BALANCING THE BANKS This page intentionally left blank BALANCING THE BANKS Global Lessons from the Financial Crisis MATHIAS DEWATRIPONT, JEAN-CHARLES ROCHET, AND JEAN TIROLE Translated by Keith Tribe Princeton University Press Princeton and Oxford Copyright © 2010 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540 In the United Kingdom: Princeton University Press, 6 Oxford Street, Woodstock, Oxfordshire OX20 1TW press. princeton. du All Rights Reserved Library of Congress Cataloging-in-Publication Data Dewatripont, M. (Mathias) Balancing the banks : global lessons from the ? nancial crisis / Mathias Dewatripont, Jean-Charles Rochet, and Jean Tirole ; translated by Keith Tribe. p. cm. Includes bibliographical references and index. ISBN 978-0-691-14523-5 (hbk. : alk. paper) 1. Banks and banking— Government policy. 2. Banks and banking—State supervision. 3. Global Financial Crisis, 2008–2009. 4. Financial crises—History—21st century. I. Rochet, Jean-Charles. II. Tirole, Jean.
III. Tribe, Keith. IV. Title. HG1573. D49 2010 332. 1—dc22 2009052389 British Library Cataloging-in-Publication Data is available This book has been composed in Sabon Printed on acid-free paper. ? Printed in the United States of America 10 9 8 7 6 5 4 3 2 1 Contents Acknowledgments chapter 1 Introduction—Mathias Dewatripont, Jean-Charles Rochet, and Jean Tirole Regulation in a Historical Perspective To Regulate or Not to Regulate? The Challenges Facing Prudential Regulation Building an Adaptive Regulatory System in a Global World Keeping a Balance ii 1 1 3 6 7 8 chapter 2 Lessons from the Crisis—Jean Tirole Part I: What Happened? Part II: How Should the Financial System Be Reformed? 10 11 47 chapter 3 The Future of Banking Regulation— Jean-Charles Rochet The Basel Accords The Breakdown of the Basel Prudential Regime The Necessary Reforms 78 78 86 100 chapter 4 The Treatment of Distressed Banks— Mathias Dewatripont and Jean-Charles Rochet Reforming Prudential Policy for Distressed Banks Macroeconomic and Systemic Considerations International Cooperation 107 110 118 122
References Index 131 137 This page intentionally left blank Acknowledgments Jean Tirole thanks Nicolas Chanut, Frederic Cherbonnier, Jacques Delpla, Mathias Dewatripont, Pierre-Olivier Gourinchas, Marc Ivaldi, Jean-Pierre Landau, Sylvie Matherat, Michel Pebereau, Jean-Charles Rochet, Philippe Trainar, and participants at various conferences and seminars for very useful discussions and comments. The author is also very grateful to Keith Tribe for translating chapter 2 from French, and to Richard Baggaley for very helpful suggestions.
Jean-Charles Rochet thanks Charles Calomiris, Patrick Honohan, Rafael Repullo, and all those who attended several seminars at the Banque de France, the Bank of Canada, and the Central Bank of Brazil, especially Sylvie Matherat and Eduardo Lundberg. He has also pro? ted from detailed comments made by Jean Tirole on a draft version of his chapter. Responsibility for the arguments advanced here lie solely with the author, of course. Mathias Dewatripont and Jean-Charles Rochet thank Janet Mitchell and Peter Praet. Chapter 2 by Jean Tirole was ? rst published as “Lecons d’une crise,” Toulouse School of Economics Notes no. 1 12/2008 (online at www. se-fr. eu/images/TSE/TSENotes/notes%281%29j . tirole%28pdf%29. pdf). Chapter 3 by Jean-Charles Rochet was ? rst published as “Le futur de la reglementation bancaire,” Toulouse School of Economics Notes no. 2 12/2008 (online at www. tse-fr. eu/images/ TSE/TSENotes/notes%282%29rochet%28pdf%29. pdf). Chapter 4 by Mathias Dewatripont and Jean-Charles Rochet, “The Treatment of Distressed Banks,” ? rst appeared as a chapter of an electronic book by CEPR and Vox, entitled Financial Regulation and Macroeconomic Stability: Key Issues for the G20 (edited by Mathias Dewatripont, Xavier Freixas, and Richard Portes; see http://www. oxeu. org/reports/G20_ebook. pdf). It was prepared for a meeting held in London on January 31, 2009. This page intentionally left blank BALANCING THE BANKS This page intentionally left blank CHAPTER 1 Introduction Mathias Dewatripont, Jean-Charles Rochet, and Jean Tirole The recent ? nancial crisis was a mix of “unique” and much more conventional events. This short book offers our perspective on what happened and especially on the lessons to be learned in order to avoid a repetition of this large-scale meltdown of ? nancial markets, industrial recession, and public de? cits.
Chapter 2 provides a diagnosis of what went wrong and discusses some key ? nancial regulation reforms. Chapter 3 takes a more detailed look at the ? aws in the prudential framework that was in place when the crisis erupted and at the required remedies, and chapter 4 focuses on the treatment of distressed banks, a key element of this prudential framework. This introduction takes a more general look at the rationale for and challenges of banking regulation. Regulation in a Historical Perspective What degree of regulation of the banking sector is appropriate has been a controversial question for almost a century.
The Great Depression, with its wave of bank failures triggered by bank runs, led in the 1930s to heavy-handed regulation, combining deposit insurance, interest-rate regulations, barriers to entry, restrictions on activities (compulsory specialization), and constraints on bank size. Although the succeeding decades witnessed a return to stability, the banking system gradually became perceived as inef? cient and poorly innovative. In order to encourage cost-cutting and innovation, and to induce banks to pass ef? ciency gains on to consumers, governments deregulated the banking industry and fostered competition from the 1970s on.
This trend was also the result of pressure from commercial banks, which were facing 2 • Chapter 1 competition from other less regulated ? nancial institutions (e. g. , money-market mutual funds and investment banks). Although details vary from country to country, the removal of interest-rate controls promoted competition at ? rst. In the turbulent macroeconomic environment of the 1970s and 1980s, though, this form of deregulation, together with an interest-rate maturity mismatch in a period of rising interest rates, resulted in the 1980s in a large-scale banking crisis in the United States (the savings and loan—S&L—crisis).
This crisis led to a mix of further deregulation and reregulation. On the one hand, diversi? cation of activities was allowed in order to reduce the specializationinduced fragility of the S&Ls. S&Ls had used short-term savings deposits to fund long-term, ? xed-rate mortgages, and were thereby exposed to yield-curve risk. On the other hand, in order to limit the exposure of deposit insurance funds, the regulation of solvency ratios became more stringent and intervention rules in case of violation of these ratios were strengthened. This emphasis on prudential regulation and the desire to harmonize country-speci? capital adequacy requirements led to the international standard embodied in the Basel system of regulation. New international regulations, including the 1988 accords, were intended to ensure a level playing ? eld in a world of increasing globalization of banking. Subsequent events made this attempt to establish a level playing ? eld, however imperfect in practice, seem prescient since large international banks have now become common in the United States, Europe, and Asia. This internationalization and the intensi? cation of competition among various marketplaces (e. g. between Wall Street and the City of London) led to a weakening of regulatory standards, fed by pressure from large banks, themselves facing competition from more lightly regulated ? nancial institutions. One can interpret the recent modi? cation of the Basel capital adequacy rules (Basel II), which allow large banks to reduce effective capital ratios if they can show that their risks are “limited,” as an outcome of lobbying by these banks. The assessment of risk under the new regulations comes from the banks themselves, through “internal models”—which represents a step toward self-regulation.
The complexity of these internal models can make it very hard for supervisors to verify what is being computed and raises con- Introduction • 3 cern, despite the requirement that those models be authorized by regulatory authorities. The trend toward weaker regulation also came from the inability of the system to cope with the pace of ? nancial innovation, itself fed by a desire to lower the amount of capital required by the regulatory agencies. Indeed, the growth of the shadow banking system, of securitization, and of structured products (backed by credit ratings that had been in? ted by the rating agencies) can be partly traced to this desire. The gradual lowering of regulatory standards predated the recent crisis. To be sure, other developments such as “irrational exuberance,” loose monetary policy, and global macroeconomic imbalances also contributed to the crisis. But underregulation or ineffective regulation is rightly blamed for playing a central role in the crisis. Not surprisingly, this has led to calls for a strengthening of regulations in a number of countries. It is worth pausing, however, to ask what the purpose and extent of regulation should be.
To Regulate or Not to Regulate? Banking is one of a handful of industries (others include insurance, ? nancial market making, and pension-fund investing) subject to prudential regulation on top of consumer protection. The focus of this book will be on the former, and more substantial and decisive, form of regulation. This is not to imply that insuf? cient consumer protection played no role in the recent crisis. Indeed, the crisis started with problems in subprime loans. Although these problems were small compared to the overall crisis that ensued, subprime lending was the release mechanism.
Subprime loans are associated with weak consumer protection regulation of banking products in the United States. Therefore the creation of an agency speci? cally dedicated to strengthening borrower protection in the United States is a welcome development. What is so unique about banks as to warrant industry-speci? c regulation? Banks ful? ll a speci? c role in the economy through their involvement in the payment/deposit system as well as in lending to households and ? rms (for a survey of models of banking, 4 • Chapter 1 see Freixas and Rochet 2008).
Although these activities are essential to the economy, they are no more essential than, say, cars or pharmaceuticals, sectors in which consumer protection regulation exists but not prudential regulation. In banking, by contrast, prudential regulation has been in place since the 1930s. One classical rationale for such regulation is the vulnerability of individual banks to depositor runs. When wholesale and uninsured retail depositors lose con? dence in a bank, their natural reaction is to withdraw their money from the bank as fast as possible.
Such bank runs stem from the banks’ transformation activity. Banks create liquidity by borrowing short and lending long. By allowing depositors to withdraw their money whenever they feel like it, banks are exposed to self-ful? lling rational panics: as shown by Diamond and Dybvig (1983), when one expects other depositors to run and thereby force the bank into costly asset liquidation, one’s dominant strategy becomes to run too. The regulator’s monitoring of the institution’s leverage (and now liquidity) positions is meant to reduce the frequency of such costly runs.
The recent crisis (as well as some previous episodes, such as the failure of the Long-Term Capital Management hedge fund) has shown that another, related rationale for subjecting the banking industry to prudential regulation could be that the failure of one bank can trigger the failure of other banks through interbank exposures or banking panics. Prudential regulation of banks—in the form of capital ratio requirements plus deposit insurance—is therefore warranted, especially for institutions that are large and interconnected and thus can generate domino effects.
In contrast with non? nancial ? rms, which are bound to bene? t when a competitor goes under, banks can be hurt both as creditors of the failed bank and also as victims of panics that follow a neighbor’s insolvency. Prudential regulation is therefore meant to protect the banking infrastructure, the ? nancial system that allows the economy to function smoothly. This warrants that speci? c attention be paid to large banks. Yet smaller and not necessarily interconnected banks, whose failure has no systemic consequences, are also subject to prudential regulation.
The main reason for this is that their debtholders are small and lacking in monitoring expertise. Deposit insurance is typically introduced in order to reduce the risk that depositors Introduction • 5 will behave erratically, but it further reduces depositors’ incentive to monitor banks. This rationale for prudential regulation—the lack of expertise and the wastefulness associated with monitoring of balance sheets by retail depositors—explains why prudential regulation is also observed for other institutions with small, dispersed debtholders such as insurance companies and pension funds.
Dewatripont and Tirole (1994) discuss in detail the speci? cs of these institutions and their differences from other ? nancial and non? nancial institutions that are much more lightly regulated. They formulate the representation hypothesis, according to which prudential regulation should aim at replicating the corporate governance of non? nancial ? rms, that is, at acting as a representative of the debtholders of banks, insurance companies, and pension funds. The ? nancial industry has recently substantially increased the magnitude of its “wholesale” liabilities, that is, liabilities held not by small depositors but by other ? ancial institutions. Does this mean that the case for regulation has been weakened? In fact not, because such liabilities, which are often short term and therefore subject to panics, create systemic problems of two sorts: (1) they imply risks for the institution’s insured depositors (a case in point is Northern Rock; see the discussion in chapter 3), and (2) even if the institution does not have formally insured deposits (as in the case of investment banks or hedge funds), its failure could create domino effects because of its high degree of interconnectedness with other ? ancial ? rms (as was the case, for example, with the investment bank Lehman Brothers). Consequently, the argument behind the representation hypothesis still holds: even if the debtholders of banks are neither small nor inexperienced, the fact that their deposits are short term means that when they expect trouble, running is the best strategy. The danger of a bank run for the banking system as a whole then typically prompts the authorities to support endangered institutions.
The expectation of this “too big to fail” or “too interconnected to fail” syndrome does prevent panics but it also makes the bank’s debtholders passive and creates the potential for excessive risk-taking, in turn implying the need for a debtholder representative to ensure discipline. The social cost of the Lehman Brothers bankruptcy has, if anything, reinforced this argument, since one can now safely expect big banking institutions to be rescued in case of ? nancial distress. 6 • Chapter 1 The Challenges Facing Prudential Regulation According to the representation hypothesis, regulation should mimic the role played by creditors in non? ancial ? rms. Since debt gives its owners the right to take control of their borrowers’ assets in bad times, regulators must take control of banks in bad times in order to limit the losses of depositors or of the deposit insurance fund. This, in turn, implies the necessity of (1) de? ning what “bad times” means, and (2) making sure that one can intervene in those circumstances. This is no easy task, even when trouble hits only a single institution; we discuss this case ? rst, and then turn to the more complicated case of multi-institution hardship prompted by negative macroeconomic shocks.
For a single institution, bad times are de? ned as times at which its capital falls below the regulatory solvency ratio, as de? ned by the Basel I and II international agreements. Such de? nitions, although increasingly complex over time, nonetheless yield only rough approximations of a bank’s riskiness; for example, they concentrate only on credit risk, and do not fully take into account portfolio risk. Moreover, even in “normal” times, it is a challenging task for the regulator to intervene early enough, given that there is always an “accounting lag” in the computation of solvency.
Moreover, this challenge is exacerbated by the fact that, in contrast to non? nancial ? rms, banks can take advantage of (explicit or implicit) deposit insurance and “hide” problems of insolvency by aggressively raising money through higher interest rates, a strategy that has been called “gambling for resurrection. ” “Market discipline” can to some extent be relied on to help provide early warning signals of a bank’s trouble. This can work, however, only if some of the bank’s debt is not explicitly or implicitly insured by the state (otherwise its risk premium is ero) or if it is privately insured, so that its insurance premium would re? ect market perceptions of its riskiness. Such market discipline can in fact precipitate a crisis by making it more expensive for a bank in trouble to remain insured, and it does not make public intervention in bad times less essential; put differently, market discipline can be only a complement to, not a substitute for, public intervention. Prompt intervention in an individual insolvency is not straight- Introduction • 7 orward, but it is even harder in the case of generalized insolvency resulting from a macroeconomic shock. Indeed, multiple factors make taking control of banks during a banking crisis much more complicated. First, banks can expect some sympathy from politicians when they argue that the responsibility is not theirs but instead that of poor macroeconomic conditions. Second, politicians may quickly be faced with a drained deposit insurance fund and be very reluctant to request money from taxpayers to cover the cost of intervention.
Third, competent staff from regulatory authorities are likely to be overwhelmed by the sudden magnitude of the task of overseeing multiple interconnected distressed ? nancial ? rms at once. In such cases, the temptation to manage the accounts of banks so as to pretend that they are not really insolvent is strong. Such forbearance has been practiced in various crises (e. g. , the S&L crisis of the 1980s) but it is dangerous: insolvent banks do misbehave (gambling for resurrection was rampant among S&Ls, for example) and experience shows that the cost to the taxpayer, though delayed, is magni? d in the end by such cover-ups. History tells us that, when a crisis hits, honest and speedy cleanups of bank balance sheets are highly desirable: real money is required; accounting tricks won’t do. A striking example is provided by the contrast between the Scandinavian and Japanese crises of the 1990s: Japan’s procrastination led to years of sluggish GDP growth while Scandinavia “bit the bullet” and came back to satisfactory growth much more quickly.
Building an Adaptive Regulatory System in a Global World One problem with regulation in recent years is that it has faced accelerating ? nancial innovation. Of course, ? nancial innovation is driven not just by the desire to serve customers better: it can be the result of pure regulatory arbitrage rather than an attempt to increase social surplus (think of structured products with originators keeping senior tranches in order to minimize capital requirements and providing huge off-balance-sheet, and therefore low-capital-requirement, liquidity support to the conduits).
More 8 • Chapter 1 generally, one can expect regulatory arbitrage by the industry (as in the case of rating agencies offering consulting services to boost the ratings of hard-to-understand structured products). As a consequence, when drafting regulation, legislators should explicitly start from the assumption that these factors will be at play, and they should be willing to adapt the system without delay to these developments. This has, unfortunately, not been the case: regulation is too often designed to “? ht the previous crisis” rather than the next one, and is typically one step behind market developments. The trend toward global banking has exacerbated regulation’s lag behind market developments. Indeed, as stressed earlier, recent years have witnessed two signi? cant trends: bigger ? nancial institutions on the domestic scene (with many domestic mergers) and accelerating globalization (partly due to cross-border mergers). On the one hand, these trends have signi? cantly increased the domestic lobbying power of ? ancial institutions, thereby giving more prominence to a laissez-faire approach. On the other hand, globalization in a world of hard-to-coordinate international regulatory policies has increased the lag between private-sector developments and regulatory responses. Taken together, these factors led to Basel II regulatory rules that were less demanding than their predecessors in terms of capital and that even started delegating bits of the actual implementation of supervision to private-sector actors, namely rating agencies or even the (big) banks themselves.
Keeping a Balance The previous trend toward decreasing capital requirements and increasing delegation of oversight to banks and credit-rating agencies clearly requires a correction, namely a strengthening of regulation. In the recent crisis, the pendulum can be expected to swing in this direction. In such complex industries, however, there are many challenges on the road to ef? cient regulation. The ? rst challenge is the need to avoid overreaction: regulation should mimic for banks the corporate governance of non? nancial ? rms, not “punish” banks just in order to place blame for the crisis. Although ? ancial institutions that are not yet regulated Introduction • 9 should be regulated if the regulated sector has large exposures to them (for example, if they are systemically important owing to their large volume of over-the-counter trades with the regulated sector), and although capital ratios should be raised in comparison to precrisis levels, it is much less clear that one should, for example, become prescriptive in terms of business models in banking: the crisis has hit some small as well as some large banks, some private as well as some state-owned banks, and some specialized as well as some universal banks.
The second challenge is the need for politicians to avoid the temptation to be especially harsh in their treatment of banks that have received a bailout—for example, by limiting their ability to pay managers in comparison to their competitors. This can be counterproductive because it means putting them at a competitive disadvantage toward those banks that have not been bailed out, at least directly (but that may nonetheless have been indirect bene? ciaries of bailouts, as creditors of bailed-out banks).
By contrast, it does make sense to promote compensation schemes that incentivize bank managers to take a long-term perspective. Finally, a danger of the recent crisis is that cross-border banking might collapse. This problem, which would be less dire for some large countries such as the United States, is of ? rst-order importance for European countries and some emerging markets. It is linked to the fact that bailout money originates from national treasuries—which are responsible to national electorates— and not from an internationally oordinated insurance fund. Therefore, it is not a surprise that bailed-out banks have in many cases been ordered to favor domestic lending. This trend can mean the end of the European Union’s Single Market in banking, which is bad news for the Single Market in general, and therefore for economic growth and ef? ciency. The challenges, thus, are numerous. The three essays in this volume discuss a number of principles to deal with these challenges, addressing the microeconomic incentives of ? ancial institutions, the impact of macroeconomic shocks, and the role of political constraints. CHAPTER 2 Lessons from the Crisis Jean Tirole This chapter aims to contribute to the debate on ? nancial system reform. In the ? rst part I describe what I perceive to be a massive regulatory failure, a breakdown that goes all the way from regulatory fundamentals to prudential implementation. Although there has been some truly shocking behavior in the world of ? nance, the universal denunciation of “? nancial madness” is pointless. Managers and employees in the ? ancial industry, like all economic agents, react to the information and incentives with which they are presented. Bad incentives and bad information generate bad behavior. Accordingly, this chapter starts by listing the principal factors that led to the crisis. Although many excellent and detailed diagnoses are now available,1 the ? rst section particularly readable one is the interesting compendium of contributions by NYU economists edited by Acharya and Richardson (2009). More concise and very useful treatments include the introductory chapter of that book as well as Hellwig (2009).
Of course, this review is bound to become dated with respect to rapidly changing events, new proposals, and meetings of one sort or another. For example, this chapter was completed before the December 2009 Basel club of regulators’ proposal of a new solvency and liquidity regime that would deemphasize banks’ internal models of risk assessment, force them to hoard enough liquidity to withstand a 30-day freeze in credit markets and to reduce their maturity mismatch, and prohibit those banks with capital close to the minimum required from istributing dividends. The chapter was also completed before President Obama’s January 21, 2010, announcement of (among other things) his desire to ban retail banks from engaging in propriety trading (running their own trading desks and owning, investing, or sponsoring hedge funds and private equity groups). More generally, Part I makes no attempt at providing an exhaustive account of the crisis or of the various reform proposals that followed it.
I think it fair to say, however, that the underlying policy issues and fundamental tensions, as discussed in the second part of my chapter and in the rest of the book, will not change so quickly. For example, a G20 meeting or two is not going to remove the problem of maturity mismatches or solve the problem of the exposure of the regulated sphere to the unregulated. 1A Lessons from the Crisis • 11 re? ects my own interpretation and is therefore key to understanding the policy conclusions I present later. Many policy makers have forgotten that effective regulation is needed for healthy competition in ? ancial markets, that economic agents should be held accountable for their actions, and that institutions and incentives should lead to a convergence of private and public interests. Although recent events do offer an opportunity for a thorough overhaul of international ? nancial regulation, it is important to strike a balance, showing appropriate political resolve while avoiding the danger of politically motivated reforms in a highly technical domain. The second part of this chapter discusses some implications of recent events for ? nancial-sector regulation.
Part I: What Happened? The crisis, originating in the U. S. home loans market, quickly spread to other markets, sectors, and countries. The hasty sale of assets at ? re-sale prices, a hitherto unprecedented aversion to risk, and the freezing of interbank, bond, and derivatives markets revealed a shortage of high-quality collateral. Starting on August 9, 2007, when the Federal Reserve (Fed) and the European Central Bank (ECB) ? rst intervened in response to the collapse of the interbank market, public intervention reached unprecedented levels.
Few anticipated on that day that many similar interventions would follow, that authorities in various countries would have to bail out entire sectors of the banking system, that the bailout of some of the very largest investment banks, a major international insurance company, and two huge government-sponsored companies guaranteeing mortgage loans would cost the American taxpayer hundreds of billions of dollars. A little more than a year later, in the autumn of 2008, the American government had already committed 50 percent of U.
S. GDP to its remedial efforts. 2 mid-November 2008, Bloomberg estimated that $7,400 billion, an amount equal to 50 percent of U. S. GDP, had been guaranteed, lent, or spent by the Fed, the U. S. Treasury, and other federal agencies. On September 2, 2009, the Federal Reserve had $2,107 billion in various assets (including mortgage-backed securities, commercial paper loans, and direct loans to AIG and banks), the Treasury $248. 8 billion in Troubled Asset Relief Program (TARP) investments in banks 2 In 12 • Chapter 2
Equally unforeseen was that American and European governments would ? nd themselves lending signi? cant sums directly to industrial companies to save them from bankruptcy. Although the crisis has macroeconomic consequences in terms of an immediate and severe recession and of a sharp increase in public debt,3 this chapter is concerned with ? nancial regulation. Policy makers and economists must have a clear understanding of what happened in order to suggest ways out of the crisis, and especially to propose reforms that will fend off future crises of a similar nature.
The proper application of standard economics would in some areas have surely allowed us to steer clear of many obvious errors; and yet the crisis provides us with prima facie evidence on how regulations are designed and evaded, and scope for new thinking about our ? nancial system. The recent ? nancial crisis will quickly become a central case study for university courses on information and incentives. The losses on the American subprime mortgage market,4 although signi? cant, were very small relative to the world economy and by themselves could not account for the ensuing “subprime crisis. In other words, the subprime market meltdown was just a detonator for what followed, namely a sequence of incentives and market failures exacerbated by bad news. At each stage in the chain of risk transfers, asymmetric information between contracting parties hampered proper market functioning. Nonetheless, market failures related to asymmetric information are a permanent feature of ? nancial markets, so the crisis cannot be explained simply in terms of market failures. Two other factors played a critical role.
First, a blend of inappropriate and poorly implemented regulation, mainly in the United States but also in Europe, gave individual actors incentives to take sizable and AIG, and the Federal Deposit Insurance Corporation (FDIC) $386 billion in bank debt guarantees and loss-share agreements (source: Wall Street Journal Europe edition). 3 Budget de? cits have reached levels unprecedented in peacetime; the steep rise in indebtedness of Western governments will limit room for maneuver in the medium term.
Sovereign debt crises might even emerge in member countries of the Organization for Economic Cooperation and Development (OECD), a contingency that was rather remote before the crisis. 4 Around $1,000 billion, or only 4 percent of the market capitalization of the New York Stock Exchange at the end of 2006 ($25,000 billion), according to the November 2008 estimates of the International Monetary Fund (IMF). Lessons from the Crisis • 13 risks, with a major portion of these risks ultimately borne by taxpayers and investors.
Second, market and regulatory failures would never have had such an impact if excess liquidity had not encouraged risk-taking behavior. A Political Resolution to Favor Real Estate The U. S. administration, Congress, and other of? cials, including some at the Fed, were eager to promote the acquisition of homes by households. 5 In addition to the incentive for purchasing a home provided by the long-standing and generous tax deductibility of interest paid on mortgages, households were encouraged to lever up their debt in order to acquire homes. 6 Consumer protection was weak, to say the least.
Many subprime borrowers were given low “teaser” rates for two or three years, with rates skyrocketing thereafter. They were told that real estate prices would continue to increase and therefore they would be able to re? nance their mortgages. Similarly, mortgages indexed to market interest rates (adjustable-rate mortgages, ARMs), which raise obvious concerns about borrowers’ ability to make larger payments when interest rates rise, were promoted in times of low interest rates. 7 Alan Greenspan himself called for an increase in the proportion of ARMs. 5 Fortunately, this was not the case in the euro area, where the ECB followed a more stringent monetary policy and authorities in a number of countries did not encourage subprime loans. Of course, loose monetary policy is only a contributing factor, as can be seen from the examples of Australia and Great Britain, two countries where the mortgage market boomed in spite of relatively normal interest rates. 6 There are several very good outlines of the excesses linked to the housing market—see, for example, Calomiris (2008), Shiller (2009), and Tett (2009). France has for the most part been spared this phenomenon. French banks have traditionally lent to solvent households, a practice reinforced by law (the Cour de Cassation ruled against a ? nancial institution that had failed in its duty of care by granting a loan incommensurate with the borrower’s present or future capacity to repay). Variable-rate loans have always played a relatively minor role in France (24 percent of outstanding loans in 2007), and completely ? exible loans, where neither interest rates nor monthly payments are capped, have always had a very small market share (less than 10 percent).
Adjustable-rate mortgages are, by contrast, very popular in Spain, the United Kingdom, and Greece. 8 According to USA Today (February 23, 2004), “While borrowers can re? nance ? xed-rate mortgages, Greenspan said homeowners were paying as much as 14 • Chapter 2 Finally, public policy encouraged institutions to lend to subprime borrowers through several channels. Fannie Mae and Freddie Mac were pushed to increase the size of their balance sheets. And loose regulatory treatment of securitization and mortgagebacked securities helped make mortgage claims more liquid.
In response to these policy and social trends, subprime lending changed in nature. Before the ? rst decade of the twenty-? rst century, lenders would carefully assess whether subprime borrowers were likely to repay their loans. By contrast, recent subprime lending involved an explosion of loans without documentation. For instance, lenders were able to base their calculations on claimed, rather than actual, income. We will return to these developments. Not surprisingly, U. S. homeownership rose over the period 1997–2005 for all regions and for all age, racial, and income groups. The fraction of owner-occupied homes increased by 11. percent over this period. Housing prices moved up nine years in a row, and across the entire United States. 9 The rise was particularly spectacular for low-income groups. Correspondingly, real estate price indexes in the lowest price tier showed the biggest increases until 2006 and the biggest drop afterward. Excessive Liquidity, the Savings Glut, and the Housing Bubble Crises usually ? nd their origin in the lack of discipline that prevails in good times. Macroeconomic factors provided a favorable context for ? nancial institutions to take full advantage of the breaches created by market and regulatory failure.
In addition to the political support for real estate ownership, there are several reasons why the origin of the crisis was located in the United States: 0. 5 to 1. 2 percentage points for that right and the protection against a potential rate rise, which could increase annual after-tax payments by several thousand dollars. He said a Fed study suggested many homeowners could have saved tens of thousands of dollars in the last decade if they had ARMs. ” Adjustable-rate mortgages made up 28 percent of mortgages in January 2004 in the United States. 9 These data are aken from Shiller (2009, 5, 36). Lessons from the Crisis • 15 a savings glut—expanding the set of borrowers and reducing margins on conforming loans A strength of the U. S. ?nancial system is that it creates large numbers of tradable securities, that is, stores of value that can easily be acquired and sold by investors trying to adapt to the lack of synchronicity between cash receipts and cash needs. The large volume of securities in the United States was attractive to investors in other countries seeking new investment opportunities and unable to ? d suf? cient amounts of stores of value at home. Surpluses in the sovereign wealth funds of oil-producing and Asian states and the foreign-exchange reserves of countries, such as China, that were enjoying export-led growth built on an undervalued currency, tended to gravitate to the United States. This cash in? ow reduced the available volume of stores of value within the United States, and the net increase in the demand for securities stimulated an accelerated securitization of debt so as to create new stores of value that were greatly in demand. 0 Thus, the international savings glut contributed to the increase in securitization that will be described shortly. Abundant liquidity in the United States led ? nancial institutions to search for new borrowers. They extended their activity in the segment of “nonconforming” or “subprime” loans, that is, loans that do not conform to the high lending standards used by the federal-government-backed Fannie Mae and Freddie Mac. But the enhanced competition associated with excess liquidity also eroded margins made on loans to safer borrowers.
This implied that the losses incurred on subprime loans could not be offset by high margins on more traditional lending. loose monetary policy The very low short- and long-term interest rates that prevailed for several years in the early 2000s (for instance, a negative Fed funds real rate from October 2002 through April 2005) made argument was developed in particular by Caballero, Farhi, and Gourinchas (2008a, 2008b). Ben Bernanke has often pointed to the excess of international savings as the cause of excess liquidity in the U. S. economy before the subprime mortgage crisis. 10 This 16 •
Chapter 2 borrowing extremely cheap. Low short-term rates sow the seeds of a potential crisis through multiple channels: First, they lower the overall cost of capital and thereby encourage leverage. Second, they make short-term borrowing relatively cheap compared to long-term borrowing, and therefore encourage maturity mismatches. Low short-term rates thus make for bigger and less liquid balance sheets. Third, low short-term rates signal the central bank’s willingness to sustain such rates, and therefore suggest that, were a crisis to come, the central bank would lower rates and facilitate re? ancing, making illiquid balance sheets less costly for ? nancial institutions. asset price bubble The crisis has revived the debate over the proper attitude of monetary authorities to an asset market-price boom. The stance of central banks in general, and of Alan Greenspan in particular, has been that their remit is limited to in? ation and growth, and does not include the stabilization of asset prices, at least insofar as these do not form an in? ationary threat. Ben Bernanke, for instance, argued in a series of in? ential articles11 that (a) it is usually hard to identify a bubble,12 and (b) bursting a bubble may well trigger a recession. 13 An auxiliary debate has focused on how authorities should burst a bubble, assuming they have identi? ed one and are willing to risk a recession. It is by no means clear that monetary policy, which controls only short-term rates, is the appropriate instrument. Regulation (by controlling the ? ow of credit to the bubble market) and ? scal policy (by issuing pubfor example, Bernanke (2000). ake a recent example, one can ask whether the extensive implicit subsidy of mortgages (through ? scal policy, through the government’s implicit backing of Fannie Mae and Freddie Mac, and through very low minimum capital requirements for liquidity support granted to vehicles resulting from securitization) did not in? ate the perception of mortgage “fundamentals. ” Ben Bernanke himself in 2005 viewed the unprecedented housing price levels as re? ecting strong economic fundamentals rather than a bubble (Tett 2009, 122). 13 See, e. g. Farhi and Tirole (2010) for a theoretical treatment of the impact of asset price bubbles and their crashes on economic activity. 12 To 11 See, Lessons from the Crisis • 17 lic debt and raising interest rates) seem to have a better chance of terminating a bubble. The alternative14 to bursting a bubble lies in the government accumulating reserves in advance of such a breakdown. When a bubble ends abruptly, losses are suffered both in the ? nancial and real sectors of the economy, and countercyclical policy becomes necessary. For countercyclical policy to have suf? ient room for maneuver, however, governments must have followed conservative ? scal policies during the upswing of the cycle, so as to be able to effectively counter the downswing. In the debate on the opportunity to stabilize asset prices, it is also important to remember that not only does the extent of the bubble need to be identi? ed, but also who is involved in it. The dotcom bubble at the end of the 1990s created only a very moderate recession when it burst in 2001 because the securities were held mainly by individual households.
By contrast, in the recent crisis, heavy losses have been suffered by a broad range of leveraged ? nancial intermediaries, creating widespread problems of liquidity and of solvency. Robert Shiller, an early and strong proponent of the view that the real estate market exhibited a bubble, has proposed that the short-selling of real estate be made easier, to facilitate stabilizing speculation by those who realize that a bubble is under way. ominous signals The unfolding of the crisis is now well known.
Macroeconomic developments led to the stagnation of house prices in 2006; prices in overheated housing markets such as Florida and California stalled; the Fed, which had decreased short-term interest rates from 2000 through 2004 (the Fed funds rate15 went from 6. 50 percent in May 2000 to 1 percent, until June 30, 2004, when it started moving up again), started raising them again (the Fed funds rate was 5. 25 percent in September 2007). In 2006–2007, Chicago Mercantile Exchange housing futures markets predicted large declines in home prices as market participants started worrying about defaults by subprime borrowers. y Ricardo Caballero in particular. is the rate at which banks lend available funds (reserves at the Fed) to each other overnight. 15 This 14 Proposed 18 • Chapter 2 It was feared that many households whose variable loans were about to reset at higher interest rates would not be able to afford the new terms as stagnating prices made re? nancing impossible. Others would go into “strategic default” and not repay their loans when they would go into negative equity (with mortgage balances larger than the total value of their homes). Although the concerns were very real, it was hard to put clear ? ures on the magnitude of likely losses. The lag between the signing of a contract and the transition to a higher variable rate, as well as traditional lags associated with downward movements in the housing market, created a real ? nancial time bomb. Furthermore, the cost of borrower default for lenders (including administrative costs, the physical deterioration of vacated homes, taxes, unpaid insurance, realtors’ commissions, and falling housing prices) is highly sensitive to the rate of decline in housing prices and other macroeconomic developments. For example, J. P. Morgan estimated in
January 2008 that for a decrease of 15 percent in house prices the losses arising from the default of an average “Alt-A adjustable-rate mortgage”16 taken out in 2006 would be around 45 percent. 17 Another reason why losses are dif? cult to forecast is uncertainty about public policy, as the rate of unrecovered debt also depends on the level of government assistance. 18 16 Alt-A mortgages have a risk pro? le between “prime” and “subprime” loans. For example, the borrower has never defaulted, but the borrowing involves a high level of debt and quite possibly incomplete documentation of ? ancial standing. 17 Cited by Calomiris (2008, 23). 18 The FDIC proposed subsidizing a revision of loan conditions, temporarily reducing the rate of interest to be paid by the borrower, and possibly extending the loan term beyond the standard thirty years. Under current law, it is by contrast much more dif? cult to reduce the principal repayable by the borrower because no such renegotiation can be done without the endorsement of those holding the debt collateralized by the mortgage loan during the process of securitization.
The FDIC proposed that the government should underwrite the losses suffered by lenders provided, among other conditions, that the renegotiation resulted in the borrowers’ not spending more than 31 percent of their income on mortgage payments. Lessons from the Crisis • 19 Excessive Securitization Although lenders had traditionally retained the bulk of their loans on their own balance sheets, more recently the underlying assets (the repayment of interest and principal on mortgages) were transferred to ? nancial intermediaries, or off-balance-sheet “structured investment vehicles” or “conduits. These intermediate structures were ? nanced mostly through short-term borrowing (say, through commercial paper with an average maturity of about one month). A key innovation was the use of “tranching,” as the revenues attached to these structures were divided into different risk classes to suit the needs of different investors. For example, some investors, for risk management or for regulatory reasons, have a high demand for safe AAA securities. 19 Others do not mind taking on more risk. The rate of securitization of housing loans grew from 30 percent in 1995 to 80 percent in 2006.
More tellingly, in the case of the subprime loans the securitized proportion went from 46 percent in 2001 to 81 percent in 2006. Securitization is a long-established practice, with clear rationales: First, it allows loan providers to re? nance themselves. With the resulting cash, they can then ? nance other activities in the economy—securitization therefore transforms “dead capital” into “live capital,” to use De Soto’s (2000) terminology. Second, when stores of value are in scarce net supply in the economy, the creation of new securities ful? ls a demand; this incentive to create new securities in reaction to the savings glut, as we have argued, played a role in the recent increase in securitization. Finally, in those cases where risks are heavily concentrated, securitization also allows lenders to diversify and spread risk. Securitization however, shifts responsibility away from the lender, whose incentive to control the quality of its lending is reduced if detailed accounts of the securitization process, see, e. g. , Franke and Krahnen (2008), Brunnermeier (2009), and Tett (2009). 19 For 20 Chapter 2 it will not suffer the consequences. 20 The lender may make marginal loans and then divest itself via securitization, without the buyers being able to detect the lack of due diligence. In fact, the rate of default on housing loans of broadly similar characteristics, but differentiated by whether they can easily be securitized or not, can increase by 20 percent according to some estimates when securitization is an option. 21 This fundamental tension between the creation of liquid assets and incentives to monitor loan quality has two corollaries.
First, the lender should not completely disengage itself and should retain part of the risk, as is done, for instance, by insurance companies when they transfer part of their risk to reinsurers. Second, securitization should be linked to certi? cation, a process obligatory for gaining market access and found in other institutions (for example, initial public offerings). Certi? cation should involve a rigorous scrutiny on the part of buyers and rating agencies. These two principles have not always been followed in the recent crisis.
First, the practice of securitization took off at a point when loans became riskier and therefore highly susceptible to informational asymmetries, whereas theory and good practice would dictate that banks should then retain a greater proportion. Lending banks, contrary to tradition, divested themselves of junior (risky) tranches, sometimes in response to the requirements of the prevailing regulatory framework. 22 A number of institutions (such as AIG, UBS, Merrill Lynch, and Citigroup) started sitting on a vast position of the so-called super-senior debt, which they either held directly or insured.
Second, buyers of these securitized loans made their purchases without paying much attention to their quality. Presumably, the fact that the loans were not retained by the original lender should have given the buyers a hint of the likely quality of these loans. But buyers had little incentive to monitor the quality of what 20 Incentive effects and the dangers of securitization have been extensively discussed in the economic literature; see, e. g. , Dewatripont and Tirole (1994). 21 See Keys et al. (forthcoming). 22 For example, for commercial banks, prudential rules require that 8 percent of assets (weighted by risk) be covered by equity.
For triple A tranches, risk is estimated at merely 20 percent, so only 1. 6 cents of equity capital is required for each dollar of such assets. Lessons from the Crisis • 21 they were buying, in part because favorable credit ratings translate into low capital requirements. Because leverage is the key to pro? tability, not to mention (for ? nanciers who are heavily egodriven) the prospect of being at the top of league tables,23 any risk that buyers were taking by buying these securitized assets was compensated by an opportunity to increase the size of their balance sheets.
Some readers may say that banks, on the whole, kept substantial exposure to the vehicles that they had created. But as we shall see later, they pledged large amounts of liquidity support in case the vehicles had trouble re? nancing on the wholesale markets. But that risk was primarily macroeconomic in nature, while the incentives to monitor loans should have been preserved by keeping more of the microeconomic risk! The Laxity of Credit-Rating Agencies Credit-rating agencies are once again under ? e. 24 In their defense, a foreshortened historical perspective has hindered proper appreciation of the risks linked with newly introduced instruments such as collateralized debt obligation (CDO) tranches or credit-default swaps. Furthermore, the weakness of the macroeconomic treatment in the agencies’ models and the departure of personnel lured by clients contributed to poor risk assessment. Yet the failure of rating agencies to ful? ll their duties is obvious.
A number of incentive misalignments have repeatedly been pointed out by critics: • The agencies provided preliminary evaluations (prerating assessments) that allowed lenders to form an idea of what their eventual rating would be, harming transparency. 25 tables rank the leaders in various areas of banking. agencies have been criticized before, for instance after the sovereign debt crises of the 1990s and after the bursting of the Internet bubble, both of which they failed to foresee. They reacted very slowly to the problems of Enron, WorldCom, and other companies that failed in 2001. 5 Such services were requested by lenders, which also did not hesitate to engage in “ratings shopping” for the most favorable rating. Calomiris (2008) notes that Congress, as well as the Securities and Exchange Commission, encouraged ratings in? ation. 24 Credit-rating 23 League 22 • Chapter 2 • In addition, the agencies explained to issuers how they should structure their tranches to barely secure a given rating, say AAA. Even if laxity had been absent, this one practice implied that an AAA tranche carried a probability of default higher than that of AAA securities that had not been the subject of such advice.
The activity of credit-rating agencies in explaining how the threshold might be minimally passed rendered the composition of such tranches marginal rather than average. • The incentives faced by rating agencies seem to have been somewhat perverse, with the commissions paid to agencies being proportional to the value of the issue, therefore generating pressure toward overrating. 26 Rating agencies would normally balance the gains from being easy on issuers against a loss of reputation which would reduce the credibility of their ratings among investors and therefore make agencies less attractive to issuers in the future. The desire to please investment banks providing an important percentage of their turnover (structured ? nance products represented a fraction of close to half of the rating agencies’ revenue at the end of the boom) no doubt had a bad in? uence. • Finally, the ratings market is very concentrated. There are only three large agencies, and two of them (Moody’s and Standard and Poor’s) share 80 percent of the market. Where a dual rating is required, these agencies ? nd themselves in a quasi-monopoly situation. An Excessive Maturity Transformation a gigantic maturity mismatch . . One essential feature of banking intermediation has always been maturity transformation. The banking system as a whole transforms short-term borrowing from depositors into long-term 26 In June 2008, the three top rating agencies signed a pact with New York’s attorney general. Under the old fee system, the agencies had a ? nancial incentive to assign high ratings because they received fees only if a deal was completed; under the new agreement, by contrast, the rating agencies receive payments for service even if a deal is not completed (source: Reuters).
Lessons from the Crisis • 23 loans to ? rms. As has long been recognized, this maturity transformation creates hazards for the ? nancial sector. If short-term borrowing is not rolled over, then the banks’ liquidity dries up, and the banking system ? nds itself in trouble. This is especially the case if the bank’s creditors panic and seek to withdraw their deposits for fear that the bank might become insolvent. Such panics have now practically vanished for small depositors covered by deposit insurance, but they remain an issue in wholesale ? ance. Moreover, even if there is no panic, a rise in the shortterm interest rate has immediate repercussions for the cost of funds for the ? nancial institution, upsetting its balance sheet. 27 Recently a number of ? nancial intermediaries—banks and nonbanks—have taken substantial risks by borrowing at very short maturities in wholesale markets (Fed funds market, commercial paper). This strategy is very pro? table when the rate of interest is low, but it exposes the ? nancial institution to a rise in the interest rate.
The leading commercial-bank illustration of this risk is Northern Rock, whose collapse proved to be very costly for the British taxpayer. The details of this banking panic have been discussed at length in newspapers28 (for the ? rst time since the 19th century a British bank suffered a run on its retail deposits), but the more fundamental problem was Northern Rock’s loss of access to wholesale markets. Three-quarters of Northern Rock deposits were secured wholesale, primarily on very short-term conditions. As already noted, transformation (borrowing short and lending long) is a traditional feature of banking activity.
More and more institutions, however, took a gamble on the yield curve, case in point is that of SIVs, which were ? nanced almost entirely with short-term liabilities and in early August 2007 saw their ? nancing costs explode as the interest rate on asset-backed commercial paper (i. e. , liabilities between one day and six months collateralized by assets) moved from 5–10 basis points above the American overnight borrowing rate to 100 basis points (Tett 2009, 182). 28 Deposit insurance in the United Kingdom was at the time poorly structured. Only ? ,000 per person was completely covered by this insurance, the next ? 33,000 being guaranteed up to 90 percent. This partial insurance provided an incentive to run, even for depositors with very little savings in the bank. By comparison, deposit insurance in the United States was temporarily raised from the standard $100,000 to $250,000 until December 2009; deposits are fully insured up to €70,000 in France. 27 A 24 • Chapter 2 betting on short-term rates remaining low and access to wholesale markets remaining easy. Several observations support this view. 9 • Commercial banks pledged substantial liquidity support to the conduits, promising to supply liquidity in case the conduits had trouble ? nding funds in the wholesale market. According to Acharya and Schnabl (2009), the ten largest conduit administrators (mainly commercial banks) had a ratio of asset-backed commercial paper to equity ranging from 32. 1 percent to 336. 6 percent in January 2007. See the accompanying table, drawn from Acharya and Schnabl’s chapter. These liquidity support pledges represented an elementary form of regulatory evasion.
Such off-balancesheet commitments carried much lower capital requirements than would have been the case had the liabilities been on the balance sheets. The increase in the market share of investment banks mechanically increased the ? nancial sector’s interest-rate fragility, as investment banks rely on repo and commercialpaper funding much more than commercial banks do. Primary dealers increased their overnight to term borrowing ratio. Leveraged buyouts have become more leveraged. Investment banks explained to their clients how to make high returns through derivative products that bet on falling interest rates. 0 • • • • Five large investment banks,31 lacking liquidity, either went bankrupt or merged with commercial banks, with the support of the U. S. government. Lehman Brothers was the biggest default in the history of the United States ($613 billion of debt, $639 billion of assets). In September 2008, Morgan Stanley and Goldman Sachs more details on increased transformation, see Adrian and Shin (2008). Tett (2009, 36). 31 A merchant bank (also called an investment bank) has two main activities: (1) portfolio management (shares, debentures, etc. , and (2) market making and acting as a counterparty in over-the-counter (OTC) trading. Unlike commercial (retail) banks, investment banks do not take retail deposits and therefore are not subject to standard banking regulation. 30 See 29 For Ten Largest Conduit Administrators by Size Conduits Number Citibank ABN Amro Bank of America HBOS JPMorgan Chase HSBC Societe Generale Deutsche Bank Barclays WestLB 23 9 12 2 9 6 7 14 3 8 CP (in ? bn) 93 69 46 44 42 39 39 38 33 30 Assets (in ? bn) 1,884 13,000 1,464 1,160 1,352 1,861 1,260 1,483 1,957 376 Administrator Equity (in ? n) 120 34 136 42 116 123 44 44 54 9 CP/Asset (%) 4. 9 5. 3 3. 1 3. 8 3. 1 2. 1 3. 1 2. 6 1. 7 8. 0 CP/Equity (%) 77. 4 201. 1 33. 7 105. 6 36. 1 32. 1 87. 2 87. 8 61. 5 336. 6 Source: Acharya and Schnabl (2009) CP = commercial paper 26 • Chapter 2 became bank holding companies. Merrill Lynch was bought by Bank of America, and Bear Stearns by JPMorgan Chase. Accordingly, all are now regulated by the Fed. Before then, the solvency and liquidity of investment banks had been subject to supervision by the Securities and Exchange Commission (SEC) since 2004, on a voluntary basis.
The SEC had assigned the task of supervising investment banks (with $4,000 billion in assets) to just seven employees! Furthermore, the concern shown by these supervisors had been simply ignored. 32 Thanks to the stability of their insured retail deposits, American commercial banks were initially slightly better able to withstand the crisis, even though various bankruptcies and the fragility of giants such as Citi and Bank of America remind us that retail banks also took gigantic risks and were highly dependant on wholesale short-term funding. 3 . . . that puts monetary authorities in a bind The generalization of risk taking through high levels of transformation puts monetary authorities in a dif? cult position. Either they do not react when interest rates rise again (risking the bottom falling out of the ? nancial system), or they yield and maintain interest rates at an arti? cially low level and indirectly bail out institutions that have taken excessive risks. Monetary authorities found themselves trapped by generalized transformation and, sure enough, the Fed funds rate fell from 5. 5 percent on September 18, 2007, to 0 percent on December 16, 2008. Farhi and Tirole (2009) show that keeping interest rates low has several costs beyond validating past excessive transformation: First, as we have seen, loose monetary policy encourages institutions to persist with the same bad behavior, paving the way for the next crisis, through two channels: low short-term rates (1) make a short liability maturity structure appealing to ? nancial institutions, and (2) boost ? nancial institutions’ leverage by lowering their overall cost of capital. Labaton (2008). comparison of capital positions of retail and investment banks at the onset of the crisis, see Blundell-Wignall and Atkinson (2008). 33 For 32 See Lessons from the Crisis • 27 Second, loose monetary policy distorts interest rates away from their natural level, discouraging savings; loose monetary policy may also distort relative prices and create in? ation. Third, a loose monetary policy transfers resources from lenders to borrowers; in particular, the recent episode has seen a sizable transfer from consumers to institutions through this channel, which is much less visible than ordinary (? cal) bailouts. To be clear—the central banks could not let institutions with excessive transformation go under by raising interest rates. They were “stuck. ” My point is that during the boom they should have prevented the emergence of this “fait accompli. ” Preventive measures were called for, as ex post toughness is neither desirable (despite the costs of leniency) nor credible. The solution in my view lies with monitoring transformation not only at the institution’s level, but also overall. It is important that multiple “strategic” ? ancial institutions do not simultaneously encounter re? nancing problems, as was the case in the crisis. Let us conclude this section with two remarks about maturity transformation and the sensitivity of balance sheets to interestrate movements. First, maturity transformation is a natural way for ? nancial institutions to correlate their risks (in this instance by betting on low interest rates), but it is by no means the only way. For example, before the crisis many ? nancial institutions were simultaneously trying to increase their exposure to the subprime market to boost their returns. 4 While that market is itself in? uenced by the interest rate, it has other drivers, and so was another source of correlated distress. Second, many observers35 extol the merits of a “market solution” to the problem of insuring deposits in the banking sector. The idea is that the fees paid by the banks for deposit insurance 34 E. g. , Tett (2009, 124). Tett (p. 102) points at another, unexpected source of correlation: the use of the same statistical techniques (Li’s Gaussian copula approach), the miscalibration of which introduced correlated errors.
The common ass
