank of $20 billion and $4 billion, respectively. On March 11, 2008, these swap lines were increased to $30 billion and $6 billion, respectively. This, I have suggested earlier, represents either the confusion of motion with action or an unwarranted subsidy to the private banks able to gain access to this foreign exchange rather than having to acquire it more expensively through the private swap markets. Banks in the euro area and Switzerland were not liquid in euros/Swiss francs but short of US dollars because the foreign exchange markets had become illiquid. These banks were short of liquidity—full stop—that is, short of liquidity in any currency.

08 Book.indb 590

2/13/09 3:59:19 PM

Central Banks and Financial Crises

591

This is unlike the case of Iceland, where the Central Bank on May 16, 2008, arranged swaps for euros with the three Scandinavian central banks. Since the Icelandic banking system is very large relative to the size of the economy and has much of its balance sheet (including a large amount of short-term liabilities) denominated in foreign currencies rather than in Icelandic kroner, the effective performance of the LLR and MMLR functions requires the central bank to have access to foreign currency liquidity. With no one interested in being long Icelandic kroner, the swap facilities are an essential line of defence for the Icelandic LLR/MMLR III.2a(iv)

The TSLF

On March 11, 2008, the Fed announced that it would expand its existing overnight securities lending program for primary dealers by creating a Term Securities Lending Facility (TSLF). Under the TSLF, the Fed will lend up to $200 billion of Treasury securities held by the System Open Market Account to primary dealers secured for a term of 28 days by a pledge of other collateral. The Facility was extended beyond the 2008 year-end in July 2008, and the maturity of the loans was increased to three months. The first TSLF auction took place on March 27, with $75 billion offered for a term of 28 days, too late to be helpful to Bear Stearns, for which the Fed had to provide extraordinary LLR support on March 14. The price is set through a single-price auction.34 The range of collateral is quite wide: all Schedule 2 collateral plus agency collateralized-mortgage obligations (CMOs) and AAA/Aaarated commercial mortgage-backed securities (CMBS), in addition to the AAA/Aaa-rated private-label residential mortgage—backed securities (RMBS) and OMO-eligible collateral.35 Until the creation of the Primary Dealer Credit Facility (PDCF, see below) the Fed could not lend cash directly to primary dealers. Instead it lends highly liquid Treasury bills which the primary dealers then can convert into cash. This facility extends both the term of the loans from the Fed to primary dealers and the range of eligible collateral. In principle this is a useful arrangement for addressing a liquidity crisis. The design, however, has one huge flaw.

08 Book.indb 591

2/13/09 3:59:19 PM

592

Willem H. Buiter

An extraordinary feature of the arrangement is that the collateral offered by a primary dealer is valued by the clearing bank acting as agent for the primary dealer.36 Apparently this is a standard feature of the dealings between the Fed and the primary dealers. Primary dealers cannot access the Fed directly, but do so through a clearing bank—their dealer. As long as the clearing bank which acts as agent for the primary dealer in the transaction is willing to price the security (say, by using an internal model), the Fed will accept it as collateral at that price. The usual haircuts, etc., will, of course, be applied to these valuations. This arrangement is far too cosy for the primary dealer and its clearer. The incentive for collusion between the primary dealer and the clearer, to offer pig’s ear collateral but value it as silk purse collateral, will be hard to resist. This invites adverse selection: The Fed is likely to find itself with overpriced, substandard collateral. Offering access to this adverse selection mechanism today creates moral hazard in the future. It does so by creating incentives for future reckless lending and investment by primary dealers aware of these future opportunities for dumping bad investments on the Fed as good collateral through the TSLF. More recently, the Fed extended the TSLF through the addition of a Term Securities Lending Facility Options Program (TOP). This rather looks to me like gilding the lily. III.2a(v)

The PDCF

On March 16, 2008, the Primary Dealer Credit Facility (PDCF) was established, for a minimum period of six months. This again was too late to be helpful in addressing the Bear Stearns crisis. Primary dealers of the Federal Reserve Bank of New York are eligible to participate in the PDCF via their clearing banks. It is an overnight loan facility that provides funding to primary dealers in exchange for a specified range of eligible collateral, including all collateral eligible for tri-party repurchase agreements arranged by the Federal Reserve Bank of New York (that is, all collateral eligible for pledge in open market operations), as well as all investment-grade corporate securities, municipal securities, mortgage-backed securities and asset-backed securities for which a price is available from the primary

08 Book.indb 592

2/13/09 3:59:19 PM

Central Banks and Financial Crises

593

dealer’s clearing bank. The rate charged is the one at the primary discount window to depositary institutions for overnight liquidity, currently 25 bps over the federal funds target rate. This facility effectively extends overnight borrowing at the Fed’s primary discount window to primary dealers, at the standard primary discount window rate. Note again the extraordinary valuation mechanism put in place for securities offered as collateral: “The pledged collateral will be valued by the clearing banks based on a range of pricing services.”37 This is the same “adverse-selection-today-leading-to-moralhazard-tomorrow-machine” created by the Fed with the TSLF. III.2a(vi)

Bear Stearns

On March 14, 2008, the Fed agreed to lend US$29 billion to Bear Stearns through JPMorgan Chase (on a non-recourse basis). Bear Stearns is an investment bank and a primary dealer. It was not regulated by the Fed (which only regulates depositary institutions) but by the SEC. Bear Stearns was deemed too systemically important (probably by being too interconnected rather than too big) to fail. It is not clear why Bear Stearns could not have borrowed at the regular Fed primary discount window. It is true that investment banks had not done so since the Great Depression, but it would have been quite consistent with the Fed’s legislative mandate. The Federal Reserve Act (1913) allows the Federal Reserve to lend, in a crisis, to just about any institution, organisation or individual, and against any collateral the Fed deems fit (see also Small and Clouse, 2004). Specifically, if the Board of Governors of the Federal Reserve System determine that there are “unusual and exigent circumstances” and at least five (out of seven) governors vote to authorize lending under Section 13(3) of the Federal Reserve Act, the Federal Reserve can discount for individuals, partnerships and corporations (IPCs) “notes, drafts and bills of exchange indorsed or otherwise secured to the satisfaction of the Federal Reserve bank….” The combination of the restriction of “unusual and exigent circumstances” and the further restriction that the Federal Reserve can discount only to IPCs “unable to secure adequate credit accommo-

08 Book.indb 593

2/13/09 3:59:19 PM

594

Willem H. Buiter

dations from other banking institutions,” fits the description of a credit crunch/liquidity crisis like a glove. So why did the Fed not determine before March 14 that there were “unusual and exigent circumstances” that would have allowed Bear Stearns direct access to the discount window? It is also a mystery why a special resolution regime analogous to that administered by the FDIC for insured depositary institutions (discussed in Section II.3a) did not exist for Bear Stearns. The experience of LTCM in 1998 should have made it clear to the Fed that there were institutions other than deposit-taking banks that might be too systemically significant to fail, precisely because, like Bear Stearns, their death throes might, through last-throw-of-the-dice asset liquidations, cause illiquid asset prices to collapse and set in motion a dangerous chain reaction of cumulative market illiquidity and funding illiquidity. An SRR could have ring-fenced the balance sheet of Bear Stearns and permitted the analogue of Prompt Corrective Action to be implemented. The entire top management could have been fired without any golden handshakes. If necessary, regulatory insolvency could have been declared for Bear Stearns. The shareholders would have lost their voting power and would have had to take their place in line, behind all other claimants. Outright nationalisation of Bear Stearns could have created a better alignment of incentives that was actually achieved, although a drawback of nationalisation would have been that all creditors of Bear Stearns would have been made whole. Instead we have a $10 per share payment for the shareholders, what looks like a sweetheart deal for JPMorgan Chase, and a $29 billion exposure for the US taxpayer to an SPV in Delaware, which has $30 billion of Bear Stearns” most toxic assets on its balance sheet. Only $1 billion of JPMorgan Chase money stands between losses on the assets and the $29 billion “loan with equity upside” provided by the Fed. III.2a(vii)

Bear Stearns’ bailout as an example of confusing the LLR and MMLR functions

The rescue of Bear Stearns represents the confusion of the lender-of-last-resort role of the traditional central bank and the market-maker-of-last-resort role of the modern central bank. Bear

08 Book.indb 594

2/13/09 3:59:19 PM

Central Banks and Financial Crises

595

Stearns was an investment bank. No investment bank is systemically important in the sense that no investment bank performs tasks that cannot be performed readily and with comparable effectiveness by other institutions. Even the primary dealer and broker roles of Bear Stearns could have been taken over promptly by the other primary dealers and brokers. Bear Stearns was rescued because it was “too interconnected to fail.” It was feared that, in a last desperate attempt to stave off insolvency, Bear Stearns would have unloaded large quantities of illiquid securities in dysfunctional, illiquid securities markets. This would have caused a further dramatic decline in the market prices of these securities. With mark-to-market accounting and through margin calls linked to these valuations, further sales of illiquid securities by distressed financial institutions would have been triggered. The losses associated with these “panic sales” would have reduced the capital of other financial institutions, requiring them to cut or eliminate dividends, raise new capital, cut new lending or reduce their investments. A vicious cycle could have been triggered of forced sales into illiquid markets triggering funding liquidity problems elsewhere, necessitating further liquidations of illiquid asset holdings. This chain of events is possible and may even have been plausible at the time. The solution, however, is to truncate the vicious downward spiral of market illiquidity and funding illiquidity right at the point where Bear Stearns was distress-selling its illiquid assets. By acting as MMLR—either by buying these securities outright or by accepting them as collateral at facilities like the TAF (extended to include investment banks as eligible counterparties), the TSLF or the PDCF—the central bank could have put a floor under the prices of these securities and would thus have prevented a vicious downward spiral of market and funding illiquidity. Whether Bear Stearns would have been able to survive with the valuations of their assets realised at these TAF-, TSLF- or PDCF-type facilities, would no longer have been systemically relevant. The arrangements for acting as MMLR for investment banks did not, unfortunately, exist when Bearn Stearns collapsed. Now that they do, they should be kept alive, on a stand-by or as-needed basis.

08 Book.indb 595

2/13/09 3:59:19 PM

596

Willem H. Buiter

They may have to be expanded to include other highly leveraged financial institutions that are too interconnected to fail. As quid pro quo, all institutions eligible for MMLR (and/or LLR) support should be subject to common regulatory requirements, including a common special resolution regime. Combined with a proper punitive pricing of securities offered for outright purchase or as collateral, moral hazard will be minimized. III.2a(viii)

Fannie and Freddie

On Sunday, July 13, 2008, the Fed, in a coordinated action with the Treasury, announced that it would provide the two GSEs, Fannie Mae and Freddie Mac, with access to the discount window on the same terms as commercial banks. The announcement was not very informative as regards the exact conditions of access: “The Board of Governors of the Federal Reserve System announced Sunday that it has granted the Federal Reserve Bank of New York the authority to lend to Fannie Mae and Freddie Mac should such lending prove necessary. Any lending would be at the primary credit rate and collateralized by U.S. government and federal agency securities. ....” It isn’t clear from this whether the two GSEs have access only to overnight collateral (at a rate 25 basis points over the federal funds target rate) or are able to obtain loans of up to 3-month maturity, as commercial banks can. As long as the collateral the Fed accepts from Fannie and Freddie consists of US government and federal agency securities only, the expansion of the set of eligible discount window counterparties to include Fannie and Freddie does not represent a material quasi-fiscal abuse of the Fed. If at some future date the maturity of the loans extended to Fannie and Freddie at the discount window were to be longer than overnight, and if lower quality collateral were to be accepted and not priced appropriately, Fannie’s and Freddie’s access to the discount window could become a conduit for quasi-fiscal subsidies. This is not, I believe, an idle concern. The Fed’s opening of the discount window to the two GSEs was announced at the same time as some potentially very large-scale contingent quasi-fiscal commit-

08 Book.indb 596

2/13/09 3:59:19 PM

Central Banks and Financial Crises

597

ments by the Treasury to these organisations, including debt guarantees and the possibility of additional equity injections. There also is the worrying matter that, even though Fannie and Freddie now have access to the discount window, there is no special resolution regime for the two GSEs to constrain the incentives for excessive risk taking created by access to the discount window. III.2a(ix)

Lowering the discount window penalty

In Section III.1, I listed the lowering (in two steps) of the discount rate penalty from 100 to 25 basis points as a stabilisation policy measure, although it is unlikely to have had more than a negligible effect, except possibly as “mood music”: it represents the marginal cost of external finance only for a negligible set of financial institutions. The discount rate penalty reductions should, however, be included in the financial stability section as an essentially quasi-fiscal measure. On August 17, 2007, there were no US financial institutions for whom the difference between able to borrow at the discount rate at 5.75 percent rather than at 6.25 percent represented the difference between survival and insolvency; neither would it make a material difference to banks considering retrenchment in their lending activity to the real economy or to other financial institutions. This reduction in the discount window penalty margin was of interest only to institutions already willing and able to borrow at the discount window (because they had the kind of collateral normally expected there). It was an infra-marginal subsidy to such banks—a straight transfer to their shareholders from the US taxpayers. It also will have boosted moral hazard to a limited degree by lowering the penalty for future illiquidity. III.2a(x)

Interest on reserves

Reserves held by commercial banks with the Fed are currently non-remunerated. As I pointed out in Section II.5, this hampers the Fed in keeping the effective federal funds rate close to the federal funds target. Commercial banks have little incentive to hold excess reserves with the central bank. If there is excess liquidity in the overnight interbank market, banks will try to lend it out overnight at any

08 Book.indb 597

2/13/09 3:59:19 PM

598

Willem H. Buiter

positive rate rather than holding it at a zero overnight rate as excess reserves with the Fed. Clearly it makes sense for interest to be paid on excess reserves at an overnight rate equal to the federal funds target rate. Under existing legislation, the Fed will have the authority to pay interest on reserves starting in October 2011. The Fed has asked Congress for this date to be brought forward. The proposal clearly makes sense, but if interest at the federal funds target rate is paid on both required and excess reserves, the proposed policy change represents a quasi–fiscal tax cut benefiting the shareholders of the banks. In a first-best world, the Fed would not collect quasi-fiscal taxes through unremunerated reserves. However, to correct this problem now, as a one-off, would look like a further reward to the banks for past imprudent behaviour and would also be distributionally unfair. The Fed should insist that interest be paid only on excess reserves held by the commercial banks, with zero interest on required reserves. Once the dust has settled, the question of the appropriate way to tax the commercial banks and fund the Fed can be addressed at leisure. III.2a(xi)

Limiting the damage of the current crisis versus worsening the prospects for the next crisis

There can be little doubt that the Fed has done many things right as regards dealing with the immediate liquidity crisis. First, it used its existing facilities to accommodate the increased demand for liquidity. It extended the maturity of its discount window loans. It widened the range of collateral it would accept in repos and at the discount window. It created additional term facilities for existing counterparties through the TAF. It increased the range of eligible counterparties by creating the TSLF and the PDCF and it extended discount window access to Fannie and Freddie. It also stopped a run on investment banks by bailing out Bear Stearns. However, the way in which some of these “putting-out-firesmanoeuvres” were executed seems to have been designed to maximise bad incentives for future reckless lending and borrowing by the institutions affected by them. Between the TAF, the TSLF, the PDCF, the rescue of Bear Stearns and the opening of the discount window to the

08 Book.indb 598

2/13/09 3:59:19 PM

Central Banks and Financial Crises

599

two GSEs, the Fed and the US taxpayer have effectively underwritten directly all of the “household name” US banking system—commercial banks and investment banks—and probably also, indirectly, most of the other large highly leveraged institutions. This was done without the extraction of any significant quid pro quo and without proportional and appropriate pain for shareholders, directors, top managers and creditors of the institutions that benefited. The privilege of access to Fed resources was extended without a matching expansion of the regulatory constraints traditionally put on counterparties enjoying this access. Specifically, the new beneficiaries have not been made subject to a Special Resolution Regime analogous to that managed by the FDIC for federally insured commercial banks. The valuation of the collateral for the TSLF and the PDCF by the clearer acting for the borrowing primary dealer seems designed to maximise adverse selection. The discount rate penalty cuts were infra-marginal transfer payments from the taxpayers to the shareholders of banks already using or planning to use the discount window facilities. Asking for the decision to pay interest on bank reserves to be brought forward without insisting that required reserved remain non-remunerated likewise represents an unnecessary boon for the banking sector. III.2a(xii)

Cognitive regulatory capture of the Fed by vested interests

In each of the instances where the Fed maximised moral hazard and adverse selection, obviously superior alternatives were available—and not just with the benefit of hindsight. Why did the Fed not choose these alternatives? I believe a key reason is that the Fed listens to Wall Street and believes what it hears; at any rate, the Fed acts as if it believes what Wall Street tells it. Wall Street tells the Fed about its pain, what its pain means for the economy at large and what the Fed ought to do about it. Wall Street’s pain was great indeed—deservedly so in many cases. Wall Street engaged in special pleading by exaggerating the impact on the wider economy of the rapid deleveraging (contraction of the size of the balance sheets) that was taking place. Wall Street wanted large rate cuts

08 Book.indb 599

2/13/09 3:59:19 PM

600

Willem H. Buiter

fast to assist it in its solvency repairs, not just to improve its liquidity, and Wall Street wanted the provision of ample liquidity against overvalued collateral. Why did Wall Street get what it wanted? Throughout the 12 months of the crisis, it is difficult to avoid the impression that the Fed is too close to the financial markets and leading financial institutions, and too responsive to their special pleadings, to make the right decisions for the economy as a whole. Historically, the same behaviour has characterised the Greenspan Fed. It came as something of a surprise to me that the Bernanke Fed, if not quite a clone of the Greenspan Fed, displays the same excess sensitivity to Wall Street concerns. The main recent evidence of Fed excess sensitivity to Wall Street concerns are, in addition to the list of quasi-fiscal features of the liquidity-enhancing measures listed in Section III.2a(xi), the excessive cumulative magnitude of cuts in the official policy rate since August 2007 (325 basis points), and especially the 75 basis points cut on January 21/22, 2008. As regards the “panic cut”, the only “news” that could have prompted the decision on January 21, 2008, to implement a federal funds target rate cut of 75 bps, at an unscheduled meeting, and to announce that cut out of normal working hours the next day was the high-frequency movement in stock prices and the palpable fear in the financial sector that the stock market rout in Europe on Monday January 21, 2008 (a US stock market holiday), and at the end of the previous week, would spill over into the US markets.38 To me, both the cumulative magnitude of the official policy rate cuts and their timing provide support for what used to be called the “Greenspan put” hypothesis, but should now be called the “GreenspanBernanke put” or “Fed put” hypothesis.39 A complete definition of the “Greenspan-Bernanke put” is as follows: It is the aggressive response of the official policy rate to a sharp decline in asset prices (especially stock prices) and other manifestations of financial sector distress, even when the asset price falls and financial distress (a) are unlikely to cause future economic activity to weaken by more than required to meet the Fed’s mandate and (b) do not convey new information about future

08 Book.indb 600

2/13/09 3:59:19 PM

Central Banks and Financial Crises

601

economic activity or inflation that would warrant an interest rate cut of the magnitude actually implemented. Mr. Greenspan and many other “put deniers” are correct in drawing attention to the identification problems associated with establishing the occurrence of a “Greenspan-Bernanke put.” The mere fact that a cut in the policy rate supports the stock market does not mean that the value of the stock market is of any inherent concern to the policy maker. This is because of the causal and predictive roles of asset price changes. Falling stock market prices reduce wealth and weaken corporate investment; falling house prices reduce the collateral value of residential property and weaken housing investment. Forward-looking stock prices can anticipate future fundamental developments and thus be a source of news. Nevertheless, looking at the available data as a historian, and constructing plausible counterfactuals as a “laboratory economist,” it seems pretty evident to me that the Fed under both Greenspan and Bernanke has cut rates more vigorously in response to sharp falls in stock prices than can be rationalised with the causal effects of stock prices on household spending and on private investment, or with the predictive content of unexpected changes in stock prices. Both the 1998 LTCM and the January 21/22, 2008, episodes suggest that the Fed has been co-opted by Wall Street—that the Fed has effectively internalised the objectives, concerns, world view and fears of the financial community. This socialisation into a partial and often distorted perception of reality is unhealthy and dangerous. It can be called cognitive regulatory capture (or cognitive state capture), because it is not achieved by special interests buying, blackmailing or bribing their way towards control of the legislature, the executive, the legislature or some important regulator or agency, like the Fed, but instead through those in charge of the relevant state entity internalising, as if by osmosis, the objectives, interests and perception of reality of the vested interest they are meant to regulate and supervise in the public interest.

08 Book.indb 601

2/13/09 3:59:19 PM

602

Willem H. Buiter

The literature on regulatory capture, and its big brother, state capture, is vast (see e.g. Stigler, 1971; Levine and Forrence, 1990; Laffont and Tirole, 1991; Hellman, et al., 2000; and Hanson and Yosifon, 2003). Capture occurs when bureaucrats, regulators, judges or politicians instead of serving the public interest as they are mandated to do, end up acting systematically to favour specific vested interests—often the very interests they were supposed to control or restrain in the public interest. The phenomenon is theoretically plausible and empirically well–documented. Its application to the Fed is also not new. There is a long-standing debate as to whether the behaviour of the Fed during the 1930s can be explained as the result of regulatory capture (see e.g. Epstein and Ferguson, 1984, and Philip, et al., 1991). The conventional choice-theoretic public choice approach to regulatory capture stresses the importance of collective action and free rider considerations in explaining regulatory capture (see Olsen, 1965). Vested interests have a concentrated financial stake in the outcomes of the decisions of the regulator. The general public individually have less at stake and are harder to organise. I prefer a more social-psychological, small group behaviour-based explanation of the phenomenon. Whatever the mechanism, few regulators have succeeded in escaping in a lasting manner their capture by the regulated industry. I consider the hypothesis that there has been regulatory capture of the Fed by Wall Street during the Greenspan years, and that this is continuing into the present, to be consistent with the observed facts. There is little room for doubt, in my view, that the Fed under Greenspan treated the stability, well-being and profitability of the financial sector as an objective in its own right, regardless of whether this contributed to the Fed’s legal macroeconomic mandate of maximum employment and stable prices or to its financial stability mandate. Although the Bernanke Fed has but a short track record, its too often rather panicky and exaggerated reactions and actions since August 2007 suggest that it also may have a distorted and exaggerated view of the importance of financial sector comfort for macroeconomic stability.

08 Book.indb 602

2/13/09 3:59:19 PM

Central Banks and Financial Crises

603

III.2b The ECB The ECB immediately injected liquidity both overnight and at longer maturities on a very large scale indeed, but, at least as regards interbank spreads, with limited success (see Chart 4), and also with no greater degree of success than the Fed or the BoE (but see Section II3b for a caution about the interpretation of the similarity in Libor-OIS spreads). The ECB’s injection of €95 billion into the Eurosystem’s money markets on August 9, 2007, is viewed by many as marking the start of the crisis.40 As regards the effectiveness of its liquidity-enhancing open market interventions on the immediate crisis (as opposed to the likelihood and severity of future crises) the ECB has been both lucky and smart. It was lucky because, as part of the compromise that created the supranational European Central Bank, the set of eligible collateral for open market operations and at the discount window and the set of eligible counterparties, were defined as the union rather than the intersection of the previous national sets of eligible collateral and eligible counterparties.41 As a result, the ECB could accept as collateral in its repos and at the discount window a very large set of securities, including private securities (even equity) and asset-backed securities like residential mortgage-backed securities. The ratings requirements were also very loose compared to those of the BoE and even those of the Fed: Eligible securities had to be rated at least in the single A category by one or more of the recognised rating agencies. The only dimension in which the ECB’s eligible collateral was more restricted than the BoE’s was that the ECB only accepts euro-denominated securities. Currently around 1700 banks are eligible counterparties of the Eurosystem for open market operations. The Fed has 20 (the primary dealers) and the BoE 40 (reserve scheme participants); around 2100 banks have access to the ECB’s discount window, as against 7500 for the Fed and 60 for the BoE. The ECB was smart in using the available liquidity instruments quite aggressively, injecting above-normal amounts of liquidity against a wide range of collateral at longer maturities (and mopping most of it up again in

08 Book.indb 603

2/13/09 3:59:19 PM

604

Willem H. Buiter

the overnight market). It is important to note that injecting X amount of liquidity at the 3-month maturity and taking X amount of liquidity out at the overnight maturity is not neutral if the intensity of the liquidity crunch is not uniform across maturities. The liquidity crunch that started in August 2007 clearly was not. Maturities of around one month were crucial for end-of-year reasons and maturities from three months to a year were crucial because that was where the markets had seized up completely. The ECB consciously tried to influence Euribor-OIS spreads to the extent that it interpreted these as reflecting illiquidity and liquidity risk rather than credit risk. No major Euro Area bank has failed so far. Some small German banks fell victim to unwise investments in the ABS markets, and some fairly small hedge funds failed, but no institution of systemic importance was jolted to the point that a special-purpose LLR rescue mission had to be organised. I have one concern about the nature of the ECB’s liquidity– oriented open market operations and about its collateral policy at the discount window. This concerns the pricing of illiquid collateral offered by banks. We know the interest rates and fees charged for these operations, and the haircuts applied to the valuations. But we don’t know the valuations themselves. The ECB uses market prices when a functioning market exists. For some of the assets it accepts as collateral there is no market benchmark. The ECB does not make the mistake the Fed makes in its pricing of the collateral offered at the PDCF and TSLF. The ECB itself determines the price/valuation of the collateral when there is no market price. But the ECB does not tell us what these prices are, nor does it put in the public domain the models or methodologies it uses to price the illiquid securities. Requests to ECB Governing Council members and to ECB and NCB officials to publish the models used to price illiquid securities and to publish, with an appropriate delay to deal with commercial sensitivity, the actual valuations of specific, individual items of collateral have fallen on deaf ears.

08 Book.indb 604

2/13/09 3:59:19 PM

Central Banks and Financial Crises

605

There is therefore a risk that banks use the ECB as lender of first resort rather than last resort, if the banks can dump low-grade collateral on the Eurosystem and have it valued as high-grade collateral.42 Since at least the beginning of 2008, persistent market talk has it that Spanish, Irish and Dutch banks may be in that game, getting an effective subsidy from the Eurosystem and becoming overly dependent on the Eurosystem as the funding source of first choice. Late May 2008, Fitch Ratings reported that Spanish banks had, during recent months, created ABS, structured to be eligible for use as collateral with the ECB (strictly, with the NCBs that make up the Eurosystem), that were riskier than the ABS structures they put together before the crisis. Accepting higher-credit–risk collateral need not imply a subsidy from the Eurosystem to the banks, as long as the valuation or pricing of these securities for collateral purposes reflects the higher degree of credit risk attached to them. One wonders whether such risk-sensitive pricing is actually taking place, especially when ECB officials publicly worry about the creditworthiness of securities accepted as collateral by the ECB when it provides liquidity to the markets or at the discount window. Although RMBS backed by mortgages originated by the borrowing bank itself are not eligible as collateral with the Eurosystem, RMBS issued by parties with whom the borrowing back has quite a close relationship (through currency hedges with the issuer or guarantor of the RMBS or by providing liquidity support for the RMBS). In principle, the higher credit risk attached to securities for which the borrower and the issuer/guarantor are close (compared the credit risk attached to similar securities issued or guaranteed by a bank that is independent of the borrowing bank) could be priced so as to reflect their higher credit risk. We have no hard information on whether such credit-risk-sensitive pricing actually takes place. I fear that if it were, we would have been told, and that the lack of information is supportive of the view that implicit subsidisation is taking place. As long as the risk-adjusted rate of return the ECB gets on its loans is appropriate, there is nothing inherently wrong with the ECB taking credit risk onto its balance sheet. But if it routinely values the

08 Book.indb 605

2/13/09 3:59:19 PM

606

Willem H. Buiter

mortgage-backed securities offered by the Spanish banks as if the mortgages backing the securities were virtually free of default risk, then the ECB is bound to be overvaluing the collateral it is offered. In the first half of 2008, Spanish commercial banks, heavily exposed to the Spanish construction and real estate sectors, are reported to have repoed at least € 46 billion worth of their assets in exchange for ECB liquidity. Participants in these repo transactions have told me that no mortgages offered to the Eurosystem as collateral have been priced at less than 95 cents on the euro. This seems generous given the dire straits the Spanish economy, and especially the construction and real estate sectors, now are in. Of course, haircuts are (as always) applied to these valuations.43 It is essential that all the information required to verify whether the pricing of collateral accepted by the Eurosystem is subsidy-free be in the public domain. That information is not available today. Because part of the collateral offered the Eurosystem is subject to default risk, there could be a case for concern even if, ex ante, the default risk is appropriately priced. In the event a default occurs (that is, if both the counterparty borrowing from the Eurosystem defaults and at the same time the issuer of the collateral defaults), the Eurosystem will suffer a capital loss. In practice, it would be one of the NCBs of the euro area that would suffer the loss rather than the ECB, as repos are conducted by the NCBs. Although the ECB’s balance sheet is small and its capital tiny, the consolidated Eurosystem has a huge balance sheet and a large amount of capital (see Table 6). The balance sheet could probably stand a fair-sized capital loss. But there always is a capital loss so large that it would threaten the ability of the Eurosystem to remain solvent while adhering to its price stability mandate. The ECB/Eurosystem would need to be recapitalised, but by which national fiscal authorities and in which proportions? Unlike the Fed and the BoE, where it is clear which fiscal authority stands behind the central bank, that is, stands ready to recapitalise the central bank should the need arise, the fiscal vacuum within which the ECB, and to some degree the rest of the Eurosystem also, operate leaves a question mark behind the question: Who would bail out the ECB?

08 Book.indb 606

2/13/09 3:59:20 PM

Central Banks and Financial Crises

607

This question may not yet be urgent now, because even euro area banks with large cross-border activities still tend to have fairly clear national identities. But this is changing. Banca Antonveneta, the fourth largest Italian bank, was owned by ABN-AMRO, a Dutch bank which is now in turn owned by Royal Bank of Scotland (UK), Fortis (Belgium) and Santander (Spain).44 Would the Italian Treasury bail out Banca Antonveneta? Soon there will be banks incorporated not under national banking statutes but under European law, as Societas Europaea. One large German financial group with banking interests, Allianz, has already done so. Given this uncertainty, it may be understandable that ECB officials are more concerned than Fed and BoE officials about carrying credit risk on the Eurosystem’s balance sheet. Although the ECB has done well in its MMLR function, albeit with the major caveat as regards the pricing of illiquid collateral, its LLR ability has not yet been tested. This is perhaps just as well. The ECB has no formal supervisory or regulatory role vis-à-vis euro area banks. The Treaty neither rules out such a role nor does it require one. In practice, no regulatory and supervisory role for the ECB has as yet evolved. Banking sector regulation and supervision in the euro area is a mess. In some countries the central bank is regulator and supervisor. Spain, France, Ireland and the Netherlands are examples. In others the central bank shares these roles with another agency. Germany is an example with the Bundesbank and BaFin (the German Financial Supervisory Authority) sharing supervisory responsibilities.45 In yet other countries the central bank has no regulatory and supervisory role at all. Austria and Belgium are examples. Since the crisis started, the ECB has complained regularly, and at times even publicly, about the lack of information it has at its disposal about potentially systemically important individual institutions. In the case of some euro area national regulators, there even exist legal obstacles to sharing information with the ECB. Compared to the Fed and the BoE, the ECB is therefore very close to the BoE which, when the crisis started, had essentially no individual institution-specific information at its disposal. The Fed, with its (shared) regulatory and supervisory role, has better information.

08 Book.indb 607

2/13/09 3:59:20 PM

608

Willem H. Buiter

On the other hand, the ECB appears much less moved by the special pleading emanating from the euro area financial sector than the Fed appears to be by Wall Street. This is not surprising. Without a supervisory or regulatory role over euro area financial institutions and markets, regulatory capture is less likely. III.2c The BoE As regards the fulfilment of its LLR and MMLR functions, the BoE missed the boat completely at the beginning of the crisis. This state of affairs lasted till about November 2007. Indeed, the Governor of the BoE did not, as far as I have been able to ascertain, use in public the words “credit crunch,” “liquidity crisis” or equivalent words until March 26, 2008 (King, 2008). The UK turned out, when the run on Northern Rock started on September 15, 2007, to have no effective deposit insurance scheme. The amounts insured were rather low (up to £30,000) and had a 10 percent deductible after the first £2,000. Worse, it could take up to six months to get your money out, even if it was insured. This is supposed to be corrected by new legislation and institutional reform. The BoE also turned out to be hopelessly (and quite unnecessarily) confused about what its legal powers and constraints were in the exercise of its LLR role. The Governor, for instance, argued on September 20, 2007, before the House of Commons Treasury Committee, that legislation introduced under an EU directive (the Market Abuse Directive) prevented covert support to individual institutions (the BoE had received legal advice to this effect). Since then what always was apparent to most has become apparent to all: Neither the MAD nor the UK’s transposition of that Directive into domestic law prevented the kind of covert support the BoE would have liked to offer to Northern Rock. Finally, there was no Special Resolution Regime for banks in the UK. There was therefore just the choice between the regular corporate insolvency regime and nationalisation. On February 18, 2008, the Chancellor announced the nationalisation of Northern Rock.

08 Book.indb 608

2/13/09 3:59:20 PM

Central Banks and Financial Crises

609

The BoE’s performance as lender of last resort, including its covert role in orchestrating private sector support for individual troubled institutions, was much more effective when Bradford & Bingley (a British mortgage lender whose exposure to the wholesale markets was second only to that of Northern Rock) got into heavy weather with a rights issue in May and June 2008.46 Neither Northern Rock nor Bradford & Bingley were in any sense systemically important institutions, but when HBOS, the fourth largest UK banking group by market capitalisation experienced trouble with its £4 billion rights issue (announced in April 2008), during June and July 2008, systemic stability was clearly at stake. The BoE and the banking and financial sector regulator, the Financial Services Authority (FSA), helped keep the underwriters on board. As noted earlier, both at its discount window (the standing lending facility) and in repos, the BoE only accepted (and accepts) the narrowest possible kind of collateral (UK sovereign debt or better). This made it impossible for the BoE to offer effective liquidity support when markets froze. For a long time, the BoE spoke in public as if it believed that what the banks were facing was essentially a solvency problem, with no material contribution to the financial distress coming from illiquid markets and from illiquid but solvent institutions (see e.g. the paper submitted to the Treasury Committee by Mervyn King on September 12, 2007, the day before the Northern Rock crisis blew up [King, 2007]). When the crisis started, the BoE injected liquidity on a modest scale, at first only in the overnight interbank market. Rather late in the day, on September 19, 2007, it reversed this policy and offered to repo at three-month maturity, and against a wider than usual range of eligible collateral, including prime mortgages, but subject to an interest rate floor 100 basis points above Bank Rate, that is, effectively at a penalty rate, regardless of the quality of the collateral. No one came forward to take advantage of this facility; fear of being stigmatised may have been as important a deterrent as the penalty rate charged.

08 Book.indb 609

2/13/09 3:59:20 PM

610

Willem H. Buiter

The Bank was extremely reluctant to try to influence, let alone target, interest rates at maturities longer than the overnight rate. It is true that, when markets are orderly and liquid, the authorities cannot independently set more than one rate on the yield curve. When the BoE sets the overnight rate, this leaves rates at all longer maturities to be market-determined, that is, driven by fundamentals such as market expectations of future official policy rates and default risk premia. When markets are disorderly and illiquid, however, there is a term structure of liquidity risk premia in addition to a term structure of default-risk-free interest rates and a term structure of default risk premia. It is the responsibility of the central bank, as MMLR, to provide the public good of liquidity in the amounts required to eliminate (most of ) the liquidity risk premia at the maturities that matter (anything between overnight and one year). Early in the crisis, the BoE’s public statements suggested that it interpreted most the spread between Libor and the OIS rate at various maturities as default risk spreads rather than, at least in part, as liquidity risk spreads. Later during the crisis, in February 2008, the BoE published, in the February Inflation Report (Bank of England, 2008), a decomposition of the one-year Libor-OIS spread between a default risk measure (extracted from CDS spreads) and a liquidity premium (the residual). It concluded that although early in the crisis most of the Libor-OIS spread was due to liquidity premia, towards the end of the sample period the importance of default risk premia had increased significantly. The decomposition is, unfortunately, flawed because the CDS market throughout the crisis has itself been affected significantly by illiquidity. The paper is, however, of interest as evidence of the evolving and changing views of the BoE as to the empirical relevance of liquidity crises. This changing view was also reflected in an evolving policy response. The BoE gradually began to act as a MMLR. At the end of 2007, the BoE initiated a number of special auctions at one-month and three-month maturities against a wider range of collateral, including prime mortgages and securities backed by mortgages.

08 Book.indb 610

2/13/09 3:59:20 PM

Central Banks and Financial Crises

611

On April 21, 2008, the BoE announced the creation of the Special Liquidity Scheme (SLS), in the first instance for £100 billion, which would lend Treasury bills for one year to banks against collateral that included RMBS, covered bonds (that is, collateralised bonds) and ABS based on credit card receivables. Technically, the arrangement was described as a swap, although it can fairly be described as a oneyear collateralised loan of Treasury bills to the banks. It is similar to the TSLF created for primary dealers in the US, although the maturity of the loans is longer (one year as against one month in the US). The BoE has made much of the fact that the SLS will only accept as collateral securities backed by “old” mortgages, that is, mortgages issued before the end of 2007. The facility is meant to solve the “stock overhang” problem but not to encourage the banks to engage in new mortgage lending using the same kind of RMBS that have become illiquid. It is, however, not obvious that without the government (not necessarily the BoE) lending a hand, securitisation of new mortgages will get off the ground any time soon. Accepting new mortgage-backed securities as collateral in repos might help revive sensible forms of securitisation, if the mortgages backing the securities satisfy certain verifiable criteria (loan to value limits, income and financial health verification for borrowers, no track record of loan default, etc.). It is true that in the UK, and a fortiori in the US, there was, prior to the summer of 2007, securitisation of home loans that ought never to have been made, including many of the US subprime loans. But the fact that, during the year since August 2007, there have been just two new residential mortgage-backed issues in the markets in the UK, suggests that the securitisation baby has been thrown out with the subprime bathwater. These securities should, of course, be valued aggressively if offered as collateral in repos, to avoid subsidies to home lenders or home borrowers. The BoE itself determines the valuation of any illiquid assets offered as collateral in the SLF. This should help it avoid the adverse selection problem created by the Fed with its PDCF and TSLF. The haircuts and other terms of the SLS were also quite punitive, judging from the howls of anguish emanating from the banking community, who nevertheless make ample use of the Facility. As with the Fed and the ECB,

08 Book.indb 611

2/13/09 3:59:20 PM

612

Willem H. Buiter

the BoE does not make public any information about the actual pricing of specific collateral or about the models used to set these prices. Without that information, we cannot be sure there is no subsidy to the banks involved in the arrangement. There can also be no proper accountability of the BoE to Parliament or to the public for the management of public funds involved. It is clear that the so-called Tripartite Arrangement between the Treasury, the BoE and the FSA did not work. It is also clear, however, that these are the three parties that must be involved and must cooperate to achieve financial stability. The central bank has the short-term liquid deep pockets and the market knowledge. The Treasury, backed by the taxpayer, has the long-term deep non-inflationary pockets. The FSA has the individual institution-specific knowledge. The problems in the UK had more to do with failures in the legal framework (deposit insurance, lender of last resort immunities, the insolvency regime and SRR for banks) than with poor communication and cooperation between the central bank, the regulator and the Treasury. IV.

Conclusion

Following a 15–year vacation in inflation targeting land with hardly a hint of systemic financial instability, the central banks in the North Atlantic region were, in the middle of 2007, faced with the unpleasing combination of a systemic financial crisis, rising inflation and weakening economic activity. Fighting three wars at the same time was not something the central banking community was prepared for. The performance of the central banks considered in this paper, the Fed, the ECB and the BoE, ranged, not surprisingly, from not too bad (the ECB) to not very good at all (the Fed). As regards macroeconomic stability, the interest rate decisions of the Fed are hard to rationalise in terms of its official mandate (sustainable growth/employment and price stability). This is not the case for the ECB and the BoE, with their lexicographic price stability mandates. The excessively aggressive interest rate cuts of the Fed reflect political pressures (the Fed is the least operationally independent of the three central bank), excess sensitivity to financial sector concerns (reflecting

08 Book.indb 612

2/13/09 3:59:20 PM

Central Banks and Financial Crises

613

cognitive regulatory capture) and flaws in the understanding of the transmission mechanism by key members of the FOMC. As regards financial stability, an ideal central bank would have combined the concern about moral hazard of the BoE with the broad sets of eligible counterparties and eligible instruments that enabled the ECB, right from the start of the crisis, to be an effective market maker of last resort, and the institution-specific knowledge that made the Fed an effective lender of last resort. The reality has been that the BoE mismanaged liquidity provision as market maker of last resort and as lender of last resort early in the crisis, and that the Fed has created moral hazard in an unprecedented way. Until the public is informed in detail about the way the three central banks price the illiquid collateral they are offered (at the discount window, in repos, and at any of the many facilities and schemes that have been created), there has to be a concern that all three central banks (and therefore indirectly the taxpayers and beneficiaries of other public spending) may be subsidising the banks through these LLR and MMLR facilities. This concern is most acute as regards the Fed, whose valuation procedures at the TSLF and PDCF are an open invitation to adverse selection. As regards the desirability of institutionally combining or separating the roles of the central bank (as lender of last resort and market maker of last resort) and that of regulator and supervisor for the financial sector, we are between a rock and a hard place. A regulator and supervisor (like the Fed) is more likely to have the institution-specific information necessary for the effective performance of the LLR role. However, regulatory capture of the regulator/supervisor is likely. Central banks without regulatory or supervisory responsibilities like the BoE (for the time being) and the ECB are less likely to be captured by vested financial sector interests. But they are also less likely to be well-informed about possible liquidity or solvency problems in systemically important financial institutions. There is unlikely to be a fully satisfactory solution to the problem of providing central banks with the information necessary for effective discharge of their LLR responsibility without at the same time exposing them

08 Book.indb 613

2/13/09 3:59:20 PM

614

Willem H. Buiter

to the risk of regulatory capture. The best safeguards against capture are openness and accountability. It is therefore most disturbing that all three central banks are pathologically secretive about the terms on which financial support is made available to struggling institutions and counterparties. Taking the official policy rate-setting decision away the central bank may reduce the damage caused by regulatory capture of the central bank by financial sector interests. Moving the rate setting authority out of the central bank could therefore be especially desirable if the central bank is given supervisory and regulatory powers. The market maker of last resort has the same position in relation to market liquidity for a transactions-oriented system of financial intermediation, as is held by the lender of last resort in relation to funding liquidity for a relationships-oriented system of financial intermediation. The central bank is the natural entity to fulfill both the LLR and MMLR functions. There is an efficiency-based case for government intervention to support illiquid markets or instruments and to support illiquid but solvent financial institutions that are deemed systemically important. As the source of ultimate domestic-currency liquidity, the central bank is the natural agency for performing both the market maker of last resort and the lender of last resort function. Liquidity is a public good that can be provided privately, but only inefficiently. There is also an efficiency-based case for government intervention to support insolvent financial institutions that are deemed systemically important. This, however, should not be the responsibility of the central bank. The central bank should not be required to provide subsidies, either through liquidity support or any other way, to institutions known to be insolvent. If institutions deemed to be solvent turn out to be insolvent, and if the central bank as a result of financial exposure to such institutions suffers a loss, this should be compensated forthwith by the Treasury, whenever such a loss would impair the ability of the central bank to pursue its macroeconomic stability objectives.

08 Book.indb 614

2/13/09 3:59:20 PM

Central Banks and Financial Crises

615

It would be even better if any securities purchased outright by the central bank or accepted as collateral in repos and other secured transactions that are not completely free of default risk, were to be transferred immediately to the balance sheet of the Treasury (say through a swap for Treasury Bills, at the valuation put on these risky securities when they were acquired by the central bank). That way, the division of labour and responsibilities between liquidity management and insolvency management (or bailouts) is clear. Each institution can be held accountable to Parliament/Congress for its mandate. If the central bank plays a quasi-fiscal role, that clarity, transparency and accountability becomes impaired. Central banks can effectively perform their market maker of last resort function by expanding traditional open market operations and repos. This means increasing the volumes of their outright purchases or loans and extending their maturity, at least up to a year in the case of repos. It means extending the range of eligible counterparties to include all institutions deemed systemically important (too large or too interconnected to fail). It also means extending the range of securities eligible for outright purchase or for use as collateral to include illiquid private securities. Regulatory instruments should be used to address financial asset market bubbles and credit booms. Specifically, supplementary capital requirements and liquidity requirements should be imposed on all systemically important highly leveraged institutions—commercial banks, investment banks, hedge funds, private equity funds or whatever else they are called or will be called. These supplementary capital and liquidity requirements could either be managed by the central bank in counter-cyclical fashion or be structured as automatic financial stabilisers, say by making them increasing functions of the recent historical growth rates of the value of each firm’s assets. To minimise moral hazard (incentives for excessive risk-taking in the future) all institutions that are eligible counterparties in MMLR operations and/or users of LLR facilities should be regulated according to common principles and should be subject to a common Special Resolution Regime allowing for Prompt Corrective Action,

08 Book.indb 615

2/13/09 3:59:20 PM

616

Willem H. Buiter

including the condition of regulatory insolvency and the possibility of nationalisation. All securities purchased outright or accepted as collateral should be priced punitively to minimize moral hazard. If necessary, the central bank should organise reverse auctions to price securities for which there is no market benchmark. The creation and proliferation of obscure and opaque financial instruments can be discouraged through the creation of a positive list (regularly updated) of securities that will be accepted by the central bank as collateral at its MMLR and LLR facilities. Securities that don’t appear on the list can be expected to trade at a discount relative to those that do. Finally, for those whose attention span is the reciprocal of the length of this paper, some do’s and don’ts for central banks. Assign specific tools to specific tasks or objectives. 1. Assign the official policy rate to the macroeconomic stability objective(s). • Do not use the official policy rate as a liquidity management tool or as a quasi-fiscal tool to recapitalise banks and other highly leveraged entities. 2. Assign regulatory instruments to the damping of asset price bubbles. • Do not use the official policy rate to target asset price bubbles in their own right. 3. Assign liquidity management tools, including the lender-of-lastresort and market-maker-of-last-resort instruments, to the pursuit of financial stability for counterparties believed to be solvent. 4. Use explicit fiscal tools (taxes and subsidies) and on-budget and on-balance-sheet fiscal resources for strengthening the capital adequacy of systemically important institutions. • Do not use the central bank as a quasi-fiscal agent of the Treasury.

08 Book.indb 616

2/13/09 3:59:20 PM

Central Banks and Financial Crises

617

5. Use regulatory instruments and the punitive pricing of liquidity to mitigate moral hazard. This past year has been the first since I left the Monetary Policy Committee of the BoE that I really would have liked to be a central banker.

Author’s note: I would like to thank my discussants, Alan Blinder and Yutaka Yamaguchi, for helpful comments, and the audience for its lively participation. Alan’s contribution demonstrated that you can be both extremely funny and quite wrong on the substance of the issues. This paper builds on material written, in a variety of formats, since April 2008. I would like to thank Anne Sibert, Martin Wolf, Charles Goodhart, Danny Quah, Jim O’Neill, Erik Nielsen, Ben Broadbent, Charles Calomiris, Stephen Cecchetti, Marcus Miller and John Williamson for many discussions of the ideas contained in this paper. They are not responsible for the end product. The views expressed are my own. They do not represent the views of any organisation I am associated with.

08 Book.indb 617

2/13/09 3:59:20 PM

618

Willem H. Buiter

Endnotes The official inflation targets are 2.0 percent per annum for the BoE and just below 2.0 percent for the ECB, both for the CPI. I assume the Fed’s unofficial centre for its PCE deflator inflation comfort zone to be 1.5 percent. Given the recent historical wedge between US PCE and CPI inflation, this translates into an informal Fed CPI inflation target of just below 2.0 percent. 1

The long-term inflation expectations data for the euro area should be taken with a pinch of salt. The reported euro area survey-based inflation expectations are the predictions of professional forecasters rather than those of a wider crosssection of the public, as is the case for the US and UK data (see European Central Bank, 2008). The euro area professional forecasters are either very trusting/gullible or know much more than the rest of us, as their 5–years–ahead forecast flat-lines at the official target throughout the sample, despite a systematic overshooting of the target in the sample. Using market-based estimates of inflation expectations, either break-even inflation rates from nominal and index-linked public debt or inflation expectations extracted from inflation swaps, would not be informative during periods of illiquid and disorderly financial markets. Even if the markets for these instruments themselves remain liquid, the yields on these instruments will be distorted by illiquidity elsewhere in the system. 2

3 The Federal Reserve Act, Section 2a. Monetary Policy Objectives, states: “The Board of Governors of the Federal Reserve System and the Federal Open Market Committee shall maintain long run growth of the monetary and credit aggregates commensurate with the economy’s long run potential to increase production, so as to promote effectively the goals of maximum employment, stable prices, and moderate long-term interest rates.” [12 USC 225a. As added by act of November 16, 1977 (91 Stat. 1387) and amended by acts of October 27, 1978 (92 Stat. 1897); Aug. 23, 1988 (102 Stat. 1375); and Dec. 27, 2000 (114 Stat. 3028)].

I can therefore avoid addressing the anomaly (putting it politely) of the exchange rate, foreign exchange reserves and foreign exchange market intervention being under Treasury authority in the US (with the Fed acting as agent for the Treasury), or of the Council of Ministers of the EU (or perhaps of the euro area?) being able to give “exchange rate orientations” to the ECB. Clearly, in a world with unrestricted international mobility of financial capital, setting the exchange rate now and in the future effectively determines the domestic short risk-free nominal interest rate as a function of the foreign short risk-free nominal interest rate (there will be an exchange rate risk premium or discount unless the path of current and future exchange rates is deterministic). If the US Treasury were really determined to manage the exchange rate, the Fed would only have an interest rate-setting role left to the extent that the US economy is large enough to influence the world short risk-free nominal interest rate. 4

08 Book.indb 618

2/13/09 3:59:20 PM

Central Banks and Financial Crises

619

The non-negativity constraint on the nominal yield of non-monetary securities is the result of (a) the arbitrage requirement that the yield on non-monetary instruments,i, cannot be less than the yield on monetary securities,iM, that is, i > iM and (b) the practical problems of paying any interest at all on currency, that is, iM ≈ 0. This is because currency is a negotiable bearer bond. Paying interest, positive or negative, on negotiable bearer securities, while not impossible, is administratively awkward and costly. This problem does not occur in connection with the payment of interest, positive or negative, on the other component of the monetary base, bank reserves held with the central bank. Reserves held with the central bank are “registered” financial instruments. The issuer knows the identity of the holder. Paying interest, at a positive or negative rate, on reserves held with the central bank is trivially simple and administratively costless. Charging a negative nominal interest rate on borrowing from the central bank (secured or unsecured, at the discount window or through open market operations) is also no more complicated than paying a positive nominal interest rate. If the practical reality that paying (negative) interest on currency is not feasible or too costly sets a zero floor under the official policy rate, this would, in my view be a good argument for doing away with currency altogether (see Buiter and Panigirtzoglou, 2003). 5

Various forms of E-money provide near-perfect substitutes for currency, even for low income households. The existence of currency is, because of the anonymity it provides, a boon mainly to the grey and black economy and to the outright criminal fraternity, including those engaged in tax evasion, money laundering and terrorist financing. The Fed has reduced its subsidisation of such illegality and criminality by restricting its largest denomination currency note to $100. The ECB practices no such restraint and competes aggressively for the criminal currency market with €200 and €500 denomination notes. When challenged on this, the ECB informs one that this is because in Spain people like to make housing transactions in cash. I am sure they do. With the collapse of the Spanish housing market, this argument for issuing euro notes in denominations larger than €20 at most, may now have lost whatever merit it had before. The complete list includes gilts (including gilt strips), sterling Treasury bills, BoE securities, HM Government non-sterling marketable debt, sterling-denominated securities issued by European Economic Area central governments and major international institutions, euro-denominated securities (including strips) issued by EEA central governments and central banks and major international institutions where they are eligible for use in Eurosystem credit operations, all domestic currency bonds issued by other sovereigns eligible for sale to the Bank. These sovereign and supranational securities are subject to the requirement that they are issued by an issuer rated Aa3 (on Moody’s scale) or higher by two or more of the ratings agencies (Moody’s, Standard and Poor’s, and Fitch). 6

The three-month OIS rate is the fixed leg of a three-month swap whose variable leg is the overnight secured lending rate. This can be interpreted (ignoring inflation risk premia) as the market’s expectation of the official policy rate over a three-month horizon. 7

08 Book.indb 619

2/13/09 3:59:20 PM

620

Willem H. Buiter

Most of the RMBS issue by Alliance & Leicester was bought by a single Continental European bank. It is therefore akin to a private sale rather than a sale to market sale to non-bank investors. 8

9 Macroeconomic theory, unfortunately, has as yet very little to contribute to the key policy issue of liquidity management. The popularity of complete contingents markets models in much of contemporary macroeconomics, both New Classical (e.g. Lucas, 1975), Lucas and Stokey (1989) and New Keynesian (e.g. Woodford, 2003) means that in many (most?) of the most popular analytical and calibrated (I won’t call them empirical) macroeconomic dynamic stochastic general equilibrium models, the concept of liquidity makes no sense. Everything is perfectly liquid. Indeed, with complete contingent markets there is never any default in equilibrium, because every agent always satisfies his intertemporal budget constraint. All contracts are costlessly and instantaneously enforced. Ad hoc cash-in-advance constraints on household purchases of commodities or on household purchases of commodities and securities don’t create behaviour/outcomes that could be identified with liquidity constraints.

The legal constraint that labour is free (slavery and indentured labour are illegal) means that future labour income makes for very poor collateral, and that workers cannot credibly commit themselves not to leave an employer, should a more attractive employment opportunity come along. This can perhaps be characterised as a form of illiquidity, but it is a permanent, exogenous illiquidity, almost technological in nature. Much of the theoretical (partial equilibrium) work on illiquidity likewise deals with the consequences of different forms of exogenous illiquidity rather than with the endogenous illiquidity problem that suddenly paralysed many asset-backed securities markets starting in the summer of 2007. The profession entered the crisis equipped with a set of models that did not even permit questions about market liquidity to be asked, let alone answered. Much of macroeconomic theorising of the past thirty years now looks like a self-indulgent working and re-working to death of an uninteresting and practically unimportant special case. Instead of starting from the premise that markets are complete unless there are strong reasons for assuming otherwise, it would have been better to start from the position that markets don’t exist unless very special institutional and informational conditions are satisfied. We would have a different, and quite possibly more relevant, economics if we had started from markets as the exception rather than the rule, and had paid equal attention to alternative formal and informal mechanisms for organising and coordinating economic activity. My personal view is that over the past 30 years, we have had rather too much Merton (1990) and rather too little Minsky (1982) in our thinking about the roles of money and finance in the business cycle. The label “market maker of last resort” is more appropriate than the alternative “buyer of last resort,” because so much of the MMLR’s activity turns out to be in collateralised transactions, especially repos, rather than in outright purchases. A 10

08 Book.indb 620

2/13/09 3:59:20 PM

Central Banks and Financial Crises

621

repo is, of course a sale and repurchase transaction, so the label “buyer of last resort” would not have been descriptively correct. 11 The Switzerland-domiciled part of the Swiss banking system (as distinct from the foreign subsidiaries which may have access to LLR and MMLR facilities in their host countries) probably owes its competitive advantage less to conventional banking prowess as to the bank secrecy it provides to the global community of tax evaders and others interested in hiding their income and assets from their domestic authorities.

For a conflicting and very positive appraisal of the Greenspan years see Blinder and Reis (2005). 12

In the case of the Fed, the legal restrictions on paying interest on reserves (about to be abolished) are a further obstacle to sensible practice. 13

For descriptive realism, I assume iM=0. 1 15 Note that EtEt-1It,t-1=Et-1It,t-1= . 1+ it 16 Central bank current expenses C b can at most be cut to zero. 14

Source: IMF http://www.imf.org/external/np/sta/ir/8802.pdf.

17

A footnote in the Federal Reserve Bulletin (2008) informs us that “This balancing item is not intended as a measure of equity capital for use in capital adequacy analysis. On a seasonally adjusted basis, this item reflects any differences in the seasonal patterns estimated for total assets and total liabilities.” That is correct as regards the use of this measure in regulatory capital adequacy analysis. For economic analysis purposes it is, however, as close to W as we can get without a lot of detailed further work. 18

The example of the failure of the Amaranth Advisors LLC hedge fund in September 2006 suggests that AUM of US$9 billion is no longer “large.” 19

Levin, Onatski, Williams and Williams (2005) contains some support for this view. They report the finding that the performance of the optimal policy in a “microfounded” model of a New-Keynesian closed economy with capital formation, assumed to represent the US, is closely matched by a simple operational rule that focuses solely on stabilizing nominal wage inflation. Admittedly, there is no financial sector or financial intermediation in the model, the model is (log-)linear and the disturbances are (I think) Gaussian. But the optimal monetary policy is derived by optimising the (non-quadratic) preferences of the representative household and there is Brainard-type parameter uncertainty about 31 parameters. 20

21 My cats, however, do indeed have fat tails, so there may be new areas of application for the PP.

Non-linearities abound in even the simplest monetary model. To name but a few: the non-negativity constraint on the nominal interest rate; the non-negativity 22

08 Book.indb 621

2/13/09 3:59:21 PM

622

Willem H. Buiter

constraint on gross investment; positive subsistence constraints on consumption; borrowing constraints; the financial accelerator (Bernanke and Gertler, 1989; Bernanke, Gertler and Gilchrist, 1999; Bernanke, 2007); local hysteresis due to sunk costs; any model in which (a) prices multiply quantities and (b) asset dynamics are constrained by intertemporal budget constraints. Although the time series used by econometricians are short (at most a couple of centuries for most quantities; a bit longer for a very small number of prices), the estimated residuals often exhibit both skew and kurtosis. From other applications of dynamic stochastic optimisation we know that different non-linearities generated huge differences in the optimal decision rule. In the theory of optimal investment under uncertainty, strictly convex costs of capital stock adjustment make gradual adjustment of the capital stock optimal. Sunk costs of investment and disinvestment make for “bangbang” optimal investment rules and for “zones of inaction.” For an exploration of some of the implications of uncertainty for optimal monetary policy outside the LQG framework, see the collection of articles in Federal Reserve Bank of St. Louis Review (2008). 23 In the previous statement I hold constant (independent of the individual household’s consumption vs. saving decision) the future expected and actual sequence of after-tax labour income, profits, interest rates and asset prices. In a Keynesian, demand-constrained equilibrium, the aggregation of the individual consumption choices, now and in the future, will in general affect the equilibrium levels of output, employment, interest rates and asset prices.

At the request of Anil Kashyap, I here provide the relevant quotes. I omitted them in the version presented at the Symposium because I felt there was no need to “rub in” the errors. All that matters is that this shared analytical error may well have led to an excessively expansionary policy by the Fed. 24

Bernanke (2007): “If the financial accelerator hypothesis is correct, changes in home values may affect household borrowing and spending by somewhat more than suggested by the conventional wealth effect because changes in homeowners’ net worth also affect their external finance premiums and thus their costs of credit.” Kohn (2006): “Between the beginning of 2001 and the end of 2005, the constantquality price index for new homes rose 30 percent and the purchase-only price index of existing homes published by the Office of Federal Housing Enterprise Oversight (OFHEO) increased 50 percent. These increases boosted the net worth of the household sector, which further fueled (sic) the growth of consumer spending directly through the traditional ‘wealth effect’ and possibly through the increased availability of relatively inexpensive credit secured by the capital gains on homes.” Kroszner (2005): “As some of the ‘froth’ comes off of the housing market—thereby reducing the positive ‘wealth effect’ of the strength in the housing sector—and people fully adjust to higher energy prices, I see the growth in real consumer spending inching down to roughly 3 percent next year.”

08 Book.indb 622

2/13/09 3:59:21 PM

Central Banks and Financial Crises

623

Kroszner (2008):“Falling home prices can have local and national consequences because of the erosion of both property tax revenue and the support for consumer spending that is provided by household wealth.” Mishkin (2008a, p.363): “By raising or lowering short-term interest rates, monetary policy affects the housing market, and in turn the overall economy, directly and indirectly through at least six channels: through the direct effects of interest rates on (1) the user cost of capital, (2) expectations of future house-price movements, and (3) housing supply; and indirectly through (4) standard wealth effects from house prices, (5) balance sheet, credit-channel effects on consumer spending, and (6) balance sheet, credit channel effects on housing demand.” Mishkin (2008a, p. 378): “Although FRB/US does not include all the transmission mechanisms outlined above, it does incorporate direct interest rate effects on housing activity through the user cost of capital and through wealth (and possibly credit-channel) effects from house prices, where the effects of housing and financial wealth are constrained to be identical.” Plossser (2007): “Changes in both home prices and stock prices influence household wealth and therefore impact consumer spending and aggregate demand.” Plosser (2007):“To the extent that reductions in housing wealth do occur because of a decline in house prices, the negative wealth effect may largely be offset for many households by higher stock market valuations.” The technically excellent recent paper by Kiley (2008) is therefore, as regards the usefulness of core inflation as the focus of the price stability leg of the Fed’s dual mandate, completely beside the point. It shows that, if you want to predict future headline inflation and you restrict your data set to current and past headline inflation and core inflation, you should definitely make use of the information contained in the core inflation data. But who would predict or target future inflation making use only of current and past headline and core inflation data? 25

26 A nation’s primary deficit is its current account deficit, excluding net foreign investment income or, roughly, the trade deficit plus net grant outflows.

In part, it may also be a peso-problem or “fake alpha” phenomenon, that is, the higher expected return is a compensation for risk that has not (yet) materialised. The market is aware of the risk, and prices it, but the econometrician has insufficient observations on the realisation of the risk in his sample. 27

A boost to public spending on goods and services or measures to stimulate domestic capital formation would help sustain demand but would prevent the necessary correction of the external account. 28

To avoid getting hoist immediately on my own “housing wealth isn’t wealth” petard, assume that the value of the first home equals the present value of the remaining lifetime housing services the homeowner plans to consume. At the end 29

08 Book.indb 623

2/13/09 3:59:21 PM

624

Willem H. Buiter

of the exercise, the reader can decide for him or herself whether this economy contains a non-homeowning renter who may be better off as the result of the fall in the price of the second home. To make the example work as stands, the second home should be a buy-to-let purchase aimed at the foreign tourist trade. 30 The open letter procedure is a useful part of the communication and accountability framework of the BoE. It requires the Governor to write an open letter to the Chancellor whenever the inflation rate departs by more than 1 percent from its target (in either direction). In that open letter, the Governor, on behalf of the Monetary Policy Committee (MPC) gives the reasons for the undershoot or overshoot of the inflation target, what the MPC plans to do about it, how long it is expected to take until inflation is back on target and how all this is consistent with the Bank’s official mandate. The current inflation target is an annual inflation rate of 2 percent for the Consumer Price Index (CPI). With actual year-on-year inflation at 3.3 percent in May 2008, an open letter (the second since the creation of the MPC in 1997) was due.

$120 per barrel would be a 240 percent increases in the 20-year average real price of oil for the US and a 170 percent increase for the euro area. 31

Perhaps the Treasury sets it? See endnote 4.

32

Complementary in the sense that an increase in the energy input raises the marginal products of labour and capital. 33

The TSLF is a single-price auction, where accepted dealer bids will be awarded at the same fee rate, which is equal to the lowest fee rate at which any bid was accepted. Dealers may submit two bids for the basket of eligible general Treasury collateral at each auction. 34

35 Schedule 1 collateral is all collateral eligible for tri-party repurchase agreements arranged by the Open Market Trading Desk (that is, all collateral acceptable in regular Fed open market operations). Schedule 2 collateral is all Schedule 1 collateral plus AAA/Aaa-rated Private-Label Residential MBS, AAA/Aaa-rated Commercial MBS, Agency CMOs and other AAA/Aaa-rated ABS.

It is revalued daily to ensure that, should the value of the collateral have declined, the primary dealer puts up the additional collateral required to restore the required level of collateralisation. With a well-designed revaluation mechanism, such “margin calls” do, of course, make sense. 36

http://www.newyorkfed.org/markets/pdcf_terms.html.

37

Apparently the French central bank President had not bothered to inform his US counterpart that a possible reason behind the stock market rout in Europe could be the manifestation of the stock sales prompted by the discovery at the Société Générale of Mr. Kerviel’s exploits. If true it is extraordinary. 38

08 Book.indb 624

2/13/09 3:59:21 PM

Central Banks and Financial Crises

625

The term was coined as a characterisation of the interest rate cuts in October and November 1998 following the collapse of Long-Term Capital Management (LTCM). 39

40 Short-term credit markets froze up after the French bank BNP Paribas suspended withdrawals from three investment funds/hedge funds it owned, citing problems in the US subprime mortgage sector. BNP said it could not value the assets in the funds, because the markets for pricing the assets had disappeared.

Eleven countries joined together to form the Eurosystem on January 1, 1999. There are 15 euro area members now and 16 on January 1, 2009, when Slovakia joins. 41

The probability of default on a collateralised loan like a repo is the joint probability of both the borrowing bank defaulting and the issuer of the security used as collateral defaulting. The probability of such a double default will be low but not zero under current circumstances. It may be quite high, when RMBS are offered as collateral, if the borrowing bank is also the bank that originated the mortgages backing the RMBS. 42

Between August 2007 and July 2008, the share of Spanish banks in the Eurosystem’s allocation of main refinancing operations and longer-term refinancing operations went up from about 4 percent to over 10.5 percent. The share of Irish banks went up from around 4.5 percent to 9.5 percent. It cannot be a coincidence that Spain and Ireland are the euro area member states with the most vulnerable construction and real estate sectors. Another measure of the increase in the scale of the Eurosystem’s lending to the Spanish banks since the beginning of the crisis in August 2007, is the value of the monthly loans extended to Spanish banks by the Banco de España. This went from a low of about €23 billion in August 2007 to a high of more than €75 billion in December 2007 (for those worried about seasonality, the December 2006 figure was just under €30 billion). 43

On May 30, 2008, Banco Santander sold Antonveneta to Banca Monte dei Paschi di Siena, an Italian bank, so the fiscal backing question mark raised by the takeovers highlighted in the main text has been erased again. This does not affect the relevance of the point that with foreign-owned banks, operating in many jurisdictions, it is not obvious which national fiscal authorities will foot the fiscal cost of a bailout. The point applies across the world, but is especially pressing for the euro area, where a supranational central bank operates alongside 15 national fiscal authorities and no supranational fiscal authority. 44

BaFin is short for Bundesanstalt für Finanzdienstleistungaufsicht.

45

Bradford & Bingley’s £400 million cash call closed on Friday, August 15, 2008. The six high street banks that, at the prompting of the BoE and the Financial Services Authority, had agreed to underwrite the rights issue are likely to be left with sizeable unplanned stakes in B&B. 46

08 Book.indb 625

2/13/09 3:59:21 PM

626

Willem H. Buiter

References Adalid, Ramón, and Carsten, Detken (2007), “Liquidity shocks and asset price boom/ bust cycles,” European Central Bank working Paper Series No. 732, February. Adrian, Tobias, and Hyun Song Shin (2007a), “Liquidity, Monetary Policy and Financial Cycles,” forthcoming in Current Issues in Economics and Finance. Adrian, Tobias, and Hyun Song Shin (2007b), “Liquidity and Leverage,” Mimeo, Princeton University, September. Bäckström, Urban (1997), “What Lessons Can be Learned from Recent Financial Crises? The Swedish Experience” Remarks made at the Federal Reserve Symposium “Maintaining Financial Stability in a Global Economy,” Jackson Hole, Wyoming, USA, August 29. Bagehot, Walter (1873), Lombard Street: a description of the money market. Bank of Canada (2008), “The Bank of Canada’s Target for the Overnight Interest Rate Policy Implementation Framework,” Bank of Canada Review article by Christopher Reid. http://www.bank-banque-canada.ca/en/pdf/target_170507.pdf. Bank of England (2008a), The Framework for the Bank of England’s Operations in the Sterling Money Markets (the “Redbook”), January. Bank of England (2008b), Inflation Report, February. Baxter, A., and R.G. King (1999), “Measuring Business Cycles Approximate BandPass Filters for Economic Time Series,” International Economic Review 81, pp. 575-93. Bernanke, Ben S. (2002), “Asset price ‘bubbles’ and monetary policy.” Remarks before the New York Chapter of the National Association of Business Economics, New York, October 15. Bernanke, Ben S. (2005), “The Economic Outlook,” Remarks at a Finance Committee luncheon of the Executives’ Club of Chicago, Chicago, Illinois, March 8, http://www.federalreserve.gov/boarddocs/speeches/2005/20050308/default.htm. Bernanke, Ben S. (2007), “The Financial Accelerator and the Credit Channel,” speech given at The Credit Channel of Monetary Policy in the Twenty-first Century Conference, Federal Reserve Bank of Atlanta, Atlanta, Georgia, June 15; http:// www.federalreserve.gov/newsevents/speech/Bernanke20070615a.htm. Bernanke, Ben S. (2008a), “Financial Markets, the Economic Outlook, and Monetary Policy,” speech given at the Women in Housing and Finance and Exchequer Club Joint Luncheon, Washington, D.C., January 10.

08 Book.indb 626

2/13/09 3:59:21 PM

Central Banks and Financial Crises

627

Bernanke, Ben S. (2008b), “Outstanding Issues in the Analysis of Inflation,” speech given at the Federal Reserve Bank of Boston’s 53rd Annual Economic Conference, Chatham, Massachusetts, June 9. http://www.federalreserve.gov/newsevents/ speech/bernanke20080609a.htm. Bernanke, Ben S., and Mark Gertler (1989). “Agency Costs, Net Worth, and Business Fluctuations,” American Economic Review, vol. 79, March, pp. 14-31. Bernanke, Ben and Mark Gertler (2001), “Should Central Banks Respond to Movements in Asset Prices?” American Economic Review no. 91, May, pp. 253-57. Bernanke, Ben S., Mark Gertler, and Simon Gilchrist (1999), “The Financial Accelerator in a Quantitative Business Cycle Framework,” in John B. Taylor and Michael Woodford, eds., Handbook of Macroeconomics, vol. 1, part 3. Amsterdam: North-Holland, pp. 1341-93. Blinder, Alan S. and Ricardo Reis (2005). “Understanding the Greenspan standard,” Proceedings, Federal Reserve Bank of Kansas City, issue Aug, pp. 11-96. Borio, C., C. Furfine and P. Lowe (2001): “Procyclicality of the financial system and financial stability: issues and policy options,” in “Marrying the macro- and microprudential dimensions of financial stability,” BIS Papers, no 1, March, pp 1-57. Bryan, Michael F., and Steven G. Cecchetti. (1994). “Measuring Core Inflation.” In Monetary Policy, edited by N. Gregory Mankiw, 195-215. Chicago: University of Chicago Press. Buchanan, Mike, and Themistoklis Fiotakis (2004), “House Prices: A Threat to Global Recovery or Part of the Necessary Rebalancing?” Goldman Sachs Global Economics Paper No. 114, July 15. Buiter, Willem H. (2004), “The Elusive Welfare Economics of Price Stability as a Monetary Policy Objective: Should New Keynesian Central Bankers Pursue Price Stability?” NBER Working Paper No. 10848, October. Buiter, Willem H. (2006), “Dark Matter or Cold Fusion?”Goldman Sachs Global Economics Paper No. 136, Monday, January 16, 2006, pp. 1-16. Buiter, Willem H. (2007a) “Central Banks as Market Makers of Last Resort 2,” August 17, FT.Com, Maverecon blog. Buiter, Willem H. (2007b), “Central Banks as Market Makers of Last Resort 3: Setting the prices,” August 21, FT.Com, Maverecon blog. Buiter, Willem H. (2007c), “Central banks as market makers of last resort 4: Liquidity, markets and mechanisms,” August 23, FT.com, Maverecon blog.

08 Book.indb 627

2/13/09 3:59:21 PM

628

Willem H. Buiter

Buiter, Willem H. (2007d), “Central banks as market makers of last resort 5: A restatement,” September 2, FT.com, Maverecon blog. Buiter, Willem H. (2007e), “Seigniorage,” economics–The Open-Access, Open-Assessment EJournal, 2007-10. http://www.economics-ejournal.org/economics/journalarticles/2007-10. Buiter, Willem H. (2007f), “Lessons from the 2007 financial crisis,” CEPR Policy Insight No. 18, December; http://www.cepr.org/pubs/PolicyInsights/PolicyInsight18.pdf. Buiter, Willem H. (2008a), “Can Central Banks go Broke?” CEPR Policy Insight No. 24, May 17. Buiter, Willem H. (2008b), “Lessons from the North Atlantic Financial Crisis”, paper prepared for presentation at the conference “The Role of Money Markets” jointly organised by Columbia Business School and the Federal Reserve Bank of New York on May 29-30, 2008. Buiter, Willem H. (2008c), “Housing Wealth Isn’t Wealth,” NBER Working Paper No. 14204, July. Buiter, Willem H, and Clemens Grafe (2004), “Patching up the Pact: Some Suggestions for Enhancing Fiscal Sustainability and Macroeconomic Stability in an Enlarged European Union,” The Economics of Transition, Volume 12 (1) 2004, pp. 67–102. Complete citation information for the final version of the paper, as published in the print edition of The Economics of Transition is available on the Blackwell Synergy online delivery service, accessible via the journal’s website at http://www.blackwellpublishing.com/journal.asp?ref=0967-0750 or at http://www. blackwell-synergy.com. Buiter, Willem H, and Nikolaos Panigirtzoglou (2003), “Overcoming the Zero Bound on Nominal Interest Rates with Negative Interest on Currency: Gesell’s Solution,” Economic Journal, Volume 113, Issue 490, October, pp. 723-746. Buiter, Willem H., and Anne C. Sibert (2007a), “The Central Bank as Market Maker of Last Resort 1,” August 12. FT.com, Maverecon blog. Buiter, Willem H., and Anne C. Sibert (2008b), “A missed opportunity for the Fed,” August 17, FT.com, Maverecon blog. Calvo, Guillermo (1983), ”Staggered Contracts in a Utility-Maximizing Framework,” Journal of Monetary Economics, September. Calvo, Guillermo, Oya Celasun and Michael Kumhof (2007), “Inflation inertia and credible disinflation: The open economy case,” Journal of International Economics, Elsevier, vol. 73(1), pages 48-68, September. Cecchetti, Stephen G. (2008), “Monetary Policy and the Financial Crisis of 2007-2008,” CEPR Policy Insight No. 21, April; http://www.cepr.org/pubs/ PolicyInsights/PolicyInsight21.pdf.

08 Book.indb 628

2/13/09 3:59:21 PM

Central Banks and Financial Crises

629

Chow, Gregory (1975), Analysis and Control of Dynamic Economic Systems, John Wiley & Sons, New York. Chow, Gregory (1981), Econometric Analysis by Control Methods, John Wiley & Sons, New York. Chow, Gregory (1997), Dynamic Economics: Optimization by the Lagrange Method, Oxford University Press. Clews, Roger (2005), “Implementing monetary policy: Reforms to the Bank of England’s operations in the money market,” Bank of England Quarterly Bulletin, Summer. Cogley, Timothy (2002), “A Simple Adaptive Measure of Core Inflation,” Journal of Money, Credit and Banking, Blackwell Publishing, vol. 34(1), pages 94-113, February. Cogley, Timothy, and Thomas J. Sargent (2001), “Evolving Post-World War II U.S. Inflation Dynamics,” NBER Macroeconomics Annual, Vol. 16, pp. 331-373, The University of Chicago Press. Cogley, Timothy, and Thomas J. Sargent (2005). “The conquest of US inflation: Learning and robustness to model uncertainty,” Review of Economic Dynamics, Elsevier for the Society for Economic Dynamics, vol. 8(2), pages 528-563, April. Counterparty Risk Management Group (2005), Toward Greater Financial Stability: A Private Sector Perspective; The Report of the Counterparty Risk Management Policy Group II, July 27. http://www.crmpolicygroup.org. Diamond D. W., (2007). “Banks and liquidity creation: A simple exposition of the Diamond-Dybvig model,” Federal Reserve Bank of Richmond Economic Quarterly 93 (2), pp. 189–200. Diamond D. W. and P. H. Dybvig (1983). “Bank runs, deposit insurance, and liquidity,” Journal of Political Economy 91 (3), pp. 401–19. Dolmas, Jim (2005), “Trimmed mean PCE inflation,” Federal Reserve Bank of Dallas Research Department Working Paper 0506, July 25. Edelstein, Robert H., and Sau Kim Lum (2004), “House prices, wealth effects, and the Singapore macroeconomy,” Journal of Housing Economics, Volume 13, Issue 4, December 2004, Pages 342-367. Epstein, Gerald A., and Thomas Ferguson (1984), “Monetary Policy, Loan Liquidation and Industrial Conflict: The Federal Reserve and The Open Market Operations of 1932,” Journal of Economic History, Vol. XLIV, No. 4, December, 1984, pp. 957-983.

08 Book.indb 629

2/13/09 3:59:21 PM

630

Willem H. Buiter

European Central Bank (2006), The Implementation of Monetary Policy in the Euro Area, September 2006; General Documentation on Eurosystem Monetary Policy Instruments and Procedures, ISSN 1725-714X (print), ISSN 1725-7255 (online). European Central Bank (2008), European Survey of Professional Forecasters, http:// www.ecb.eu/stats/prices/indic/forecast/html/index.en.html. Federal Reserve Bank of St. Louis Review (2008), Monetary Policy Under Uncertainty, Proceedings of the Thirty-Second Annual Economic Policy Conference of the Federal Reserve Bank of St. Louis, Volume 90, Number 4, July/August. Federal Reserve System (2002), Alternative Instruments for Open Market and Discount Window Operations. Feldstein, Martin (2008), “Concluding Remarks,” in Housing, Housing Finance and Monetary Policy, a Symposium Sponsored by The Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming, August 30-Spetember 1, 2007, pp. 489-500, Federal Reserve Bank of Kansas City. Ferguson, Roger W. (2005), “Asset Prices and Monetary Liquidity,” remarks given to the Seventh Deutsche Bundesbank Spring Conference, Berlin, Germany, May 27. Flemming, John (1976), Inflation. London: Oxford University Press, 1976. Gollier, Christian, and Nicolas Treich (2003), “Decision-Making Under Scientific Uncertainty: The Economics of the Precautionary Principle,” The Journal of Risk and Uncertainty, 27:1; 77–103. Goodhart, Charles (2002), “Myths about the Lender of Last Resort,” in Goodhart, Charles and Gerhard Illing (eds.) (2002) Financial Crises, Contagion, and the Lender of Last Resort, a Reader, Oxford University Press, pp. 227–248. Goodhart, Charles (2004), “Bank Regulation and Macroeconomic Fluctuations,” Oxford Review of Economic Policy, Oxford University Press, vol. 20(4), pages 591-615, Winter. Goodhart, Charles and Gerhard Illing (eds.) (2002) Financial Crises, Contagion, and the Lender of Last Resort, a Reader, Oxford University Press. Goodhart, Charles and Avinash Persaud (2008a), “How to avoid the next crash,” Financial Times, Comment, January 20. Goodhart, Charles and Avinash Persaud (2008b), “A party pooper’s guide to financial stability”, Financial Times, Comment, June 4. Gordy, Michael B., and Bradley Howells (2004), “Procyclicality in Basel II: Can We Treat the Disease Without Killing the Patient?” Board of Governors of the Federal Reserve System; First draft: April 25, 2004. This draft: May 12, 2004.

08 Book.indb 630

2/13/09 3:59:21 PM

Central Banks and Financial Crises

631

Gourinchas, Pierre-Olivier, and Hélène Rey. (2007), “From World Banker to World Venture Capitalist: U.S. External Adjustment and the Exorbitant Privilege.” In G7 Current Account Imbalances: Sustainability and Adjustment, edited by Richard H. Clarida. Chicago: University of Chicago Press (for NBER). Greenspan, Alan (2002): “Opening Remarks,” speech at a symposium in Jackson Hole, Wyoming, August 30, 2002. Greenspan, Alan (2005), “Opening Remarks,” speech at the Jackson Hole Symposium, Wyoming. Greiber, Claus, and Ralph Setzer (2007), “Money and housing—evidence for the euro area and the US,” Deutsche Bundesbank, Discussion Paper, Series 1: Economic Studies, No 12/2007. Hanson, Jon, and David Yosifon, “The Situation: An Introduction to the Situational Character, Critical Realism, Power Economics, and Deep Capture,” 152 U. Pa. L. Rev. 129 (2003). Hausmann, Ricardo, and Federico Sturzenegger (2007), “The missing dark matter in the wealth of nations and its implications for global imbalances,” Economic Policy, Volume 22, Issue 51, July, pp. 469-518. Hellman, Joel S., Geraint Jones and Daniel Kaufmann (2000), “Seize the State, Seize the Day: State Capture, Corruption and Influence in Transition,” World Bank Policy Research Working Paper No. 2444, September. IIF (2008), “IIF Final Report of the Committee on Market Best Practices (.pdf ),” Institute of International Finance, Inc. July 17. Ingves, S. and Lind, G., (1996), “The management of the bank crisis—in retrospect,” Quarterly Review, No. I, pp. 5-18, Sveriges Riksbank. Kashyap, A. K., and J. C. Stein (2004): “Cyclical implications of the Basel II capital standards,” Economic Perspectives, Federal Reserve Bank of Chicago, First Quarter, pp 18-31. Kiley, Michael T. (2008), “Estimating the common trend rate of inflation for consumer prices and consumer prices excluding food and energy prices,” Finance and Economics Discussion Series 2008-38, Divisions of Research & Statistics and Monetary Affairs, Federal Reserve Board, Washington, D.C. King, Mervyn (2007), paper submitted to the Treasury Committee, September 12. King, Mervyn (2008), “TC Opening Statement March 26, 2008”, http://www. bankofengland.co.uk/publications/other/treasurycommittee/ir/tsc080326.pdf. Kohn, Donald L. (2006), Economic Outlook, speech given at the Money Marketeers of New York University, New York, New York, October 4. http://www. federalreserve.gov/newsevents/speech/kohn20061004a.htm.

08 Book.indb 631

2/13/09 3:59:21 PM

632

Willem H. Buiter

Kroszner, Randall S. (2005), “‘Rodney Dangerfield’ redux: Still no respect for the economy,” speech given at the annual business forecast luncheon of the University of Chicago Graduate School of Business, Wednesday, December 7. Kroszner, Randall S. (2008), testimony on the Federal Housing Administration Housing Stabilization and Homeownership Act before the Committee on Financial Services, U.S. House of Representatives, April 9, 2008. http://www. federalreserve.gov/newsevents/testimony/kroszner20080409a.htm. Laffont, J. J., and J. Tirole, (1991), “The politics of government decision making: A theory of regulatory capture,” Quarterly Journal of Economics, 106(4): 1089-1127. Levin, Andrew, Alexei Onatski, John C. Williams and Noah Williams (2005), “Monetary Policy under Uncertainty in Micro-Founded Macroeconometric Models,” in Mark Gertler and Kenneth Rogoff, eds., NBER Macroeconomics Annual 2005. Cambridge, Mass.: MIT Press, pp. 229-88. Levine, M. E., and J. L. Forrence (1990), “Regulatory capture, public interest, and the public agenda: Toward a synthesis,” Journal of Law Economics Organization, 6: 167-198. Lucas, Robert E. (1975), “An Equilibrium Model of the Business Cycle,” Journal of Political Economy. Lucas, Robert E., and Nancy L. Stokey (1989), Recursive Methods in Economic Dynamics. Merton, Robert C. (1990, 1992), Continuous-Time Finance, Oxford, U.K.: Basil Blackwell, 1990. (Rev. ed., 1992.) Minsky, Hyman (1982), “The Financial-Instability Hypothesis: Capitalist processes and the behavior of the economy,” in C. Kindleberger and Laffargue, editors, Financial Crises. Mishkin, Frederic S. (2008a), “Housing and the Monetary Transmission Mechanism,” in Housing, Housing Finance and Monetary Policy, a Symposium Sponsored by The Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming, August 30-September 1, 2007, pp. 359- 413, Federal Reserve Bank of Kansas City. Mishkin, Frederic S. (2008b), “Monetary Policy Flexibility, Risk Management, and Financial Disruptions,” speech given at the Federal Reserve Bank of New York, New York, January 11. http://www.federalreserve.gov/newsevents/speech/ mishkin20080111a.htm. Muellbauer, John N. (2008), “Housing, Credit and Consumer Expenditure,” in Housing, Housing Finance and Monetary Policy, a Symposium Sponsored by The Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming, August 30-Spetember 1, 2007, pp. 267-334, Federal Reserve Bank of Kansas City.

08 Book.indb 632

2/13/09 3:59:21 PM

Central Banks and Financial Crises

633

Mundell, Robert A. (1962.), “The Appropriate Use of Monetary and Fiscal Policy for Internal and External Stability,” IMF Staff Papers, reprinted in Robert A. Mundell, International Economics, 1968. OECD (2008), “The implications of supply-side uncertainties for economic policy,” OECD Economic Outlook, May, Chapter 3. Olson, Mancur (1965) [1971]. The Logic of Collective Action: Public Goods and the Theory of Groups, Revised edition, Harvard University Press. Philip, R., P. Coelho and G. J. Santoni (1991), “Regulatory Capture and the Monetary Contraction of 1932: A Comment on Epstein and Ferguson,” The Journal of Economic History, Vol. 51, No. 1, March, pp. 182-189.   Plosser, Charles I. (2007), “House Prices and Monetary Policy,” Lecture, European Economics and Financial Centre Distinguished Speakers Series, July 11, London, UK. http://www.philadelphiafed.org/publicaffairs/speeches/plosser/2007/07-11-07_euroecon-finance-centre.cfm. Quah, Danny, and Shaun P. Vahey, (1995). “Measuring Core Inflation?” Economic Journal, Royal Economic Society, vol. 105(432), pages 1130-44, September. Rich, Robert, and Charles Steindel (2007). “A comparison of measures of core inflation,” Economic Policy Review, Federal Reserve Bank of New York, issue Dec, pages 19-38. Small, David H., and James A. Clouse (2004), “The Scope of Monetary Policy Actions Authorized under the Federal Reserve Act,” Board of Governors of the Federal Reserve System Research Paper Series, FEDS Papers 20004-40, July. Spaventa, Luigi (2008), “Avoiding Disorderly Deleveraging,” CEPR Policy Insight No. 22, May; http://www.cepr.org/pubs/PolicyInsights/PolicyInsight22.pdf. Stigler, G. (1971), “The theory of economic regulation,” Bell J. Econ. Man. Sci. 2:3-21. Summers, Lawrence (2008), “Why America Must Have a Fiscal Stimulus,” Financial Times, Op-Ed Column, January 6. Trichet, Jean-Claude (2008), Introductory statement, August 7. http://www.ecb. int/press/pressconf/2008/html/is080807.en.html. Wolf, Martin (2008), “How imbalances led to credit crunch and inflation,” Financial Times column, June 17. Woodford, Michael (2003), Interest & Prices; Foundations of a Theory of Monetary Policy, Princeton University Press, Princeton and Oxford.

08 Book.indb 633

2/13/09 3:59:21 PM

08 Book.indb 634

2/13/09 3:59:21 PM

Commentary: Central Banks and Financial Crises Alan S. Blinder

Buiter papers don’t pull punches. They have attitude. They often feature an alluring mix of brilliant insights and outrageous statements. And they tend to be verbose. This tome displays all those traits. But since it runs 141 pages, I have about 6 seconds per page. So I must be selective. I will therefore concentrate on two big issues: Generically, what are the proper functions of a central bank? Specifically, has the Fed’s performance in this crisis really been that bad? Starting with the second. Does the Fed deserve such low marks? Willem’s critique of the Fed boils down to saying it was both too soft-hearted and extremely muddled in its thinking. Its attempts to avoid painful adjustments that were necessary, appropriate, and in many ways inevitable have planted moral hazard seeds all over the financial landscape. And its entire framework for conducting monetary policy is fundamentally wrong. Other than that, it did well! Now, you have to give credit to a guy with the nerve to come here, with black bears on the outside and the FOMC on the inside, and be so critical of the Fed—which has earned kudos in the financial community. But those very kudos, Willem says, are symptomatic of a deep problem. In his words, “a key reason [for the policy errors] is 635

08 Book.indb 635

2/13/09 3:59:21 PM

636

Alan S. Blinder

that the Fed listens to Wall Street and believes what it hears…the Fed is too close to the financial markets and leading financial institutions, and too responsive to their special pleadings, to make the right decisions for the economy as a whole” (pg. 599-600). There is a valid point here. I am, after all, the one who warned that central banks must be as independent of the markets as they are of politics­—that they must “listen to the markets” only in the sense that you listen to music, not in the sense that you listen to your mother— and that central banks sometimes fail to do so.1 But has the Fed really done as badly as Willem says? I think not. While the Fed’s performance has not been flawless, I think it’s been pretty good under the circumstances. Those last three words are important. Recent circumstances have been trying and, in many respects, unique. Unusual and exigent circumstances, to coin a phrase, require improvisation on the fly—and improvisation is rarely perfect. So I give the Fed high marks while Willem gives them low ones. Let me illustrate the different grading standards with a short, apocryphal story that Willem may remember from his childhood in Holland (even though it’s based on an American story). One day, a little Dutch boy was walking home when he noticed a small leak in the dike that protected the town. He started to stick his finger in the hole. But then he remembered the moral hazard lessons he had learned in school. “Wait a minute,” he thought. “The companies that built this dike did a terrible job. They don’t deserve a bailout, and doing so would just encourage more shoddy construction. Besides, the foolish people who live here should never have built their homes on a flood plain.” So the boy continued on his way home. Before he arrived, the dike burst and everyone for miles around drowned—including the little Dutch boy. Perhaps you’ve heard the Fed’s alternative version of the story. In this kinder, gentler version, the little Dutch boy, somewhat desperate and worried about the horrors of a flood, stuck his finger in the dike and held it there until help arrived. It was painful and not guaranteed to work—and the little boy would rather have been doing other

08 Book.indb 636

2/13/09 3:59:21 PM

Commentary

637

things. But he did it anyway. And all the people who lived behind the dike were saved from the error of their ways. While you decide which version you prefer, I will take up three of Willem’s six criticisms of the Fed’s monetary policy framework. I don’t have time for all six. The risk management approach First, methinks the gentleman doth protest too much about the difference between optimization with a quadratic loss function and the Fed’s risk management approach, which allegedly focused exclusively on output while ignoring inflation. Many of you will recall that, at the 2005 Jackson Hole conference, some of us debated whether these two approaches were different at all.2 I think they are different. But the truth is that, with a quadratic loss function, any shock that raises the unemployment forecast and lowers the inflation forecast should induce easier monetary policy. You don’t need minimax or anything fancy to justify rate cuts. Welcoming a recession? Second, the spirit of Andrew Mellon apparently lives on in the person of Willem Buiter. Mellon’s famous advice to President Hoover in 1931 was: Liquidate labor, liquidate stocks, liquidate the farmers, liquidate real estate. It will purge the rottenness out of the system... . People will work harder, live a more moral life…and enterprising people will pick up the wrecks from less competent people. Willem’s advice to Chairman Bernanke in 2008 is that the U.S. economy needs a recession—and the sooner the better. Why? Because a recession is the only way to whittle the current account deficit down to size—you might say, to “purge the rottenness out of the system.” Is that really true? What about expenditure switching at approximate full employment? Isn’t that what we did, approximately, during the Clinton years—using a policy mix of fiscal consolidation and easy money?

08 Book.indb 637

2/13/09 3:59:21 PM

638

Alan S. Blinder

Core or headline inflation? Third, I still think the FOMC is correct to focus more on core than headline inflation. Let me explain with the aid of a quotation from Willem’s paper and some charts from a forthcoming paper by Jeremy Rudd and me.3 Willem observes that, “Core inflation is relevant to the price stability leg of the Fed’s mandate to the extent that it is a superior predictor of future headline inflation, over the horizon that the Fed can influence headline inflation” (pg. 559). Exactly right.4 Let’s apply that idea. Chart 1 depicts the simplest and most benign case: an energy price spike like OPEC II. The relative price of energy shoots up but then falls back to where it began. The right-hand panel, based on an estimated monthly pass-through model, shows that such a shock should, first, boost headline inflation way above core, but subsequently push headline well below core. The effects on both headline and core inflation beyond two years are negligible. It seems clear, then, that a rational central bank would focus on core inflation and ignore headline. Chart 2 shows a less benign sort of energy shock: The relative price of energy jumps to a higher plateau and remains there. OPEC I was a concrete example. Once again, the right-hand panel shows that headline inflation leaps above core, but then converges quickly back to it. However, this time core and headline wind up permanently higher. They are also substantially identical after about seven months. So, over the relevant time horizon, it seems that the central bank should again concentrate on core, not headline. Chart 3 depicts the nastiest case which, unfortunately, may apply to the years since 2002. Here the relative price of oil keeps on rising for years. As you can see, both headline and core inflation increase, and there is no tendency for headline to converge back to core. In this case, one can make a coherent argument that the central bank should focus on headline inflation. So is Willem’s criticism correct? Well maybe, but only with the wisdom of hindsight. When there are big surprises, you can be right

08 Book.indb 638

2/13/09 3:59:21 PM

Commentary

639

Chart 1 Effect of a temporary spike in energy prices 1.4

A. Level of real energy price

1.4

1.3

1.3

1.2

1.2

1.1

1.1

1.0

1.0

0.9

-4

0

4

8

12

16

20

24

28

32

36

B. Path of headline and core inflation (monthly change at AR)

3.5

3.5

2.5

0.9

2.5

Headline

1.5

1.5 Core

0.5

0.5

- 0.5

08 Book.indb 639

-4

0

4

8

12

16

20

24

28

32

36

40

44

48

- 0.5

- 1.5

- 1.5

- 2.5

- 2.5

2/13/09 3:59:22 PM

640

Alan S. Blinder

Chart 2 Effect of a permanent jump in energy prices 1.4

A. Level of real energy price

1.4

1.3

1.3

1.2

1.2

1.1

1.1

1.0

1.0

0.9

6.0

-4

0

4

8

12

16

20

24

28

32

36

B. Path of headline and core inflation (monthly change at AR)

0.9

6.0

5.0

5.0 Headline

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0 Core

0.0

0.0 -4

08 Book.indb 640

0

4

8

12

16

20

24

28

32

36

40

44

48

2/13/09 3:59:22 PM

Commentary

641

Chart 3 Effect of a steady rise in energy prices A. Level of real energy price

1.4

1.4

1.3

1.3

1.2

1.2

1.1

1.1

1.0

1.0

0.9

6.0

4

0

-4

8

12

16

20

24

32

28

36

B. Path of headline and core inflation (monthly change at AR)

0.9

6.0

5.0

5.0

4.0

4.0

3.0

3.0

2.0

2.0 Core Headline

1.0

0.0

08 Book.indb 641

-4

0

4

8

12

16

1.0

20

24

28

32

36

40

44

48

0.0

2/13/09 3:59:22 PM

642

Alan S. Blinder

ex ante but wrong ex post. It is well known that the Fed does not attempt to forecast the price of oil but uses futures prices instead. It is also well known that futures prices underestimated subsequent actual prices consistently throughout the period, regularly forecasting either flat or declining oil prices. Thus the Fed inherited and acted upon the markets’ mistakes­­—a forgivable sin, in my book. Remember also that no relative price can rise without limit. Oil prices are finally plateauing or coming down, which will restore the case for core. While I could spend more time defending the FOMC against Willem’s many charges, I think I’ve now said enough to ingratiate myself to our hosts. So let’s proceed to the more generic issues. What should a central bank do? On our first day in central banking kindergarten, we all learned that a central bank has four basic functions: 1. to conduct macroeconomic stabilization policy, or perhaps just to create low and stable inflation; let’s call this “monetary policy proper;” 2. to preserve financial stability, which sometimes means acting as lender of last resort; 3. to safeguard what is often called the financial “plumbing”; and 4. to supervise and regulate banks. Willem doesn’t much care for this list. In previous incarnations, he has argued that the central bank should pursue price stability and nothing else, including no responsibility for either unemployment or financial stability.5 But here he changes his mind and focuses on the lender of last resort (LOLR) function, number 2 on the list. In doing so, he ignores the plumbing issue entirely; he argues that central banks should not supervise banks; and he even suggests—heavens to Betsy!—transferring responsibility for monetary policy decisions elsewhere. I respectfully disagree on all counts.

08 Book.indb 642

2/13/09 3:59:22 PM

Commentary

643

Monetary policy proper On the second day of central banking kindergarten, we all learn that most central banks have multiple monetary policy instruments, including the policy interest rate (in the U.S., the federal funds rate) and lending to banks, which itself includes price (in the U.S., the discount rate), any sort of quantity rationing (including “moral suasion”), and the LOLR function. Willem muses about separating the responsibility for interest rate from this other stuff, which would be quite a radical step. Why? He explains that while the central bank will “have to implement the official policy rate decision…it does not have to make the interest rate decision” because it is “not at all self-evident” that the same skills and knowledge are needed to set the interest rate as to manage liquidity (pg. 530). “Not at all self-evident” seems a pretty thin basis for such a momentous change. On behalf of all the current and past central bankers in the room, may I suggest that it is self-evident that the lender of last resort should also set the interest rate? Reason #1: Emergency liquidity provision occasionally becomes an integral and vital part of monetary policy just as they taught us in central banking kindergarten. Having the Fed set the discount rate while someone else sets the funds rate is akin to putting two sets of hands on the steering wheel. Reason #2: Aren’t we really concerned about financial stability because of what financial instability might do to the overall economy? Who, after all, cares about even wild gyrations in small, idiosyncratic financial markets that have negligible macro impacts? Reason #3: If we take interest rate setting away from the central bank, to whom shall we give it? To a decisionmaking body without the means to execute its decision? To an agency that will almost certainly be less independent than the central bank? Safeguarding the financial plumbing To my way of thinking, but apparently not to Willem’s, one reason central banks have LOLR powers is precisely to enable them to keep the plumbing working during crises. And indeed, central banks

08 Book.indb 643

2/13/09 3:59:22 PM

644

Alan S. Blinder

throughout history have used the window lending for precisely this purpose. I submit that this connection is also self-evident. Bank supervision I come, finally, to the most controversial function. Whether or not the central bank should supervise banks has been vigorously debated for years now, and there are arguments on both sides. Or perhaps I should say there were arguments on both sides until the Northern Rock debacle showed us what can happen when a central bank doesn’t know what’s happening inside a bank to which it might be called upon to lend. Yet, somehow, Willem reaches the opposite conclusion. Why? Because he claims that “cognitive regulatory capture” led the Fed astray. Yet he himself acknowledges that “institution-specific knowledge…made the Fed an effective lender of last resort” (pg. 613). I could rest my case on that statement. It would seem peculiar to leave the lender of last resort ignorant of the creditworthiness of potential borrowers. Market maker of last resort While Willem generally wants to clip the central bank’s wings, he does want to expand the LOLR function to what he calls acting as the MMLR. I don’t much care for the name, since market making normally means buying and selling to smooth or profit from price fluctuations. But what Willem means by MMLR makes sense: “during times when systemically important financial markets have become disorderly and illiquid…the market maker of last resort either buys outright…or accepts as collateral…systemically important financial instruments that have become illiquid” (pg. 525). Ironically, that description fits the Fed’s recent lending policies to a tee. However, Willem raises two legitimate criticisms. First, the Fed values the collateral it takes at prices provided by the clearing banks— which seems rather too trusting. I agree. Second, the Fed has ignored Bagehot’s advice to charge a penalty rate. Lending below market is like making a fiscal transfer—which Willem justifiably questions. But I part company when he argues that central banks should lend only at appropriate risk-adjusted rates. Because the LOLR serves a

08 Book.indb 644

2/13/09 3:59:22 PM

Commentary

645

social purpose broader than profit maximization, it is easy to justify expected risk-adjusted losses in an emergency. In sum, while there is surely room for improvement around the edges, I don’t believe that either the structure or framework of U.S. monetary policy needs the kind of wholesale overhaul that Willem recommends. Cosmetic surgery, maybe. But not a lobotomy.

08 Book.indb 645

2/13/09 3:59:23 PM

646

Alan S. Blinder

Endnotes See Alan S. Blinder, Central Banking in Theory and Practice (MIT Press, 1998), pp. 59-62, which was expanded upon in Alan S. Blinder, The Quiet Revolution (Yale, 2004), Chapter 3. These first of these books was the Robbins Lectures given at the LSE in 1996, which were in turn based on my Marshall Lectures, given at Cambridge in 1995 and hosted by Professor Willem Buiter! 1

See Alan S. Blinder and Ricardo Reis, “Understanding the Greenspan Standard”, in Federal Reserve Bank of Kansas City, The Greenspan Era: Lessons for the Future (proceedings of the August 2005 Jackson Hole conference), pp. 11-96 and the ensuing discussion. 2

3 Alan S. Blinder and Jeremy Rudd, “The Supply-Shock Explanation of the Great Stagflation Revisited”, paper under preparation for the NBER conference on The Great Inflation, September 2008.

Neither Buiter nor I mean to imply that past inflation is the only variable relevant to forecasting future inflation. 4

Willem Buiter, “Rethinking Inflation Targeting and Central Bank Independence,” inaugural lecture, London School of Economics, October 2006. 5

08 Book.indb 646

2/13/09 3:59:23 PM

Commentary: Central Banks and Financial Crises Yutaka Yamaguchi

I want to thank first the Kansas City Fed for inviting me to this splendid symposium. I have found Dr. Buiter’s paper long, comprehensive, thought-stimulating and, of course, provocative. It is an interesting read unless you belonged to one of the targeted institutions. In what follows, I will talk more about my own observations mostly on the Fed, rather than offer direct comments on the paper, but I hope my remarks will cross the path of Dr. Buiter’s here and there. The author is highly critical of the Fed’s performance in the past year, particularly in monetary policy. The sharp contrast between the Fed and the ECB (and the Bank of England) in monetary policy raises a legitimate question of why the Fed has been so aggressively easing. The Fed’s trajectory since last summer appears to me broadly consistent with the weakening U.S. economic growth and the Fed’s dual mandate. But the Fed would not have eased as much as it has if it had not adhered to the “risk management” aspect of monetary policy. The relevant risks here are twofold: a financial systemic instability and inflation. They are both hard to reverse once set in motion or embedded in the system. They point to different policy responses.

647

08 Book.indb 647

2/13/09 3:59:23 PM

648

Yutaka Yamaguchi

The Fed must have weighed the relative importance of these threats and “gambled,” to borrow the word used by Martin Feldstein, to place higher emphasis on the risk of financial disruptions leading to even weaker economic activity. I am sympathetic to this decision and therefore to the Fed’s monetary policy trajectory since last summer. Now let me examine this “risk management” approach in a broader perspective. As I do so, I’ll be a bit less sympathetic. This approach is a key component of the so-called “clean up the mess after a bubble bursts” argument. It has been a conventional wisdom in recent years among many central bankers around the world. But the ongoing crisis prompts me to revisit the argument. Three questions come to my mind. The first is about the timing of such “clean up” operation. Taking a look at the Japanese episode first, the Tokyo stock market peaked at the end 1989. The Bank of Japan began to cut the policy rate oneand-a-half years later in July 1991. The lag from the property market peak is a bit ambiguous, given the nature of the market, but it was probably a little shorter. Twenty-some years later, the U.S. housing market peaked in the second half of 2005. The Fed started to ease two years later. Thus, there is striking similarity between the two countries in the timing of the first interest rate cut after a major bubble burst. The similarity has good reasons: It is difficult to recognize on real time if a bubble has in fact burst or not; it is also difficult to ease monetary policy when economic growth still looks robust and financial markets still stable. Yet if the central bank waited till a turbulence has erupted, it might well be too late. When should the central bank start the mopping-up operation? The second question relates to the exceptional uncertainty in the post-bubble period. We observe in the U.S. economy today unique and substantial uncertainty over the extent of housing price decline, magnitude of losses incurred by the financial system, strength of financial “headwind” against the economy, inflationary potential and so on. These special forces tend to cloud the economic and price picture and, if anything, should make it more difficult for the central bank to take “decisive” actions. Such uncertainty is not new nor is limited to the current U.S. scene. We went through a similar phase of extraordinarily low visibility in the early 1990s. In fact, concern

08 Book.indb 648

2/13/09 3:59:23 PM

Commentary

649

about a resumption of asset price inflation was rather prevalent even a few years after the stock and property market peaks. It is sometimes argued that Japanese monetary policy failed to take early on some decisive easing actions, such as large and permanent interest rate reduction. The failure to do so, the argument goes, led the economy to deflation. Such argument is totally negligent of the then-existing uncertainty and seems to me quite unrealistic. Uncertainty over the state of the financial system is particularly relevant for the central bank. When the financial system gets badly impaired in terms of its capital, it becomes vulnerable to shocks. Sentiment shifts often and false dawn arrives a number of times. Above all, monetary policy transmission seriously weakens if not totally breaks down. In the case of Japan, systemic stability was restored only when significant capital was injected into the banking system using public funds. In my view, the lesson to draw from the Japanese episode should be, above all, the importance of an early and large-scale recapitalization of the financial system. How it can be done should vary according to the given circumstances and national context. The third and last question as regards the “clean up the mess strategy” is that it is inappropriately generalizing one specific experience of addressing the collapsing tech bubble by aggressive rate cuts. But the tech bubble was not after all a major credit bubble. It did not leave behind a massive pile of nonperforming assets. The U.S. financial system was able to emerge from the bubble’s aftermath relatively unscathed. The tech bubble and its aftermath was, if I may say so, an easier type to “clean up” ex-post; it does not have universal applicability to other episodes. From the standpoint of securing financial stability, credit bubbles should be the focus of attention. This brings me to the final segment of my remarks: the role of monetary policy vis-à-vis credit cycle. Proposals abound these days on how to restrain excessive credit growth in times of upswing. Most of them, including Dr. Buiter’s, advocate some regulatory measures. Few are in favor of “leaning against the wind” by monetary policy— so had I thought until I listened to Prof. Shin yesterday.

08 Book.indb 649

2/13/09 3:59:23 PM

650

Yutaka Yamaguchi

Some proposals to use regulatory measures appear sensible. However, I remain skeptical if a regulatory approach alone would work. I happen to belong to the dying species of former central bankers who have had experiences in the past in direct credit controls. Even in the days of heavily regulated banking and financial markets, outright controls tended to invite serious distortions in credit flows. Bank of Japan’s guidance on bank lending, for example, was clearly more effective when supported by higher interest rates, as higher funding cost partially offset the banks’ incentive to lend. That was then. The world has vastly changed, and we now live in highly sophisticated financial markets. Still, importance of affecting the incentives has not much changed. For instance, if we look at the sequence of what happened in the run-up to the current crisis, there was a sustained easy money and low interest rate environment, which drove market participants to search for yield, which resulted in much tighter credit spreads, which then prompted many players to raise leverage, and things collapsed. Simply capping on leverage, for instance, might invite circumventions and distortions unless the root cause of credit expansion was not addressed. I believe a more balanced and symmetric approach to address credit cycles, including “leaning against the wind” by monetary policy, is worth considering in the pursuit for both monetary and financial stability.

08 Book.indb 650

2/13/09 3:59:23 PM

General Discussion: Central Banks and Financial Crises Chair: Stanley Fischer

Mr. Fischer: We all quote Bagehot selectively and forget he operated in a fixed exchange rate environment. Willem says the U.S. has to get the current account adjusted and at the same time should be running higher interest rate policies. The dollar must be an essential part of any of that adjustment, and higher U.S. interest rates don’t help in that regard. The Bagehot rules don’t translate exactly to a system where the exchange rate is flexible. Secondly, about Mundell’s Principle of Effective Market Classification. One of the first things that we learned in micro is about constrained optimization. Sometimes you have one constraint and two objectives, and you have to trade off between them. That’s micro. In macro and in the Mundell Principle—incidentally I learned of it as being Tinbergen’s Principle—rhetoric tends towards the view that you need as many instruments as targets, and that tradeoffs somehow are not allowed. We all frequently say, “Well, the Fed’s only got one instrument. It has to fix the inflation rate.” There may be reasons of political economy to say that, but it’s not true in general that you can’t optimize unless you have as many instruments as targets. Mr. Barnes: In criticizing the Fed for being too sensitive to perceived downside risks in the economy, Willem asserted it’s easier for a central bank to respond to a sharp downturn in activity than to 651

08 Book.indb 651

2/13/09 3:59:23 PM

652

Chair: Stanley Fischer

respond to embedded inflation expectations. That may be true a lot of the time, but it is not clear to me it is true in the context of a postcredit boom when you have high risk of negative feedback loops. I would argue the experience of Japan suggests it can be very difficult to get out of an economic downturn in that kind of environment. Mr. Makin: I very much enjoyed all three presentations. I wanted to very quickly ask the question regarding the little boy with his finger in the dike. First, is the little boy the Fed, the Treasury, or some other institution? Secondly, I think you said, “He keeps his finger in the dike until help arrives and everybody is better off.” What if it takes a really long time for help to arrive in the sense he stuck his finger in the dike and a big wave came along called a recession? What would he do then? Those become critical issues. Finally, in order to influence the answer, I would suggest the bad wall construction was probably the fault of the commercial banks and the people. Silly to be living on the flood plain are the real estate speculators. Maybe with that richer texture, you could comment. Mr. Frenkel: At this conference, we have discussed issues on housing, financial markets, regulation, incentives, moral hazard, etc., but we have discussed very little the macro picture. That is also the way I see Willem’s paper. Three years ago at this conference, we said the current account deficit of the United States is too big, it is not sustainable, and it must decline. The U.S. dollar is too strong, it is not sustainable, and it must decline. The housing market boom is not sustainable; prices must decline. The Chinese currency, along with other Asian currencies, is too weak; they must rise. Some even said interest rates may be too low and pushing us into more risky activities, so we must think about risk management. Here we are three years later and all of these things have happened. We may have had too much of these good things. There are a lot of spillover effects, negative things or whatever. But what we have had is a massive adjustment that was called for, needed, and recognized.

08 Book.indb 652

2/13/09 3:59:23 PM

General Discussion

653

Within this context, the question is, How come all of these disruptions have not yet caused a deeper impact on the U.S. real output? There the answer is the foreign sector. We have had a fantastic cushion coming from the foreign sector. In fact, if you look at U.S. growth, you see all the negative contributions that came from the housing shrinkage were offset by the positive contribution that came from exports. That positive contribution was induced among others by the declining dollar and all of the things we knew had to happen. In fact, we are in a new paradigm in which last year 70 percent of world growth came from emerging markets and only 30 percent from the advanced economies. Within this context, when the dust settles and the financial crisis is behind us, and the lessons are learned, let’s remember one thing. This cushion of the foreign sector is essential for the era of globalization. All of these calls for protectionism that are surfacing in Washington and elsewhere, including the U.S. election debate, would be a disaster. The only reason why the United States is not in a recession today, in spite of the fact there is a significant slowdown, is the foreign sector. We can talk about extinguishing fires and all of these other things, but we need to remember the macro system must produce current account deficits and imbalances that do not create incentives for protectionism. Let’s bring the discussion back to the macro issues. Mr. Mishkin: When I read this paper, I said this paper has a lot of bombs, but maybe a better way to characterize it is there are a lot of unguided missiles that have been shot off now in this context. I only want to deal with one of them, which is the issue of the risk management precautionary principle approach. Willem is even stronger in his statement because he just called it “bogus” in the paper, but actually calls it “bogus science” in his presentation. His reasoning here is the only reason you would use a precautionary principle, or this risk management approach, which many know I advocated, is because of potential for irreversibility in terms of something bad happening. He goes to the literature on environmental risk to discuss this. I wish he had actually read some of the literature on optimal monetary

08 Book.indb 653

2/13/09 3:59:23 PM

654

Chair: Stanley Fischer

policy because it might have been very helpful in this context. Indeed, the literature on optimal monetary policy does point out when you have nonlinearities, where you can get an adverse feedback loop, in particular the literature I am referring to­—which has been very well articulated—is on the zero lower bound interest rate literature. In fact, it argues what you need to do is act more aggressively in order to deal with the potential for a nonlinear feedback loop. On that context, the issue of science here does have something to say, and we do have literature on optimal monetary policy that I think is important to recognize in terms of thinking about this. One other thing is that Mr. Yamaguchi mentioned the AdrianShin paper. I didn’t make a comment on that before, but one little comment here. What that paper does—which is very important—is show there is another transmission mechanism of monetary policy. That was very important. It indicates you should take a look at that in terms of assessing what the appropriate stance of monetary policy should be. It does not argue you have to go and lean against the wind in terms of asset price bubbles. We should be very clear in terms of what the contribution of the paper was. In this case, I am agreeing with Willem, just so we even it up. Mr. Trichet: I thought the session was particularly stimulating. Alan, you said it was not Willem’s habit to pull punches. Well, I think we had a demonstration because we had our own punches, too. I would like to make two points. The first point is to see in which universe the various central banks are placed. For us, things are very clear. We have—as I have often said—one needle in our compass. We don’t have to engage in any arbitrage between various goals. We have a single goal, which is to deliver price stability in the medium term. It is true that at the very beginning of the turmoil and turbulences in mid-2007, we thought it was very important to make this point as clearly as possible. It was nothing new there, of course, because it was only a repetition of what we had always said. It was understood quite correctly that we had one needle in our compass, and we were very clear in saying that we then would strictly separate between what was

08 Book.indb 654

2/13/09 3:59:23 PM

General Discussion

655

needed for monetary policy to deliver price stability in the medium term and what was needed to handle the operational framework in a period of very high tensions in the money market. My second point relates to the remark by Willem or Peter before, namely that the ECB did pretty well in the circumstances of turmoil in terms of the handling of the operational framework. After further reflection, and taking due account of the very special natural environment of Jackson Hole that is full of biodiversity, it seems to me that the notion to consider regarding the origin of our operational framework is diversity. We had to merge a lot of various frameworks in order to have our system operate from the very start of the euro. Three elements stand out: first, in contrast to the Bank of England or the Fed, we accepted private paper from the very beginning in our operational framework, which was a tradition in at least three countries, including Germany, Austria, France, and others. Second, we could refinance over three months because again it was a tradition which had been a useful experience in a number of countries, again including Germany. And third, we had a framework with a very large number of counterparties, which appears to have been, in the circumstances, extraordinarily useful because we could provide liquidity directly to a very broad set of banks and did not need to rely on a few banks to onlend liquidity received from the central bank. All this, I would say, was the legacy of the start of the euro. It permitted us to go through the full period without changing our operational framework. Of course, we continuously reviewed this framework, as we have done in the past before the turmoil as well. Again, I believe that the diversity of the origin of our operational framework, due to the fact we had to merge a large number of traditions and a large number of experiences, proved very valuable. That being said, we have exactly the same problems as all other central banks. We still are in a market correction. For a long time, I hesitated to mention the word “crisis” myself and preferred to label it “a market correction of great magnitude with episodes of a high level of

08 Book.indb 655

2/13/09 3:59:23 PM

656

Chair: Stanley Fischer

volatility and turbulences.” I remained with this characterization until, I would say, Bear Stearns. Now I am prepared to speak of a crisis. Let me conclude by saying how useful I find interactions like this one. We need a lot of collegial wisdom to continue to handle the situation, and I will count on our continuous exchange of experiences and views. Mr. Sinai: Of the many, many points in this interesting paper, there are two I want to comment on. One is in support of Professor Buiter, and the other is not. One is on core inflation, and the other is on the asset bubbles and whether central banks should intervene earlier. On this last one, I don’t really see how the consequences of asset bubbles are in current existing policy approaches, looking back over the last few bubbles we have had, either in the policy framework at the time and policy rates or in financial markets. For example, the U.S. housing boom-bust cycle and housing priceasset bubble bursting. It is a bust and I would argue that we are in the midst, and still are, of an asset-price bubble bursting. We also have a credit and debt bubble, and those prices and those securities that represent that have been bursting and declining as well. We see that all around us all the time. I don’t think that was in the approaches of any central bank a year or two ago—the consequences of what we see today and of what is showing up in terms of the impacts. Similarly so, the dot-com stock market bubble’s bursting—and some people call the general U.S. stock market bubble bursting in 2000-01—that wasn’t in the existing approach to monetary policy, and its consequences surely affected the future distribution of outcomes. For an issue of not leaning against the wind and not acting preemptively in an insipient bubble, these two examples in the recent history convince me we ought to seriously consider alternatives to waiting, to waiting until after a bubble bursts—that is, what you call the Greenspan-Bernanke way—and what I just heard Rick Mishkin continue to support. Of the choices available, there is a lot to be

08 Book.indb 656

2/13/09 3:59:23 PM

General Discussion

657

said for finding methods to intervene earlier when you have insipient bubbles. That would be true for all central banks, and we have an awful lot of them. There was sentiment here last year, I think Jacob Frenkel and Stan Fischer, increasing sentiment in the central bank community to think about intervening before a bubble bursts. So, I don’t agree with you at all on that one. On the issue of core inflation, I really do agree with you in terms of central banks and what they should focus on. The case of the U.S. core versus headline inflation rate is an example. Core inflation in the United States provides the lowest possible reading on inflation of all possible readings—that is the core consumption deflator. If you follow that one you are going to get the lowest reading on inflation of all the possible measures that exist on inflation. This means that you are going to run a lower interest rate regime, if you focus on that as the key inflation barometer. We did run a very low interest rate regime based on that for quite a long time, and we see the consequences of that today in what’s going on in the highly leveraged events off the housing boom and bust. Second, Alan, you showed us three charts. The third one, to me, is the most relevant because crude oil prices on average have been rising now for seven years, so it’s hardly a temporary spike or a transitory spike. I think we would all agree it’s part of a global demand-supply situation. Finally, in taking those charts and making conclusions that core inflation will be a good predictor of headline inflation may have been true in the past, but given the changed structure of inflation and the global component of it this time, the econometrics of the backwardlooking approach that is implicit in looking at those charts and drawing conclusions are subject to some concern. Mr. Hatzius: I’d like to address Willem’s assertion the Fed eased far too much, given the inflation risks. From a forward-looking perspective, which I think is the perspective that matters, the Fed’s influence on inflation primarily works via its ability to generate slack in the economy. Even with the 325 basis points of cumulative easing, the economy is already generating very significant amounts of slack and

08 Book.indb 657

2/13/09 3:59:23 PM

658

Chair: Stanley Fischer

that is most clearly visible in the increase in the unemployment rate, from 4.4 percent early last year to 5.7 percent now. Most forecasters expect the unemployment rate to increase further to somewhere between 6 and 7 percent over the next six to 12 months. That would resemble the levels we saw at the end of the last two recessions. In other words, we are already generating the very disinflationary forces that higher interest rates are supposed to generate, despite 325 basis points of monetary easing. My question to Willem is, What is wrong with that analysis in your view? Is it that you disagree with the basic view of how Fed policy affects inflation—namely, by generating slack? Or is it that you think the sustainable level of output has fallen so sharply that a 6 to 7 percent unemployment rate will be insufficient to combat inflationary pressures? Or is it that you think these expectations of a 6 or 7 percent unemployment rate are simply wrong and the economy is going to bounce back in a fairly major way? Mr. Harris: I wanted to underscore the idea that we can’t make this simple comparison between European and U.S. monetary policy—Willem said in the paper there are rather similar circumstances in Europe and the United States. However, the U.S. economy has gone into this downturn much faster than Europe. The shocks to the U.S. economy are greater. We know the economy would have been in even worse shape if the Fed hadn’t eased interest rates, and we also know it is not over. It is not over in the United States, and it is not over in Europe. It may turn out that what happens is that Europe just lags the Fed in terms of rate cuts going forward. I don’t understand the idea there are rather similar circumstances in Europe and the United States. I have the same question as Jan Hatzius. With the unemployment rate headed well above 6 percent, what level of the unemployment rate would restore the Fed’s credibility here? Mr. Kashyap: There is a sentence in your paper I encountered on the airplane, so I did not have the Internet to check this. It says, “Ben Bernanke, Don Kohn, Frederic Mishkin, Randall Kroszner,

08 Book.indb 658

2/13/09 3:59:23 PM

General Discussion

659

and Charles Plosser all have made statements to the effect that credit, mortgage equity withdrawal, or collateral channel through which house prices affect consumer demand is on top of the normal (pure) wealth effect.” I don’t remember all of those speeches, but I have read the Mishkin one pretty recently. There is a long passage in it directly contradicting this statement. If you are going to have these really tough comments, you need to have footnotes where you quote them verbatim. You can’t say he essentially said this. For instance, in Rick’s paper there are a couple of pages where he has this analogy that going to the ATM may Granger-cause spending even if it is only an intermediate step between your income and spending. In the same way, mortgage equity withdrawal may only be an intermediate step between greater household wealth and higher consumer spending. Maybe there is some other part of his story that I forgot, but it just doesn’t seem to be fair because this is Fed publication and people will assume that it must have been fact checked—I doubt people are going to go back to read the speeches themselves. If you are going to say something like that, given you are already at 140 pages, what’s the cost of going 170 pages and documenting it so that we could see? [Note: Following the symposium, the author added an extended footnote as requested by Professor Kashyap.] Mr. Muehring: The panel certainly lived up to its billing. I particularly wanted to note Mr. Yamaguchi’s heartfelt commentary, which was something to think about on the way home. I wanted to ask a question that goes to the one theme that seems to run throughout this conference, namely, is the central role of assetbased repo financing in the current crisis that Peter Fisher mentioned? It was also in the Shin paper, and several of the others, and can be seen in the liquidity hoarding by banks, who wouldn’t accept somebody else’s collateral and vice versa and thus this central critical importance of the haircuts in this crisis. One is to ask, so, one, do the panelists think there is a way to restrain the leverage generated through the repo financing during the upswing? And, two, if they could make just a general comment on

08 Book.indb 659

2/13/09 3:59:23 PM

660

Chair: Stanley Fischer

the merits of the various term facilities the central banks—the Fed in particular—have created, do they see limits in what can be achieved through the term liquidity facilities and how do they envision the future place of the facilities if the central banks are required to be market makers of last resort going forward? Mr. Weber: I only have a comment on one section of the paper, which deals with the collateral framework: Willem and I have discussed this in the past. He appears to have the misperception that the price or value of an illiquid asset is zero. This is why he believes that there is a subsidy implied in our collateral framework. But here are the facts. We value illiquid assets at transaction prices, and it could be the price of a distressed sales or a value taken from indices, such as the ABX index that was discussed yesterday. In addition, we then take a haircut from that price and we are in the legal position to issue margin calls and ask for a submission of additional collateral to cover the value of the repo. In the euro system, for example, the Bundesbank has banks pledge a pool of assets to the repo window, which is usually used between 10 to 50 percent. To cover the value of the outstanding repos, the entire pool is pledged to the central bank, and we can seize all that collateral to cover the amount due. Thus, I disagree with the statement that there is an implicit subsidy implied because the repo is well-covered due to these institutional provisions. Let me make a second point. If you have a pool of collateral pledged and the use of that pool moves between 10 to 50 percent in normal times to a much higher use of collateral, it is a very good indication that banks need more backup liquidity, in the sense of central bank liquidity, and the bank may be in distress. Thus, the endogenous increase in the percentage use of the pool for us is a very good early indicator of potential liquidity problems of that bank in refinancing in the market because, as a consequence, it switches from market liquidity to repo liquidity. To sum up, Willem, some of the allegations you make do not really hold up.

08 Book.indb 660

2/13/09 3:59:23 PM

General Discussion

661

Mr. Buiter: First of all, I want to address the culturally sensitive issue—which is the little boy with his finger in the dike. That story was, of course, written by an American. No Dutchman would have written it because it is based on a wrong model. That hole in the dike that you can plug with your finger, you can leave alone quietly. It will not cause a flood. There was no threat. It is also good to know that, despite the length of the paper, some people want to lengthen it. All I can say is, it’s only this long because I didn’t have time to write a shorter paper. Very briefly, my point is not that circumstances weren’t unusual and exigent and difficult for central banks, but even at the time the choices were made there was knowledge and other choices that could have been made. They are options available that would have been superior to the methods chosen. One of them obviously is the way in which—take the Fed as an example—the PDCF and TCLF securities are priced. That is just crazy. You don’t let borrowers (or the agent of the borrowers) determine the value of the collateral they offer you especially if it is illiquid. There are other options. In the case of Bear Stearns, one wonders why exigent and unusual circumstances weren’t invoked to allow it to borrow directly at the discount window. There are options that were open. In the case of the Bank of England, of course, the list of why did they wait so long, for the first few months when there was no lender of last resort. The facility accepts ad hoc ones when there turned out to be no deposit insurance worth anything and there was no insolvency regime for banks. It is quite extraordinary. So there were options that should have been used at the time. On risk management: I fully agree with Alan. You don’t need risk management, or whatever it is, to justify cutting rates. However, risk management was used to provide justification for cutting rates and especially the nonlinearities’ irreversibility soft or light version of risk management. We all have our nonlinearities. You can put it at zero for the normal interest rates. Gross investment can’t be negative either. But that is not a nonlinearity. That bias goes the other way.

08 Book.indb 661

2/13/09 3:59:23 PM

662

Chair: Stanley Fischer

So the notion that plausible systemically important nonlinearities would create a bias in favor of putting extraordinary weight on preventing a shock collapse of output rather than safeguarding it against high and rising inflation is not at all obvious to me. If the arguments aren’t really strong, one shouldn’t arbitrage the words from serious science into social science. Alan selectively ended his quote on the core inflation at a point it would have contradicted what he said: “Core inflation is relevant to the price stability leg of the Fed’s mandate to the extent that it is a superior predictor of future headline inflation, over the horizon the Fed can influence headline inflation.” And then it goes on: “a better predictor not only than headline inflation itself, but than any readily available set of predictors.” So whether or not core inflation is a better predictor of headline inflation, headline inflation itself is neither here nor there. It’s the best or necessary condition for being relevant, not as a sufficient condition. That anybody should use univariant predictors for future inflation to formulate policy is a mystery to me. So, I just find that framework doesn’t make any sense. On core inflation, the key message is to statisticians especially: “Get a life!” Get away from the monitor. Get away from the keyboard. Open the window. See whether there might be a structural break in the global economy that is not in the data—2.5 billion Chinese and Indians entering the world economy systematically raising the relative price of non-core goods and services to core goods and services is not something that has been happening on a regular basis in samples that are at our disposal. You have to be very creative and intelligent, not bound by whatever time series your research assistant happens to have loaded into your machine. Can the central bank get timely information about liquidity and solvency of individual institutions without being supervisor and regulator? That is a key question. If there is a way of getting the information, without the regulatory and supervisory powers, which make

08 Book.indb 662

2/13/09 3:59:23 PM

General Discussion

663

an interesting subject for capture, then we are in the game. In the United Kingdom—it was supposed to work this way with the Bank of England—tagging along was the FSA. It didn’t work. There are institutional obstacles to the free, unconstrained, and timely flow of relative information. So, this is a deep problem. I would think, if the central bank were not subject to capture, then I would prefer the interest rate decision be with the central bank. It is only when the central bank has to perform market maker and lender-of-last-resort functions is there is a serious risk of the official policy rate being captured, as I think it was in the U.S. That would be reason for moving it out. It is the second-best argument of institutional design. On the quotes, I cited all the papers that I quoted. They are in there. I have the individual quotes, if you want them. I can certainly put them in, but especially your representation of the Mishkin paper, which I assume is the Mishkin paper I cited at length in the paper, is a total misrepresentation of that paper. There are two sets of simulations. One is just a regular wealth effect and the other is part of the wealth effect or financial assets. It is doubled to allow for a credit channel effect. There is very clearly in that particular paper a liquidity effect, a credit channel, or collateral effect on top of the standard wealth effect. I will append the paper, if that makes you happy. Mr. Blinder: I wanted to square something Jean-Claude Trichet said and then just react to a couple of questions. The legal mandates of the ECB and the Federal Reserve are different. It follows from that, that even if the circumstances were identical, you would expect different decisions out of the ECB governing counsel and the Federal Reserve. I wanted to underscore that. Secondly, about the little Dutch boy: Willem is correct. It is an American tale, but I can tell him that, if I ever see a leak in the Lincoln Tunnel, I call the cops. John Makin asked if it was the Fed or whoever was supposed to put the finger in the dike. Yes, it was the Fed because the Fed can and did act fast. Waiting for help? Yes, the Fed could have used more

08 Book.indb 663

2/13/09 3:59:23 PM

664

Chair: Stanley Fischer

help from the U.S. Treasury, for example, and over a longer time lag from the U.S. Congress, which it is going to get—grudgingly and slowly—and I might say from the industry. Let’s leave it at that. John asked the question, If there were a recession, then what would happen? If I can paraphrase Andrew Mellon, this is my answer. Liquefy labor, liquefy stocks, liquefy the farmers, liquefy real estate. It will purge the recession out of the system. People will have work and live a better life. Finally, on core versus headline inflation: I really want to disagree with Allen Sinai and implicitly again with Willem. At the end of this, I am going to propose a bet with 150 witnesses. Core inflation is only below headline inflation when energy is rising fast. When energy is rising slowly, it is above. Over very long periods of time, there is no trend difference between the two. Now there was between 2002 and 2008, I think. It looks like it’s over, but who really knows if it’s over? But I do want to cite the theorem that no relative price can go to infinity. So, we know Chart 3 that I sketched just can’t go on forever, no matter whether there is China, India, or what. It just cannot happen. The concrete bet that I would propose to either Willem or Allen is that over the next 12 months—and you can pick the inflation rate (I don’t care if it’s PC or CPI)—the headline will be below the core. If you’ll give me even odds on that, I’ll put up $100 against each of you.

08 Book.indb 664

2/13/09 3:59:23 PM

Concluding Remarks Stanley Fischer

When we met at this conference a year ago the financial crisis was just beginning and it was far from clear how serious it would be. By now, it is generally described as the worst financial crisis in the United States since World War II, which is to say, since the Great Depression. Further, as Chairman Bernanke told us in his opening address, the financial storm is still with us, and its ultimate impact is not yet known. As usual, the Kansas City Fed has put together an excellent and timely program, both in the choice of topics and authors, and also in the choice of discussants. Before getting to the substance of the discussions of the last two days, I would like to make a number of preliminary points. First, although this is widely described as the worst financial crisis since World War II, the real economy in the United States is still growing, albeit at a modest rate.1 The disconnect between the seriousness of the financial crisis and the impact—so far—on the real economy is striking. At least three possibilities suggest themselves: first, the worst of the real effects may yet lie ahead; second, the vigorous policy responses, both monetary and fiscal, may well have had an impact; and third, perhaps, that although all of us here are inclined to believe the financial system plays a critical role in the economy, 665

08 Book.indb 665

2/13/09 3:59:23 PM

666

Stanley Fischer

that may not have been true of some of the financial innovations of recent years, a point that was made by Willem Buiter. Second, the losses from this crisis, as a share of GDP, to the financial system and the government are likely to be small relative to those suffered by some of the Asian countries during the 1990s.2 That may make it clearer why those crises have left such a deep impact on the affected countries. Third, about warnings of the crisis: At policy-related conferences in recent years, the most commonly discussed potential economic crisis related to the unwinding of the U.S. current account deficit. That crisis scenario was based on the unsustainability of the U.S. current account deficit and the corresponding surpluses of China and other Asian countries, and more recently also of the oil-producing countries. In such scenarios, the potential crisis would have come about had the dollar decline needed to restore equilibrium become disorderly or rapid, creating inflationary forces that the Fed would have to counteract by raising its interest rate. But there were also those who described a scenario based on a financial sector crisis resulting from the reversal of the excessively low risk premia that prevailed in 2006 and 2007, and in the case of the United States and a few other countries from the collapse of the housing price bubble. Among those warning about all or parts of this scenario were the BIS, with Chief Economist Bill White and his colleagues taking the lead, Nouriel Roubini, Bob Shiller, Martin Feldstein, the late Ned Gramlich, Bill Rhodes, and Stephen Roach. As in the case of most crises and intelligence failures, the question was not why the crisis was not foreseen, but why warnings were not taken sufficiently seriously by the authorities—and, I should add, the bulk of policy economists. In his opening address, Chairman Bernanke noted the Fed’s three lines of response to the crisis: sharp reductions in the interest rate; liquidity support; and a range of activities in its role as financial regulator. In his lunchtime speech yesterday, Mario Draghi, Governor of the Banca d’Italia, mentioned briefly the six areas on which the Financial Stability Forum’s report, published in April, focuses. They

08 Book.indb 666

2/13/09 3:59:23 PM

Concluding Remarks

667

are: capital requirements; liquidity; risk management; transparency; credit rating agencies; and asset valuation (including the difficult and tendentious topic of mark-to-market accounting). All these topics received attention during the conference, and all of them are of course receiving attention from the authorities as they deal with the crisis, and begin to institute reforms intended to reduce the extent and frequency of similar crises in the future. Rather than try to take up these topics one-by-one, it is easier to describe the conference by focusing on three broad questions, similar but not identical to those raised in the paper by Charles Calomiris: • What are the origins of the crisis? • What is likely to happen next, in the short run of a year or two, and when will growth return to potential? • What structural changes should and are likely to be implemented to prevent the recurrence of a similar crisis, and to significantly reduce the frequency of financial crises in the advanced countries? A fourth topic, the evaluation of central bank behavior in this crisis, was implicit in the discussion in much of the conference and explicit in the last paper of the conference, by Willem Buiter. I.

The Origins of the Crisis

The immediate causes of the financial crisis were an irrationally exuberant credit boom combined with financial engineering that (i) led to the creation of and reliance on complex financial instruments whose risk characteristics were either underestimated or not understood, and (ii) fueled a housing boom that became a housing price bubble, and (iii) led to a worldwide and unsustainable compression of risk premia. The bursting of the U.S. housing price bubble and the beginnings of the restoration of more normal risk premia set off a downward spiral in which a range of complex financial instruments rapidly lost value, causing difficulties for leading financial institutions and for the real economy. These developments gradually brought the Fed and the major central banks of Europe into action as providers

08 Book.indb 667

2/13/09 3:59:23 PM

668

Stanley Fischer

of liquidity to imploding financial institutions and markets, and later led to lender-of-last-resort type interventions to restructure and/or save financial institutions in deep trouble. It has become conventional to blame a too easy monetary policy in the U.S. during the years 2004-2007 for the excessive global liquidity, but this issue was not much mentioned during the conference. The Fed may have taken a long time to raise the discount rate from its one percent level in June 2003 until it reached 5.25 percent three years later. But it should be remembered that the concern over deflation in 2003 was both real and justified. More important in the development of the bubble in the housing market was the availability of financing that required very little— if any—cash down and provided low teaser rates on adjustable rate mortgages. As is well known, the system worked well as long as housing prices were rising and mortgages could be refinanced every few years. The fact that the housing finance system developed in this way reflects a major failure of regulation, a result in part of the absence of uniform regulation of mortgages in the United States, and in some parts of the system, the absence of practically any regulation of mortgage issuers. This was and is no small failure, whose correction is widely seen as one of the most pressing areas of reform needed as the U.S. financial regulatory system is restructured. The first line of defense for the financial system should be internal risk management in banks and other financial institutions. These systems also failed, and their failure is even more worrisome than the failure of the regulators—for after all, it is very difficult to expect regulators, with their limited resources and inherent limits on how much they can master the details of each institution’s risk exposure, to do better than internal risk management in fully understanding the risks facing an institution. Based on my limited personal experience—that is to say on just one data point—I do not believe the risk managers were technically deficient. Rather their ability to envisage extreme market conditions, such as those that emerged in the last year in which some sources of financing simply disappeared, was limited. Perhaps that is why we seem to have perfect storms, once in a century events, so regularly.

08 Book.indb 668

2/13/09 3:59:24 PM

Concluding Remarks

669

There is a delicate point here. If risk managers are required to assign high probabilities to extreme scenarios, such as those of the last year, the volume of lending and risk-taking more generally might be seriously and dangerously reduced. Thus it is neither wise nor efficient for the management of financial firms or their regulators to require financial institutions to become excessively risk averse in their lending. But if these institutions pay too little attention to adverse events that have a reasonable probability of occurring, they contribute to excesses of volatility and crises. The hope is that despite the moral hazard that will be enhanced by the authorities’ justified reactions in this crisis, there is a rational expectations equilibrium that ensures a financial system that is both stable and less crisis prone— even though we all know we will not be able entirely to eliminate financial crises. As Tobias Adrian and Hyun Song Shin stated in their paper, this is the first post-securitization financial crisis. With so much of the financial distress related to securitization, the “originate to distribute” model of mortgage finance has come under close scrutiny. Views are divided. Some see the loss of the incentive to scrutinize mortgages (or whatever assets are being securitized) closely as a major factor in the crisis, suggesting that the crisis would not have been so severe had the originators of the mortgages expected to hold them to maturity. This is clearly true. Others pointed out that securitization has been very successful in other areas, especially the securitization of credit card receivables, and that it would be a mistake to reform the system in ways that make it harder to continue the successful forms of securitization—another view that has merit. A few years ago Warren Buffett described derivatives as financial weapons of mass destruction, at the same time as Alan Greenspan explained that new developments in the financial system, including ever-more sophisticated derivatives and securitization, enabled a better allocation of risks. It seems clear that in this crisis financial engineers invented instruments that were too sophisticated—at this point it is obligatory to refer to “CDOs squared”—for both their own risk managers and their customers to understand fully, and that this is part of the explanation for the depth and complexity of the

08 Book.indb 669

2/13/09 3:59:24 PM

670

Stanley Fischer

crisis. That is to say that the Buffett view is a better guide to the role of financial super-sophistication, at least in this crisis. But as with securitization, it would be a mistake to overreact and try to regulate extremely useful techniques out of existence. The role of the rating agencies in this crisis has received a great deal of criticism, including in this conference. However, in considering reforms of the system, we should focus on the particular conflicts of interest that the rating agencies faced in rating the complex financial instruments whose nature was not well understood by many who bought them, and try to deal with those conflicts, while recognizing that external ratings by an independent agency will continue to be necessary for risk management purposes despite all the difficulties associated with that fact. Let me turn now to leverage and liquidity, the latter the topic of the paper by Franklin Allen and Elena Carletti. It has repeatedly been said that this crisis was in large measure due to financial firms becoming excessively leveraged. This must have been said in one way or another about every financial crisis for centuries—and it was certainly said during the financial crises of the 1990s, including the LTCM crisis. Most financial institutions, notably including banks, make a living off leverage. Nonetheless, there should be leverage constraints —required capital ratios—for any financial institutions that receive or are likely to receive protection from the public sector. Of course, one element of the regulatory game is that regulators impose regulations and the private sector seeks ways around them. So regulators have to be on their toes. Perhaps the worst breach in the regulation of bank leverage comes from the existence of off-balance sheet financing. There is no good reason to permit off-balance sheet financing, particularly when, as in the current crisis, items that many thought were off-balance sheet return to the balance sheet when they become problematic. Liquidity shortages have been a central feature of this crisis, but that too is typically the case in financial crises. In their paper Allen and Carletti focus on the role of liquidity—particularly the hoarding of liquidity—in explaining several features of market behavior during the

08 Book.indb 670

2/13/09 3:59:24 PM

Concluding Remarks

671

crisis: the phenomenon that the prices of many AAA-rated tranches of securitized products other than subprime mortgages fell; that interbank markets for even relatively short-term maturities dried up; and the fear of contagion. There is little doubt that required liquidity ratios will be imposed on financial institutions following this crisis, but it also has to be recognized that instruments that appear liquid during good times become illiquid during crises. Thus few instruments other than shortterm government paper should be eligible as liquid for purposes of the liquidity ratio. Several speakers and discussants raised the issue of compensation systems for traders and managers in the financial system. There is little doubt that the heads I win, tails you lose, nature of bonus payments contributes to excessive risk taking by traders. It remains to be seen whether it will be possible to change the compensation system to provide incentives that will more closely align private and social benefits and costs. II.

What Next?

As Chairman Bernanke noted in his opening remarks, the financial crisis is not yet over. At the time of the conference the most immediate problem on the agenda was the future of the GSEs, particularly Fannie Mae and Freddie Mac. As the financial crisis has deepened, as the housing market has deteriorated and housing prices have fallen, and as risk aversion has increased, the situation of these two massive housing sector financial institutions has worsened, to the point where the widespread belief that the government would stand behind them if they ever got into trouble was essentially confirmed by the authorities in July. Because of a lack of clarity of the plan announced in July, the U.S. Treasury issued a more far-reaching plan in the first half of September. The two GSEs had become too big to fail, not only because of their role in the U.S. housing market, not only because of their political power in Washington, but also because their bonds constituted a significant share of the reserves of China, Japan and other countries.

08 Book.indb 671

2/13/09 3:59:24 PM

672

Stanley Fischer

A default on the liabilities of the GSEs would have had a major immediate impact on the exchange rate of the dollar, and long-lasting effects on market confidence in the dollar and its role as a reserve currency, and those were risks that the U.S. authorities rightly were not willing to take. The GSE rescues in July and September followed the Bear Stearns intervention in March, and raised the question of what more it would take to stabilize the U.S. financial system, as well as the financial systems of Switzerland and the U.K., and possibly other countries. The special liquidity operations of the major central banks are part of the answer. Beyond that, there were suggestions to give more help to mortgage borrowers who now have negative equity in their houses. And more than one speaker referred to the need for a new Reconstruction Finance Corporation, without specifying what such an organization would be expected to do—probably if established it would be expected to help recapitalize the financial system. Capital raising by stressed financial institutions is another component, though several speakers expressed doubts about the banks’ capacity to raise capital at an affordable price at this time. Anil Kashyap, Raghu Rajan and Jeremy Stein suggested a scheme whereby banks would buy insurance that would provide capital in downturns or crises, with the insurance policy being one that makes a given amount of capital available to a bank in a well-defined event in which the overall condition of the banking system—for moral hazard reasons, not the condition of the bank itself—deteriorates. This is an interesting proposal, whose institutional details need to be worked out, but it is probably not relevant to the resolution of the current crisis. The end of the housing price bubble and its impact on the financial system marked the start of the financial crisis, and the contraction of house-building activity was the main factor reducing the growth rate of the economy as the financial sector difficulties mounted. Martin Feldstein in his introductory remarks suggested that U.S. house prices still have 10-15 percent to fall to reach their equilibrium level, but that they may well overshoot on the downside, and thus prolong the crisis. He emphasized the negative effect of the decline in housing wealth on consumption and aggregate demand. Willem

08 Book.indb 672

2/13/09 3:59:24 PM

Concluding Remarks

673

Buiter argued that to a first approximation there is no wealth effect from a rise or decline in the price of housing for people who expect to continue to live in their house—or to put the issue another way, that the perfect hedge against a change in the cost of housing is to own a house. Nonetheless Buiter agreed that the availability of financing based on the owners’ equity in the house would have an effect on aggregate demand. A year after the start of the crisis, with the financial situation not yet stabilized, many ventured guesses as to how severe the downturn would be and how long it would continue. There seemed to be near unanimity that the recovery would not begin this year, and a majority view that growth in the U.S. would resume after mid-year 2009. The dynamics of recovery are complicated, for so long as the financial system continues to deteriorate, it will negatively affect the real economy, and the real economic deterioration in turn will have a negative effect on the financial crisis. That is why some conference participants believed that recovery in the U.S. would not take place until 2010. III.

Longer-term Reforms

The agenda for longer-term reform of the financial system to reduce the frequency and intensity of financial crises was laid out in the speech by Mario Draghi, which drew on the excellent report of the Financial Stability Forum which he chairs, published in April.3 Several other noteworthy reports, including the Treasury’s report on the reorganization of financial sector supervision in the United States,4 two reports by the private sector Countercyclical Risk Management group, headed by Gerry Corrigan,5 and the report of the IIF, the Institute of International Finance,6 have also been published in the last several months. The reform agenda suggested by the Financial Stability Forum has already been described, to reform capital requirements; liquidity; risk management; transparency; credit rating agencies; and asset valuation. In presenting a summary of the FSF Report, Mario Draghi emphasized the role that poor risk management, fueled by inappropriate incentives, had played in generating the crisis. He argued

08 Book.indb 673

2/13/09 3:59:24 PM

674

Stanley Fischer

that the strengthening and implementation of the Basel II approach would significantly align capital requirements with banks’ risks. He also discussed ways of reducing the pro-cyclicality of the behavior of the banking system, and the need in formulating monetary policy to take account of financial sector developments—the latter a point developed in the persuasive paper by Adrian and Shin. The reports of the Counterparty Risk Management Group have presented a set of recommendations to improve the plumbing of the financial system, particularly in trading and dealing with sophisticated and by their nature closely interlinked derivative contracts. Among the recommendations are to attempt to move more contracts to organized markets, and to impose some form of regulation. Further, in light of the huge volume of outstanding derivative contracts, the unwinding of a major financial company is bound to be extremely difficult and costly, despite the existence of netting contracts that in principle could make that process much less difficult. Hence there can be little doubt about the need for further work on market infrastructure. In addition, this crisis has led to a rethinking of the structure of financial market regulation, centered on the role of the central bank in regulation. The apparent failure of coordination in the United Kingdom among the Treasury, the Bank of England, and the FSA in dealing with the Northern Rock case at a time when the central bank was called upon to act as lender of last resort, has led to a reexamination of the FSA model, that of a single independent regulator over the entire financial system, separate from and independent of the central bank. The Fed’s role in the rescue of Bear Stearns, and the apparent extension of the lender of last resort safety net to investment banks has led many to argue that the Fed should supervise all financial institutions for whom it might act as lender of last resort­—and the Fed has already reached an agreement with the SEC on cooperation in supervising the major investment banks, which have not until now been under the Fed’s supervision.7 Historically supervision has been structured along sectoral lines—a supervisor of the banks, a supervisor of the insurance companies, and so forth. More recently the approach has been functional, in particular distinguishing between prudential and conduct-of-business supervision.

08 Book.indb 674

2/13/09 3:59:24 PM

Concluding Remarks

675

In the twin-peaks Dutch model, prudential supervision of the entire financial system is located in the central bank, and conduct of business supervision in a separate organization, outside the central bank. In the Irish model, both functions are located in the central bank.8 In Australia, prudential and conduct-of-business supervision are located in separate organizations, both separate from the central bank. As is well known, in the UK the FSA—the Financial Services Authority—is responsible for supervision of the entire financial system, and is located outside the central bank. Sometimes a third function is added—that of supervision of (stability of) the entire financial system, a responsibility that is typically assigned to the central bank.9 It is absolutely certain that the structure of supervisory systems will be revisited as a result of this crisis. One conclusion, I strongly believe, will be that prudential supervision should be located within the central bank. Another issue that will be reexamined is the role of the lender of last resort, and how far the central bank’s safety net should extend. The analytic distinction between problems of liquidity and solvency is helpful in thinking through the role of the lender of last resort, but the judgment of whether an institution faces a liquidity or a solvency problem is rarely clear in the heat of the moment. Traditionally it has been thought that the central bank should operate as lender of last resort only for banks,10 but as the Bear Stearns case showed, the failure of other types of institutions may also have serious consequences for the stability of the financial system.11 And of course, the moral hazard issue has always to be borne in mind in discussing the depth and breadth of the security blanket provided by the lender of last resort. In the financial crises of the 1990s, particularly those in Asia in 1997-98, the IMF argued that countries could avoid financial crises by (i) ensuring that their macroeconomic framework was sound and sustainable, and (ii) that the financial system was strong. To what extent does the current crisis validate or contradict that conclusion? The macroeconomic situation of the United States in recent years has not been sustainable, in that the current account deficit clearly had to be corrected at some point; similarly longer-run budget projections point to the need for a substantial correction in future. This does not necessarily mean that the U.S. macroeconomic framework

08 Book.indb 675

2/13/09 3:59:24 PM

676

Stanley Fischer

was not sustainable. It is clear however that the financial system was not strong, and that in particular, the supervisory system was not a system, but a collection of separate and not well coordinated authorities, with substantial gaps and shortcomings in its coverage. The question of the connection between the unsustainability of the macroeconomic situation and the financial crisis remains a key question for research. IV.

Evaluating Policy Performance So Far

In his interesting and provocative paper, Willem Buiter criticizes, among other things, the Fed’s “rescue” of Bear Stearns, and its failure to control inflation.12 The Bear Stearns rescue still looks sensible, in light of the fragile state of the financial system when it took place, and in light of the fact that the existing owners were not protected but rather saw the value of their shares massively marked down. As to the inflation point, Buiter in part argues that the Fed was too slow in raising interest rates in the period 2003-2006, and in addition that it was obvious that the entry of Chinese and Indian producers and consumers into the world economy would be inflationary, and should have been anticipated by the Fed. With regard to the latter point, we should remember that until about a year ago the predominant view about the entry of China and India into the global economy, was that it was a deflationary force, pushing down on wages in the industrialized countries. Why the changed view? That must be a result of the overall balance of macroeconomic forces in the global economy, which switched from deflationary to inflationary as the rapid global growth of the last four years continued. It remains to be analyzed where the inflationary impulses were centered, and what role was played by China’s exchange rate policy. More generally, whether the ongoing integration of China and India into the global economy will lead to deflation or ongoing inflation as the relative prices of goods consumed directly or indirectly by them—middle class goods—rise will also be determined by the overall balance of global macroeconomic policy.

08 Book.indb 676

2/13/09 3:59:24 PM

Concluding Remarks

V.

677

Concluding Comment

Typically, the question the returning traveler is asked after attending an international conference as well known as this one is “Were they optimistic or pessimistic?” This time the answer for the short run of up to a year is obvious: “pessimistic.” But if the authorities in the U.S. and abroad move rapidly and well to stabilize the financial situation, growth could be beginning to resume by the time we meet here again next year.

Author’s note: This is an edited version of concluding comments delivered at the Federal Reserve Bank of Kansas City conference, “Maintaining Stability in a Changing Financial System,” Jackson Hole, Wyoming, August 21-23, 2008. In light of their importance, I have had to mention some of the financial developments that occurred after the Jackson Hole conference. However I have tried to minimize the use of hindsight in preparing the written version of the comments and have tried to keep them close to the concluding comments delivered on August 23, 2008.

08 Book.indb 677

2/13/09 3:59:24 PM

678

Stanley Fischer

Endnotes This comment was made before the upward revision (in late August, after the Kansas City Fed conference) of second quarter GDP. 1

Whether this statement turns out to be true depends on the ultimate cost to the public of the many rescue measures announced after the Jackson Hole conference. 2

3 “Report of the Financial Stability Forum on Enhancing Market and Institutional Resilience,” Financial Stability Forum, April 2008.

“The Department of the Treasury Blueprint for a Modernized Financial Regulatory Structure,” U.S. Treasury, March 2008. 4

“Containing Systemic Risk: The Road to Reform,” Counterparty Risk Management Policy Group III (CRMPG III), August 6, 2008. “Toward Greater Financial Stability: A Private Sector Perspective,” Counterparty Risk Management Policy Group II (CRMPG II), July 27, 2005. 5

“IIF Final Report of the Committee on Market Best Practices: Principles of Conduct and Best Practice Recommendations,” Institute of International Finance, July 2008. 6

This was written before the disappearance of the major investment banks in the U.S. 7

8 More accurately, the organization is known as the “Central Bank and Financial Services Authority of Ireland.”

The U.S. Treasury’s March 2008 report on the reform of the supervisory system, op. cit. adopts this tri-functional approach. 9

The current Bank of Israel law (passed in 1954) allows the central bank to lend only to banks. In cases of liquidity, the central bank can do that on its own authority; in solvency cases, it needs the approval of the government. 10

This point is reinforced by the Fed’s decision in September to extend a loan to AIG, to prevent its immediate collapse. 11

Alan Blinder’s discussion of Willem Buiter’s paper provides a more comprehensive analysis of the major points raised by Buiter. 12

08 Book.indb 678

2/13/09 3:59:24 PM

The Participants Tobias Adrian Senior Economist Federal Reserve Bank of New York

Jeannine Aversa Economics Writer Associated Press

Shamshad Akhtar Governor State Bank of Pakistan

Martin H. Barnes Managing Editor, The Bank Credit Analyst BCA Research

Lewis Alexander Chief Economist Citi Hamad Al-Sayari Governor Saudi Arabian Monetary Agency Sinan Alshabibi Governor Central Bank of Iraq David E. Altig Senior Vice President and Director of Research Federal Reserve Bank of Atlanta Shuhei Aoki General Manager for the Americas Bank of Japan

Charles Bean Deputy Governor Bank of England Steven Beckner Senior Correspondent Market News International C. Fred Bergsten Director Peterson Institute for International Economics Richard Berner Co-Head, Global Economics Morgan Stanley, Inc. Brian Blackstone Special Writer Dow Jones Newswires

679

08 Book.indb 679

2/13/09 3:59:24 PM

680

The Participants

Alan Bollard Governor Reserve Bank of New Zealand

José R. De Gregorio Governor Central Bank of Chile

Hendrik Brouwer Executive Director De Nederlandsche Bank

Servaas Deroose Director European Commission

James B. Bullard President and Chief Executive Officer Federal Reserve Bank of St. Louis

William C. Dudley Executive Vice President Federal Reserve Bank of New York

Maria Teodora Cardoso Member of the Board of Directors Bank of Portugal Mark Carney Governor Bank of Canada John Cassidy Staff Writer The New Yorker Luc Coene Deputy Governor National Bank of Belgium Lu Córdova Chief Executive Officer Corlund Industries Andrew Crockett President JPMorgan Chase International Paul DeBruce President DeBruce Grain, Inc.

08 Book.indb 680

Robert H. Dugger Managing Director Tudor Investment Corporation Elizabeth A. Duke Governor Board of Governors of the Federal Reserve System Charles L. Evans President and Chief Executive Officer Federal Reserve Bank of Chicago Mark Felsenthal Correspondent Reuters Miguel Fernández OrdÓÑez Governor Bank of Spain Camden R. Fine President and Chief Executive Officer Independent Community Bankers of America

2/13/09 3:59:24 PM

The Participants

Jacob A. Frenkel Vice Chairman American International Group, Inc.

Jan Hatzius Chief U.S. Economist Goldman Sachs & Company

Ingimundur Fridriksson Governor Central Bank of Iceland

Thomas M. Hoenig President and Chief Executive Officer Federal Reserve Bank of Kansas City

Timothy Geithner President and Chief Executive Officer Federal Reserve Bank of New York Svein Gjedrem Governor Norges Bank Austan Goolsbee Professor Graduate School of Business, University of Chicago Tony Grimes Director General Central Bank and Financial Services Authority of Ireland Krishna Guha Chief U.S. Economics Correspondent Financial Times Craig S. Hakkio Senior Vice President and Special Advisor on Economic Policy Federal Reserve Bank of Kansas City Ethan Harris Chief U.S. Economist Lehman Brothers, Inc.

08 Book.indb 681

681

Robert Glenn Hubbard Professor Graduate School of Business, Columbia University Greg Ip U.S. Economics Editor The Economist Neil Irwin National Economy Correspondent The Washington Post Radovan Jelasic Governor National Bank of Serbia hugo frey jensen Director Danmarks Nationalbank Thomas Jordan Member of the Governing Board Swiss National Bank Sue Kirchhoff Reporter USA Today Donald L. Kohn Vice Chairman Board of Governors of the Federal Reserve System

2/13/09 3:59:24 PM

682

George Kopits Member of the Monetary Council National Bank of Hungary Randall Kroszner Governor Board of Governors of the Federal Reserve System Jeffrey M. Lacker President and Chief Executive Officer Federal Reserve Bank of Richmond Jean-Pierre Landau Deputy Governor Bank of France Scott Lanman Reporter Bloomberg News Ju-Yeol Lee Deputy Governor The Bank of Korea Mickey D. Levy Chief Economist Bank of America Steven Liesman Senior Economics Reporter CNBC Erkki Liikanen Governor Bank of Finland Justin Yifu Lin Senior Vice President and Chief Economist The World Bank

08 Book.indb 682

The Participants

Lawrence Lindsey President and Chief Executive Officer The Lindsey Group Dennis P. Lockhart President and Chief Executive Officer Federal Reserve Bank of Atlanta Philip Lowe Assistant Governor Reserve Bank of Australia Brian Madigan Director, Division of Monetary Affairs Board of Governors of the Federal Reserve System John H. Makin Principal Caxton Associates, Inc. Philippa Malmgren Chairman Canonbury Group Limited Tito T. Mboweni Governor South African Reserve Bank Paul McCulley Managing Director Pacific Investment Management Company Henrique De Campos Meirelles Governor Central Bank of Brazil

2/13/09 3:59:24 PM

The Participants

Allan Meltzer University Professor of Political Economy Carnegie Mellon University Yves Mersch Governor Central Bank of Luxembourg Mário Mesquita Deputy Governor Central Bank of Brazil Loretta Mester Senior Vice President and Director of Research Federal Reserve Bank of Philadelphia Laurence H. Meyer Vice Chairman Macroeconomic Advisers Frederic S. Mishkin Governor Board of Governors of the Federal Reserve System Rakesh Mohan Deputy Governor Reserve Bank of India Michael H. Moskow Vice Chairman and Senior Fellow for the Global Economy The Chicago Council on Global Affairs Kevin Muehring Senior Managing Director Medley Global Advisors

08 Book.indb 683

683

Kiyohiko G. Nishimura Deputy Governor Bank of Japan Peter Orszag Director Congressional Budget Office Sandra Pianalto President and Chief Executive Officer Federal Reserve Bank of Cleveland Charles I. Plosser President and Chief Executive Officer Federal Reserve Bank of Philadelphia Diane M. Raley Vice President and Public Information Officer Federal Reserve Bank of Kansas City Robert H. Rasche Senior Vice President and Director of Research Federal Reserve Bank of St. Louis Sudeep Reddy Economics Reporter The Wall Street Journal Martin Redrado President Central Bank of Argentina Irma Rosenberg First Deputy Governor Sveriges Riksbank

2/13/09 3:59:24 PM

684

Harvey Rosenblum Executive Vice President and Director of Research Federal Reserve Bank of Dallas

The Participants

Michelle A. Smith Assistant to the Board and Division Director Board of Governors of the Federal Reserve System

Eric S. Rosengren President and Chief Executive Officer Federal Reserve Bank of Boston

Mark S. Sniderman Executive Vice President Federal Reserve Bank of Cleveland

Fabrizio Saccomanni Deputy Governor Bank of Italy

Gene Sperling Senior Fellow for Economic Studies Council on Foreign Relations

Klaus Schmidt-Hebbel Chief Economist Organisation for Economic Co-operation and Development

David J. Stockton Director, Division of Research and Statistics Board of Governors of the Federal Reserve System

Gordon H. Sellon, Jr. Senior Vice President and Director of Research Federal Reserve Bank of Kansas City Nathan Sheets Director, International Finance Division Board of Governors of the Federal Reserve System Allen Sinai Chief Global Economist Decision Economics, Inc. Slawomir Skrzypek President National Bank of Poland

Daniel Sullivan Senior Vice President and Director of Research Federal Reserve Bank of Chicago Lawrence H. Summers Managing Director D.E. Shaw Phillip L. Swagel Assistant Secretary for Economic Policy U.S. Treasury Department Josef Tosŏvský Chairman, Financial Stability Institute Bank for International Settlements Umayya S. Toukan Governor Central Bank of Jordan

08 Book.indb 684

2/13/09 3:59:24 PM

The Participants

Joseph Tracy Executive Vice President and Director of Research Federal Reserve Bank of New York Jean-Claude Trichet President European Central Bank ZdenĔk TŮma Governor Czech National Bank Gertrude Tumpel Gugerell Member of the Executive Board European Central Bank

685

Janet L. Yellen President and Chief Executive Officer Federal Reserve Bank of San Francisco Durmuş Yilmaz Governor Central Bank of the Republic of Turkey Peter Zoellner Executive Director Austrian National Bank

Louis Uchitelle Economics Writer The New York Times José Darío Uribe General Manager Bank of the Republic, Colombia Julio Velarde Governor Central Reserve Bank of Peru Kevin M. Warsh Governor Board of Governors of the Federal Reserve System Axel A. Weber President Deutsche Bundesbank John A. Weinberg Senior Vice President and Director of Research Federal Reserve Bank of Richmond

08 Book.indb 685

2/13/09 3:59:24 PM