Datasets:

smarquie

/

fincrisis

like 0

F I X I N G F I N A N C E S E R I E S – PA P E R 3 | N O V E M B E R 2 0 0 8 The Origins of the Financial Crisis Martin Neil Baily, Robert E. Litan, and Matthew S. Johnson The Initiative on Business and Public Policy provides analytical research and constructive recommendations on public policy issues affecting the business sector in the United States and around the world.The Origins of the Financial Crisis Martin Neil Baily, Robert E. Litan, and Matthew S. Johnson The Initiative on Business and Public Policy provides analytical research and constructive recommendations on public policy issues affecting the business sector in the United States and around the world.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008  ConTenTS Summary 7 Introduction 10 housing Demand and the Perception of Low risk in housing Investment 11 The Shifting Composition of Mortgage Lending and the erosion of Lending Standards 1 economic Incentives in the housing and Mortgage origination Markets 20 Securitization and the Funding of the housing Boom 22 More Securitization and More Leverage—CD os, SIVs, and Short-Term Borrowing 27 Credit Insurance and Tremendous growth in Credit Default Swaps 32 The Credit rating Agencies 3 Federal reserve Policy, Foreign Borrowing and the Search for Yield 36 regulation and Supervision 0 The Failure of Company risk Management Practices 2 The Impact of Mark to Market 3 Lessons from Studying the origins of the Crisis  references 6 About the Authors 7T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 7The financial crisis that has been wreaking havoc in markets in the U.S. and across the world since August 2007 had its origins in an asset price bubble that interacted with new kinds of financial innovations that masked risk; with compa - nies that failed to follow their own risk management procedures; and with regulators and supervisors that failed to restrain excessive risk taking. A bubble formed in the housing markets as home prices across the country increased each year from the mid 1990s to 2006, moving out of line with fun - damentals like household income. Like traditional asset price bubbles, expectations of future price increases developed and were a significant factor in inflating house prices. As individuals witnessed rising prices in their neighborhood and across the country, they began to expect those prices to con - tinue to rise, even in the late years of the bubble when it had nearly peaked. The rapid rise of lending to subprime borrowers helped inflate the housing price bubble. Before 2000, subprime lending was virtually non-existent, but thereafter it took off exponentially. The sus - tained rise in house prices, along with new financial innovations, suddenly made subprime borrowers — previously shut out of the mortgage markets — attractive customers for mortgage lenders. Lend - ers devised innovative Adjustable Rate Mortgages (ARMs) — with low “teaser rates,” no down-pay - ments, and some even allowing the borrower to postpone some of the interest due each month and add it to the principal of the loan — which were predicated on the expectation that home prices would continue to rise. But innovation in mortgage design alone would not have enabled so many subprime borrowers to access credit without other innovations in the so- called process of “securitizing” mortgages — or the pooling of mortgages into packages and then sell -SUMMA rY ing securities backed by those packages to inves - tors who recei ve pro rata payments of principal and interest by the borrowers. The two main govern - ment-sponsored enterprises devoted to mortgage lending, Fannie Mae and Freddie Mac, developed this financing technique in the 1970s, adding their guarantees to these “mortgage-backed securities” (MBS) to ensure their marketability. For roughly three decades, Fannie and Freddie confined their guarantees to “prime” borrowers who took out “conforming” loans, or loans with a principal below a certain dollar threshold and to borrowers with a credit score above a certain limit. Along the way, the private sector developed MBS backed by non- conforming loans that had other means of “credit enhancement,” but this market stayed relatively small until the late 1990s. In this fashion, Wall Street inve stors effectively financed homebuyers on Main Street. Banks, thrifts, and a new industry of mortgage brokers originated the loans but did not keep them, which was the “old” way of financ - ing home ownership. Over the past decade, private sector commercial and inves tment banks developed new ways of se - curitizing su bprime mortgages: by packaging them into “Colla teralized Debt Obligations” (sometimes with other asset-backed securities), and then divid - ing the cash flows into different “tranches” to ap - peal to diffe rent classes of investors with different tolerances for risk. By ordering the rights to the cash flows, the developers of CDOs (and subse - quently other sec urities built on this model), were able to convince the credit rating agencies to assign their highest ratings to the securities in the high - est tranche, or risk class. In some cases, so-called “monoline” bond insurers (which had previously concentrated on insuring municipal bonds) sold protection insurance to CDO investors that would pay off in the event that loans went into default. In other cases, especially more recently, insurance companies, inves tment banks and other parties did T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 8 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on the near equivalent by selling “credit default swaps” (CDS), which were similar to monocline insurance in principle but different in risk, as CDS sellers put up very little capital to back their transactions. These new innovations enabled Wall Street to do for subprime mortgages what it had already done for conforming mortgages, and they facilitated the boom in subprime lending that occurred after 2000. By channeling funds of institutional investors to support the origination of subprime mortgages, many households pr eviously unable to qualify for mortgage credit became eligible for loans. This new group of eligible borrowers increased housing demand and helped inflate home prices. These new financial innovations thrived in an en - vironment of easy monetary policy by the Fed - eral Reserve and poor regulatory oversight. With interest rates so low and with regulators turning a blind eye, fina ncial institutions borrowed more and more money (i.e. increased their leverage) to finance their pu rchases of mortgage-related securi - ties. Banks created off-balance sheet affiliated enti - ties such as Structured Investment Vehicles (SIVs) to purchase mortgage-related assets that were not subject to regulatory capital requirements Finan - cial institutions also turned to short-term “collater - alized borro wing” like repurchase agreements, so much so that by 2006 investment banks were on average rolling over a quarter of their balance sheet every night. During the years of rising asset prices, this short-term debt could be rolled over like clock - work. This tenuous situation shut down once panic hit in 2007, however, as sudden uncertainty over as - set prices caused lenders to abruptly refuse to roll - over their debts, and over-leveraged banks found themselves exposed to falling asset prices with very little capital. While ex post we can certainly say that the system- wide increase in borrowed money was irresponsible and bound for catastrophe, it is not shocking that consumers, would-be homeowners, and profit- maximizing banks will borrow more money when asset prices are rising; indeed, it is quite intuitive. What is especially shocking, though, is how insti - tutions along each link of the securitization chain failed so grossly to perform adequate risk assess - ment on the mortgage-related assets they held and traded. From the mortgage originator, to the loan servicer, to the mortgage-backed security issuer, to the CDO issuer, to the CDS protection seller, to the credit rating agencies, and to the holders of all those securities, at no point did any institution stop the party or question the little-understood com - puter risk models, or the blatantly unsustainable deterioration of the loan terms of the underlying mortgages. T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 A key point in understanding this system-wide fail - ure of risk assessment is that each link of the secu - ritization chain is plagued by asymmetric informa - tion – that is, one party has better information than the other. In such cases, one side is usually careful in doing business with the other and makes every effort to accurately assess the risk of the other side with the information it is given. However, this sort of due diligence that is to be expected from markets with asymmetric information was essentially absent in recent years of mortgage securitization. Com - puter models took the place of human judgment, as originators did not adequately assess the risk of borrowers, mortgage services did not adequately assess the risk of the terms of mortgage loans they serviced, MBS issuers did not adequately assess the risk of the securities they sold, and so on. The lack of due diligence on all fronts was partly due to the incentives in the securitization model itself. With the ability to immediately pass off the risk of an asset to someone else, institutions had lit - tle financial incentive to worry about the actual risk of the assets in question. But what about the MBS, CDO, and CDS holders who did ultimately hold the risk? The buyers of these instruments had every incentive to understand the risk of the underlying assets. What explains their failure to do so?One part of the reason is that these investors — like everyone else — were caught up in a bubble men - tality that enveloped the entire system. Others saw the large profits from subprime-mortgage related assets and wanted to get in on the action. In addition, the sheer complexity and opacity of the securitized financial system meant that many people simply did not have the information or capacity to make their own judgment on the securities they held, instead relying on rating agencies and complex but flawed computer models. In other words, poor incentives, the bubble in home prices, and lack of transparency erased the frictions inherent in markets with asym - metric information (and since the crisis hit in 2007, the extreme opposite has been the case, with asym - metric information problems having effectively frozen credit markets). In the pages that follow, we tell this story more fully. T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 10 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on tion of mortgage lending and the erosion of lending standards; economic incentives in the housing and mortgage origination markets; securitization and the funding of the housing boom; the innovations in the securitization model and the role of leveraged financial institutions; credit insurance and growth in credit default swaps; the credit rating agencies; federal reserve policy and other macroeconomic factors; regulation and supervision; the failure of company risk management practices; and the im - pact of mark to market accounting. The paper con - cludes with a preview of subsequent work in the Fixing Finance series by describing some lessons learned from studying the origins of the crisis. 1. There exists much literature that also seeks to explain the events leading up to the crisis. Also see Ashcraft and Schuermann (2008), Calomiris (2008), Gerardi, Lenhart, Sherlund, and Willen (2008). Gorton (2008), Demyanyk and Hemert (2008), among many others.The financial crisis that is wreaking havoc in financial markets in the U.S. and across the world has its origins in an asset price bubble that interacted both with new kinds of financial in - novations that masked risk, with companies that failed to follow their own risk management pro - cedures, and with regulators and supervisors that failed to restrain excessive taking. In this paper, we attempt to shed light on these factors.1 The paper is organized as follows: the first section addresses the bubble that formed in home prices over the decade or so up to 2007 and the factors that affected housing demand during those years. The following sections address: the shifting composi -InTroDUCTI onT hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 11The driving force behind the mortgage and fi - nancial market excesses that led to the current credit crisis was the sustained rise in house prices and the perception that they could go no - where but up. Indeed, over the period 1975 through the third quarter of 2006 the Office of Federal Housing Enterprise Oversight (OFHEO) index of house prices hardly ever dropped. Only in 1981-82 did this index fall to any significant extent—5.4 per - cent—and that was the period of the worst recession in postwar history. From 1991 through the third quarter of 2007, the OFHEO house price index for the U.S. showed increases in every single quarter, when compared to the same quarter in the prior year. Rates of price increase moved above 6 percent in 1999, accelerating to 8 and then 9 percent before starting to slow at the end of 2005. Karl Case and Robert Shiller (2003) report that the overwhelming majority of persons surveyed in 2003 agreed with or strongly agreed with the statement that real estate is the best investment for long-term holders. Respon - dents expected prices to increase in the future at 6 to 15 percent a year, depending on location. The continuous advance of nominal house prices has not always translated into real price increases, after taking into account general inflation. Figure 1 shows that, between 1975 and 1995, real home prices went through two cyclical waves: ris - ing after 1975, falling in the early 1980s, and then rising again before falling in the early 1990s. From 1975 until 1995, housing did increase faster than inflation, but not that much faster. After the mid 1990s, however, real house prices went on a sus - tained surge through 2005, making residential real estate not only a great investment, but it was also widely perceived as being a very safe investment.2 A variety of factors determine the demand for resi - dential housing, but three stand out as important in driving price increases. The first factor was just described. When prices rise, that can increase the pace of expected future price increases, making the effective cost of owning a house decline. The ex - pected capital gain on the house is a subtraction from the cost of ownership. As people witness price increases year after year — and witness those around them investing in homes — a “contagion” of expectations of future price increases can (and did) form and perpetuate price increases. The second is that when household income rises, this increase allows people to afford larger mortgages and increases the demand for housing. Over the period 1995-2000, household income per capita rose substantially, contributing to the increased de - mand. However, Figure 1 shows that the increase in house prices outpaced the growth of household income starting around 2000. One sign that house prices had moved too high is that they moved ahead much faster than real household income. People were stretching to buy houses.3 The third factor is interest rates. After soaring to double digits and beyond in the inflationary surge of the 1970s and early 1980s, nominal rates started to come down thereafter, and continued to trend down until very recently. Real interest rates (ad - justed for inflation) did not fall as much, but they fell also. From the perspective of the mortgage market, nominal interest rates may be more rel - evant than real rates, since mortgage approval typi - 2. The Case-Shiller Index is also widely used to measure housing prices. It has a broadly similar pattern to the one shown here, but does not go back as far historically. 3. The relation between household income and housing demand is not exact. See, for example, Gallin (2004). For a more in-depth and dis - aggregated look at the ratio of home prices to income over the past decades, see Case et al (2008). Shiller (2008) shows that for over 100 years (from as far back as 1880 to the early 1990s), house prices moved proportionally to fundamentals like building costs and population. The subsequent boom was out of line with each of these fundamentals.housing Demand and the Perception of Low Risk in housing investmentT hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 12 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on figuRE 1: Real home Prices and Real household income (1976=100); 30-year conventional Mortgage Rate Source: ohFeo; Federal reserve; Bureau of the Census. home Prices and Income are deflated by CPI less Shelter. cally depends upon whether the borrower will be able to make the monthly payment, which consists mostly of the nominal interest charge. Regardless, with both real and nominal interest rates lower than they had been for many years, the demand for mortgage-financed housing increased. Asset price bubbles are characterized by a self-rein - forcing cycle in which price increases trigger more price increases, but as the level of asset prices moves increasingly out of line with economic fundamen - tals, the bubble gets thinner and thinner and finally bursts. At that point the cycle can work in reverse as people hurry to get rid of the asset before prices fall further (see Box 1). This was the pattern of the dot com bubble of the late 1990s, when investors were enthralled by the promise of new technologies and bid up the prices of technology stocks beyond any reasonable prospect of earnings growth. There were some crashes of particular stocks and finally prices of most technology stocks plunged. In the case of the housing bubble, prices in some markets moved so high that demand was being choked off. Eventually, suspicions increased that price rises would slow down, which they did in 2005, and that prices would ultimately fall, which happened in 2007 according to both the Case-Shiller and the OFHEO indexes.4 The rise in housing prices did not occur uniformly across the country, a fact that must be reconciled with our story of the origins of the bubble. If there were national or international drivers of the price boom, why did these not apply to the whole mar - ket? In some parts of the country there is ample land available for building, so that as mortgage in - terest rates fell and house prices started to rise, this prompted a construction boom and an increase in the supply of housing. Residential housing starts in - creased from 1.35 million per year in 1995 to 2.07 1976 1980 1984 1988 1992 1996 2000 200480100120140160180200 024681012141618Index (1976 = 100) PercentAnnualized 30-year Conventional Mortgage Fixed Rate (Right Axis ) Mean Real Household Income (1976 = 100) Left Axi s Real Home Price Index (1976 = 100) Left Axi sT hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 134. The Case-Shiller index started to decline a little earlier than OFHEO and has fallen by substantially more. That is to be expected since the Case-Shiller 10-city index follows the markets that have seen big price declines. 5. As an illustration, Case et al (2008) show that the behavior of the ratio of home prices to per capita income varied substantially across cities, rising substantially in metropolitan areas like Miami and Chicago but staying relatively flat in cities like Charlotte and Pittsburgh. 6. Germany is the exception, where there was a huge building boom following reunification, resulting in an oversupply of housing.million in 2005, with 1.52 of the two million built in the south and west. Demand growth outstripped supply, however, in very fast growing areas like Las Vegas and in California and East Coast cities where zoning restrictions limited the supply of land. In the Midwest, there was only a modest run up in house prices because the older cities that were dependent on manufacturing were losing jobs and population. So the answer to the puzzle is that while the factors encouraging price increases applied broadly (espe - cially the low interest rates), the impact on prices and the extent to which a bubble developed also depended largely on local conditions.5An additional note on this issue comes from look - ing at other countries. The decline of interest rates was a global phenomenon and most of the advanced countries saw corresponding rises in housing pric - es.6 For example, home prices in the UK rose nearly 70 percent from 1998 to 2007. In some of these countries, there have been subsequent price declines, suggesting a price bubble like that in the U.S. In general, the experience of other countries supports the view that the decline in mortgage in - terest rates was a key factor in triggering the run up of housing prices (see Green and Wachter (2007)). T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 1 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on As the economy recovered from the 2001 re - cession, the expansion of mortgage lending was in conformable and other prime mort - gages, but as the boom proceeded, a larger frac - tion of the lending was for so-called “non-prime” lending that consists of subprime, Alt-A and home equity lending. The definition of what constitutes a “subprime” borrower is not precise, but it generally refers to a borrower with a poor credit history (i.e. a FICO score below 620 or so) that pays a higher rate of interest on the loan. Alt-A borrowers, deemed a bit less risky but not quite prime, had better credit scores but little or no documentation of income. Figure 2 illustrates the recent shift into “non- prime” lending. In 2001 there were $2.2 trillion worth of mortgage originations, with 65 percent of these in the form of conventional conforming loans and Federal Housing Administration (FHA) and Department of Veterans Affairs (VA) loans. An ad - ditional 20 percent were prime jumbo mortgages, The Shifting composition of Mortgage Lending and the Erosion of Lending Standards issued to those with good credit buying houses that were too expensive to be conforming, meaning that 85 percent of originated loans in 2001 were prime quality. There was a huge expansion of mortgage lending over the next couple of years, and in 2003 nearly $4 trillion worth of loans were issued, but the share of prime mortgages remained steady at 85 percent as the volume of conformable mortgages soared. The total volume of mortgage lending dropped af - ter 2003, to around $3 trillion a year in 2004-06 but the share of subprime and home equity lend - ing expanded greatly. Prime mortgages dropped to 64 percent of the total in 2004, 56 percent in 2005 and 52 percent in 2006, meaning that nearly half of mortgage originations in 2006 were subprime, Alt-A or home equity. It is clear that there was a significant change in lending patterns apparent in the composition of loans going back to 2004. T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 1figuRE 2: Total Mortgage Originations by Type: with share of each product; billions, percent Source: Inside Mortgage Finance. heL is home equity Loan.2001 2002 2003 2004 2005 2006 2007 2007Q4 (annualized )Total= 2,215 2,885 3,945 2,920 3,120 2,980 2,430 1,800 (annualized ) FHA/V A Conformin g JumboHEL Alt-A SubprimeT hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 16 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on The events leading up to the current crisis were very much in line with some common theories on how bubbles form. For example, Bikhchandani, Hirshleifer and Welch (1992) developed a theory on why ratio - nal people exhibit herding behavior that can lead to a bubble. Bikhchandani et al constructed a game theory model where individuals base their decisions both on their own judgment and on the actions of others. If an individual observes everyone around her choosing one way, she may conclude they are all correct, even if she herself may believe the opposite is true. The authors refer to this phenomenon – where by observing the actions of others, an individual discards her own judg - ment – as an “information cascade.” In a marketplace where individuals observe the actions of others, herding behavior may trump the judgment of rational individu - als. This kind of “social contagion” can go a long way in describing how homeowners, mortgage originators, holders of mortgage-backed securities, regulators, rat - ings agencies – indeed everyone – could get swept up in a bubble that ex post was clearly bound to burst. Another bubble theory that had received attention in the press was developed by the economist Hyman Minsky, who argued that financial markets are inherently un - stable, and he developed a theory of a bubble cycle that aptly describes the recent bubble in housing markets. Minsky theorized that a bubble had five steps. Step 1 was displacement: investors start to get excited about some - thing – whether it be dot-com companies, tulip bulbs in 17th century Holland, or subprime mortgages. Step 2 is a boom: speculators begin to reap high returns and see - ing their returns, more investors enter the market. Step 3 is euphoria: as more and more people crowd into the market, lenders and banks begin to extend credit to more dubious borrowers and lower lending standards (i.e. lend to borrowers with no documentation of income, or offer loans with high loan to value ratios), financial engineers create new instruments through which they can increase their exposure to the market (i.e. CDOs, CDS), and there is a general desperate surge by new participants to get “a piece of the action.” Indeed, Step 3 could be largely framed in terms of the “information cascades” and the herding behavior it entails.Step 4 is profit-taking: the bubble reaches its peak, and smart investors cash out of the market. This profit-taking unleashes the final step, which is Panic. Once the bubble begins to contract, pessimism immediately replaces exu - berance, and investors try to get rid of their now ill-fated assets as quickly as possible. In the context of the current crisis, banks see their asset values plummet and see their lenders refuse to rollover debt, forcing them to de-lever - age even further to make good on their liabilities. A so- called “Minsky moment” occurs when banks and lenders are forced to fire-sell even their safe assets in order to pay off their outstanding liabilities. Minsky went even further in a 1992 piece where he outlined his “financial instability hypothesis” and ar - gued market economies will inevitably produce bubbles. When times are good, banks will increase the riskiness of their assets to capture high returns, and they will borrow more and more to finance and increase the profitability of these assets. Minsky’s view is that financial markets are inherently unstable. There is, of course, an alternative, efficient markets view, which says that individuals are independent-minded in - vestors, and that asset prices reflect information that is known to everyone. It follows that the aggregate market is wiser than any one individual. In that view, excessive risk taking is not an inherent outcome of markets, but rather is a moral hazard problem that is the responsibility of government policies that insure deposits and bail out banks that get into trouble. While failures of govern - ment policy contributed to what happened, we judge that failures by private market participants were at the heart of this crisis, a viewpoint expressed by Alan Greenspan in Congressional testimony on October 23, 2008. Robert J. Shiller has studied speculative bubbles, analyz - ing stock market and other asset price cycles, based upon “irrational exuberance” in markets. He wrote about the risks of a real estate bubble well before the crisis hit and offers an analysis of the current crisis in Shiller (2008).BOx 1: The Mortgage Boom in the context of Theories of BubblesT hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 17The period of 2001-07 was one of rather modest growth in household income, but household con - sumption continued to grow as the personal saving rate, already low, continued to decline. Americans were tapping into the rising wealth they had in their homes in order to finance consumption. Greens - pan and Kennedy (2007) estimate that homeown - ers extracted $743.7 billion in net equity from their homes at the peak of the housing boom in 2005 — up from $229.6 in 2000 and $74.2 in 1991. The increase in house prices allowed a borrowing spree. The spree was largely financed by a boom in Home Equity Loans (illustrated in Figure 2) that allowed homeowners to borrow against the rising value of their home.7 In addition, there was an expansion of loans to lower-income, higher-credit risk families, including from the Government Sponsored Enter - prises, Fannie and Freddie, as they sought to expand home ownership for the benefits it brings in terms of sustaining neighborhoods. There was a deterioration in lending standards gen - erally dated to 2004 or 2005. Families that lacked the income and down payment to buy a house under the terms of a conforming mortgage were encour - aged to take out a mortgage that had a very high loan to value ratio, perhaps as high as 100 percent (often using second or even third mortgages), mean - ing that they started with no initial equity — and thus no true financial stake — in the house Such borrowing typically requires a rather high interest rate and high monthly payment, one that likely vio - lates the usual rules on the proportion of household income needed to service the debt. Originators got around this problem by offering Adjustable Rate Mortgages (ARMs), which had low initial payments that would last for two or three years, before reset - ting to a higher monthly amount. These so-called “teaser” interest rates were often not that low, but low enough to allow the mortgage to go through.8 Borrowers were told that in two or three years the price of their house would have increased enough to allow them to re-finance the loan. Home pric - es were rising at 10 to 20 percent a year in many locations, so that as long as this continued, a loan to value ratio of 100 percent would decline to 80 percent or so after a short time, and the household could re-finance with a conformable or prime jum - bo mortgage on more favorable terms. There is a lively industry in the United States that offers guides for people who want to make money by buying residential real estate and then re-selling it at a profit. The Miami condominium market was a favorite place for real estate speculation as inves - tors bought condos at pre-construction prices and then sold them after a short time at a profit. Specu - lative demand—buying for the purpose of making a short-term profit—added to overall housing de - mand.9 By their, nature fraudulent practices are hard to as - sess in terms of the volume of outright fraud, but based on press reports and interviews, it seems clear that shading the truth and outright fraud became important in the real estate boom (and in the sub - sequent bust). According to the Financial Crimes Enforcement Network, the number of reported cases of mortgage fraud increased every year since the late 1990s, reaching nearly 53,000 in 2007, compared with roughly 3,500 in 2000.10 Some bor - rowers lied about their income, whether or not they were going to live in the house they were buying, and the extent of their debts. Credit scores can be manipulated, for example, by people who become signatories on the credit accounts of friends or rela - tives with good credit ratings. Without having to make regular payments on a loan themselves, they 7. Indeed, Home Equity Loans have boomed since the 1980s, when banks first began to advertise them to homeowners as a way to “extract wealth” from their homes. See Louise Story, “Home Equity Frenzy was a Bank Ad Come T rue,” The New York Times ; August 15, 2008. 8. As mortgage rates are typically linked to the Federal Funds rate, the loose monetary policy during 2001-2004 helped keep these ARM rates down at an “unnaturally” low level. 9. Since pretty much anyone who buys a house factors in the expected capital gain on the house, everyone is subject to speculative demand. The reference here refers to people or companies that bought houses they did not intend to live in or use as vacation homes. 10. T aken from Barth and Yago (2008) T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 18 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on can acquire the high credit rating of the other per - son. Another fraudulent practice occurred with speculators. Mortgage lenders want to know if a household will actually occupy a house or unit be - ing purchased; or if it will be rented out or re-sold. This knowledge affects the probabilities of default or of early repayment, both of which can impose costs on the lender. We do not know how many delinquent mortgages are on properties that are not owner-occupied, but we have heard figures in the 40 to 50 percent range. Misrepresentation by borrowers and deceptive practices by lenders were often linked together. A mortgage broker being paid on commission might lead the borrower through an application process, suggesting places the borrower might change the answer or where to leave out damaging informa - tion. Sometimes the line will be fuzzy between a situation where broker helps a family navigate the application process so they can buy a house they really can afford, and a situation where the broker and the applicant are deliberately lying. Looking at the data, the deterioration in lending standards over the course of the boom is remark - able. The share of subprime loans originated as ARMs jumped from 51 to 81 percent from 1999 to 2006; for Alt-A loans, the share jumped from 6 to 70 percent during the same time period. A similar deterioration happened in combined loan to value ratios (the CL TV combines all liens against a property): the average CL TV ratio for originated subprime loans jumped from 79 to 86 percent. Fur - thermore, the share of full-doc subprime origina - tions fell from 69 to 58 percent; for Alt-A loans it dropped from 38 to 16 percent.11 Figure 3 provides further illustration of the shift into riskier lending as the boom progressed. It shows the proportion of mortgage originations for home purchase that were made based on interest only or negative amortization loan provisions (“re -fis” are excluded from this data). Someone borrow - ing with an interest-only loan pays a slightly lower monthly payment because there is no repayment of principal. Since the principal repayment in the first few years of a mortgage are usually very small, this is not a big issue in the short run, although the impact mounts up over the years. A negative amortization loan goes even further, and borrow - ers do not even pay the full amount of the interest accruing each month, so the outstanding balance rises over time. Such a mortgage might make sense for families whose incomes are rising over time and where home prices are rising, but it adds a signifi - cant amount of risk for both borrower and lender. In summary, the boom in mortgage borrowing was sustained by low interest rates and easier lending practices. As households cashed in the wealth in their property for consumption, less credit-worthy families were able to buy houses, and speculators purchased property in hopes of making money by reselling them. The increasingly lax lending stan - dards are characteristic of classic behavior during bubbles. Fraud, lack of due diligence, and deceptive practices occurred on both sides of the mortgage transactions, but as long as house prices continued to rise at a good pace, the whole structure could continue, and even the fraud and deception were buried as people were able to refinance and were unlikely to default on their mortgages and lose the equity (if they had any) that they had built up. With the benefit of hindsight we can look back and see that some of the innovative mortgage products have contributed to the default mess we have now. However, we would like to note that this analysis is not meant to be construed as a call to restrict finan - cial innovations. There were substantial benefits to those who used the products properly. Y oung families often face a tough situation in trying to buy a home. They are at an early stage in their careers, earning moderate incomes while they have the expenses of young children. Owning a home 11. Data taken from Ashcraft and Schuerman (2008). The drop in the share of full-doc loans for Alt-A loans is relatively unsurprising, as Alt-A loans were by definition made to borrowers with little or no documentation of income.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 1figuRE 3: interest-Only and Negative amortization Loans, Share of Total Mortgage Originations Used to Purchase a Home (excludes refis): 2000-2006; percent Source: Credit Suisse (2007), LoanPerformance in a good neighborhood with good schools is a very desirable and natural wish, but many families lack the down payment necessary and the monthly mortgage payment may be out of reach, especially in high-cost regions such as California or the East Coast. Based on their expected lifetime family in - come, they can afford a house, but at this early stage of their life-cycle, they are liquidity constrained. Some such families rely on older family members for help, but not all can do this. Mortgages with low payments for the first few years and low down payments provide a way to deal with this problem. Lending standards need to be restored to sanity in the wake of the mortgage crisis, but that should not mean, for example, the abolition of adjustable rate mortgages or low down payments for borrowers with the right credit.2000 2001 2002 2003 2004 2005 200605101520253035 21462529 23T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 20 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on The legal and institutional arrangements that prevail in the U.S. housing market produced a pattern of incentives that contributed to what happened. First, there are important protections given to households. These vary by state, but in many states it is possible to repay a mortgage early without penalty. This option meant that house - holds were encouraged to take out mortgages with terms that looked good in the short run, but were unfavorable in future years. They expected to refi - nance later on better terms, and without incurring a pre-payment penalty. In some states the mortgage contract is “without recourse to the borrower,” meaning that if a house - hold stops paying on a mortgage and goes into de - fault, the lender can seize the house (the collateral on the loan) but cannot bring suit to recover losses that are incurred if the sale of the property does not yield enough to pay off the mortgage and cover the selling and legal costs. In principle, this encourages households to walk away when they are unable or unwilling to cover a mortgage payment. This can be an important protection for families facing un - employment or unexpected medical expenses, but it can lead to abuse by borrowers and encourage over- borrowing. In a significant percentage of defaults in the current crisis, borrowers are simply mailing in the keys to the house and are not even contact - ing the lender to try and work out a settlement that would avoid default. There is debate about the importance of this issue. On the one hand, there are reports that the states that have had the most problem with mortgage defaults are the ones that are non-recourse to the borrower. On the other hand, lenders rarely find it profitable to pursue de - faulting borrowers—big bank suing poor family in trouble is not a situation most banks want to take to a court. The most perverse incentive in the mortgage origi - nation market though, is the ability of originators to immediately sell a completed loan off their books to another financial institution. Currently, most mort - gage loans are originated by specialists and brokers who do not provide the funding directly. One insti - tution provides the initial funding of the mortgage but then quickly sells it off to another financial in - stitution, where either it is held on a balance sheet or packaged with other mortgages to be securitized (see below).12 The key issue here is that the institu - tion that originates the loan has little or no financial incentive to make sure the loan is a good one. Most brokers and specialists are paid based on the volume of loans they process. They have an incentive to keep the pace of borrowing rolling along, even if that meant making riskier and riskier loans. Mian and Sufi (2008) provide evidence that many of largest increases in house prices 2001-2005 (and subsequently large crashes in prices and foreclo - sures 2005-07) happened in areas that experienced a sharp increase in the share of mortgages sold off by the originator shortly after origination, a process they refer to as “disintermediation” (but is synony - mous to the first stage of securitization, which we discuss shortly). These areas were also characterized by high “latent demand” in the 1990s, meaning that a high share of risky borrowers had previously been denied mortgage applications. The “disintermedia - tion” process, by allowing originators to pass off the risk of their loans, increased the supply of credit and encouraged them to lend to risky borrowers who previously were ineligible for loans (the authors also find that these areas experienced relatively high de - linquency rates once house prices began to fall after 2004). Thus, by increasing the availability of credit to riskier borrowers, disintermediation increased housing demand and house prices during the boom Economic incentives in the housing and Mortgage Origination Markets 12. Mortgage sales contracts often allowed the buyer to “put” back the mortgage to the seller for a limited period, a year or two. But in an era of rising housing prices and thus low delinquencies, originators did not view these “puts” as a serious risk.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 21years. The authors find that some of these areas that experienced high house price appreciation did so despite experiencing negative relative income and employment growth over the period. Miam and Sufi (2008) thus show empirically that the abil - ity to securitize subprime mortgages was key factor in inflating the housing bubble. The adverse incentives in the originate-to-distrib - ute model for mortgages occur in other markets where there is asymmetric information—when one party to the transaction knows more than the other. For example, most drivers know little about me - chanical issues, so when they have a problem with their car they take it to an auto mechanic. That me - chanic will know much more about the cause of the difficulty than the owner, so he or she can tell the owner that there are expensive problems that must be fixed, even if that is not the case. The mechanic has an economic incentive to exaggerate problems in order to make a profit on the repair. This does not necessarily tell you that there is a market fail - ure, however, because there are market responses to information asymmetries—people in business for a long time want to develop a reputation for honesty and reliability. Publications like Consumer Reports or services like Angie’s List can be used to find qual - ity products and services. In the mortgage origina - tion market, there were similar market responses to the asymmetric information. There were provisions intended to provide information to and protect the interests of the ultimate holders of the default risk. For example, anyone selling a mortgage loan had to provide information on the credit score of the borrower, the loan to value ratio, and other infor - mation that the buyer of the mortgage could use to assess its value. Many of the originating finan - cial institutions had been providing mortgages for many years and had built up reputations for sound practices. Unfortunately, the market responses to asymmetric information in the mortgage market did not solve the problem. It is somewhat puzzling why this was the case in the secondary market where mortgages were re-sold. One would have expected that the in - stitutions that ultimately ended up with the default risk knew about the incentive problems in the origi - nation process and would have taken the necessary steps to counteract them. It is hard to get a full answer as to why they did not, but the key issue is the one given earlier. The long upward movement of house prices convinced nearly all stakeholders that these prices had nowhere to go but up, so the level of monitoring and the standards of lending in mortgage origination eroded. Default rates had remained low for many years and so there did not seem to be much risk involved. Another issue, as we will discuss below, is that the securitization process created an enormous gap between the origination of the loan and the investors who ultimately held the underlying risk, making sound risk analysis ex - tremely difficult.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 22 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on In the old model, mortgage loans were made by Savings & Loans institutions (S&Ls) and the funds for them came from the savings deposits of retail customers. The S&Ls themselves vetted the mort - gages and took on the three risks involved: the risk of default; the risk of pre-payment (which reduces returns); and the risk of changes in interest rates. By keeping a stake in the health of their loans, origina - tors had a financial incentive to monitor their quality and investigate whether or not the borrower could feasibly repay the mortgage. However, it was also quite expensive for these institutions to keep loans on their books, and it limited the volume of loans they could originate. This system broke down in the S&L crisis of the mid- 1980s for complex reasons that link to the era when financial institutions and interest rates were much more heavily regulated.13 T o oversimplify, the cri - sis stemmed from both interest rate risk and default risk. As market interest rates rose, the S&Ls had to pay higher rates on their deposits but could not raise the rates on their stock of mortgages by enough to compensate. They tried to avoid insolvency by in - vesting in much riskier assets, including commercial real estate that promised higher returns but then suf - fered serious default losses. Because of regulations limiting interstate banking, the mortgage portfo - lios of the S&Ls were geographically concentrated, which made them riskier—the residential mortgage markets in T exas and California suffered high default rates in the 1980s. There were also some fraudulent practices at that time; for example in the Lincoln Savings collapse, the CEO Charles H. Keating was convicted and served time in jail. In response to the losses in the S&Ls, the federal government created Securitization and the funding of the housing Boom the Resolution T rust Corporation to take the assets off the banks’ books, and then sold them off. In the process, there were large losses that were covered by taxpayers — roughly $150 billion. Securitization was seen as a solution to the problems with the S&L model, as it freed mortgage lenders from the liquidity constraint of their balance sheets. Under the old system, lenders could only make a limited number of loans based on the size of their balance sheet. The new system allowed lenders to sell off loans to a third-party, take it off their books, and use that money to make even more loans. The Government Sponsored Enterprises (GSEs), notably Fannie Mae and Freddie Mac, were created by the federal government in 1938 and 1970, respectively, to perform precisely this function: the GSE’ s bought mortgage loans that met certain conditions (called “conforming loans”) from banks in order to facilitate mortgage lending and (theoretically) lower mortgage interest rates.14 The GSEs initially funded their mortgage purchases by issuing bonds, but they were pioneers in securiti - zation — or where a pool of geographically dispersed mortgages is re-packaged and sold as mortgage- backed securities (MBS) to investors (see box 2 be - low). Freddie Mac issued the first ever modern mort - gage backed security in June 1983. The returns of an MBS reflect the returns on the underlying mortgage pool. Those who held the GSE-issued MBS took on some of the risks, notably the interest rate risk. Importantly, however, the GSEs retained the default risk of the mortgages that underlined the MBS they sold. They guaranteed investors against default losses and pre-payment losses (by including a guarantee fee 13. One of these was the result of regulation (Regulation Q) that limited the interest rate that S&Ls could pay on their deposits and led de - positors to withdraw funds when market rates rose. That regulation, in an era of double digit market interest rates, exposed the thrifts to a massive potential outflow of funds in the 1979-1981 period, which was avoided when Congress lifted Regulation Q. But even after this occurred, the loss in asset value on the S&Ls balance sheets meant that most had little or no capital at risk. 14. There are different estimates of the extent to which the GSEs provided lower interest rates for borrowers. Most suggest the impact on mortgage rates is fairly small. See Passmore, Shurland and Burgess (2006), for example. Presumably without the GSEs, other financial institutions would have had a bigger role.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 23in the price of the MBS), or at least losses above an expected amount built into the rate of return of the MBS when it was issued. Investors in GSE-issued MBS were thus shielded from the default risk of the underlying loans. The GSEs could then either sell the MBS on the open market, or they could issue their own bonds, use the revenue to buy the MBS and hold them on their own books. They could also buy MBS issued by private in - stitutions to further increase the size of their books. They earned a profit because they earned a higher in - terest return on the mortgage assets than they would pay on the bonds that they have issued. This has some similarity to the S&L model, except that Fannie and Freddie can hold much larger pools of mortgages that are geographically dispersed. In addition, the GSEs were seen as implicitly guaranteed by the federal gov - ernment (a guarantee that has since become explicit) so they paid only a few basis points above T reasury yields on their bond issuance. This implicit government backing lowered their cost of borrowing and allowed them to inflate their balance sheets enormously. Over the years, this line of business was very profitable for the GSEs, and the size of their internally-held mort - gage portfolios ballooned until they faced regulatory restrictions pushed by Alan Greenspan, then Federal Reserve Chairman, and others. The GSEs have been major participants in the mort - gage market and by 2008, Fannie and Freddie held or guaranteed $5.4 trillion in mortgage debt. The T rea - sury was forced to nationalize them in September 2008 and guarantee their liabilities because they would oth - erwise have been driven into bankruptcy. Fannie and Freddie combined had nearly $5.5 billion in losses in the first two quarters of 2008, according to their statements. How did they get into trouble? Mostly because they behaved like so many other people and believed that default rates were stable and predictable and that, at most, there would be only regional price declines and not national price declines. When the price bubble burst, they faced much higher default rates than expected and they did not have enough capital to cover their losses. Their unstable “govern - ment sponsored” status allowed them to skirt around capital requirements, and they became overleveraged – indeed their leverage ratio in 2007 was estimated to be over twice that of commercial banks.15 In part, their problems also came from their efforts to meet the affordable housing goals set by Congress. Congress pushed them to provide more loans to low- income borrowers to justify the capital advantage they had because of the implicit federal guarantee. The rules under which they operated required that they not buy subprime whole loans directly. But they faced no limits on the amount of subprime MBS they could buy from private issuers that they then kept on their books. Indeed, the two of them bought between $340 and $660 billion in private-label subprime and Alt-A MBS from 2002-2007.16 The losses they now face on their mortgage portfolio include both prime mortgages and the lower quality mortgages on their books. House prices have fallen so much that even many prime mortgages are defaulting. Many have pointed to the GSEs as one of the main, culprits in the financial crisis because the implicit government guarantee allowed them to inflate their balance sheets by borrowing at below-market rates. Is this perception correct? Starting in 2004, they did begin to buy riskier loans in the face of pressure from Congress, but this was late in the game, after private subprime lending had already taken off. Further, while the GSEs purchased private-label subprime MBS to hold on their books, they by no means “led the charge.” For example, in 2002 Fannie Mae pur - chased just over 2 percent of private-label subprime and Alt-A MBS. In 2004, once the market was al - ready booming, it bought 10 percent of the total, and in 2007 it bought 4.5 percent.17 Fannie and Freddie did not catalyze the market for subprime MBS; rath - er, they started to hold such mortgages in the pools they purchased, perhaps because of shareholder pres - sure or to regain market share. 15. Greenlaw et al (2008), page 35. 16. OFHEO (2008). The wide range is because data for Freddie Mac’s purchases of subprime and Alt-A MBS only goes back to 2006, so its purchases are estimated 2002-2005. 17. The data for Freddie Mac’s purchases of subprime MBS does not go back as far, but it is probable that Freddie played a bigger role than Fannie in the market. In 2006 and 2007, for example, Freddie bought 12 percent of all subprime MBS issued in those years.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 2 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on Figure 4 illustrates the way that MBS repackaged mortgage loans in order to increase the funds avail - able to the mortgage market, as well as to generate fees for the re-packagers. While some of the underlying mortgages would inevitably default, they are selected from geographically diverse areas which, it was once believed, would protect the health of the overall pool from any local default shocks; prior to the current tur - moil in housing markets, there had never been a hous - ing downturn on a national scale. Still, an asset based on a simple pool of subprime mortgages would carry a credit rating below or well below AAA. Rather than sell one asset based on the entire pool, though, an MBS issuer could issue securities with vary - ing risk and return by tranching the securities into different groups based on exposure to the underlying risk of the pool. After buying the receivables of thou - sands of mortgage loans, an issuer then transfers them to what is called a Special Purpose Vehicle (SPV), an off-balance sheet legal entity, which “holds” the re - ceivables in a pool and issues the securities. The se - curities are typically separated into senior, mezzanine (junior), and non-investment grade (equity) tranches. A senior tranche has preferred claim on the stream of returns generated by the mortgages; once all the senior tranche securities are paid, the mezzanine holders are paid next, and the equity tranche receive whatever is left. A portion of the mortgages can go into delin - quency, but various forms of protection should mean there is still enough income coming into the pool to keep paying the holders of at least the senior tranche. Thus, the holders of the senior tranche have an asset that is less risky than the underlying pool of mortgages, and they were deemed so safe that credit rating agen - cies were willing to give them AAA ratings.The safety of a senior tranche, or any tranche, mainly depends on two concepts (other than the health of the mortgage loans themselves): the degree of subordina - tion under it and the level of credit enhancement in the MBS.18 Subordination of a tranche refers to the to - tal size of the tranches junior to it. The higher the subordination, the safer the tranche. If, for example 75 percent of a set of MBS is senior, then the senior tranche benefits from 25 percent of subordination, plus any over-collateralization.19 Over-collateralization, or when the face value of the mortgage assets in the pool is higher than the face value of the re-packaged securi - ties, is a form of credit enhancement used to reduce the exposure of the debt investors to the underlying risk of the pool. The over-collateralized part of the MBS is the “equity” tranche, as its holders are the first to lose money in case of default and receive whatever money is “left over” if there are below-than-expected defaults. If, for example, 1.5 percent of an MBS is equity, then 1.5 percent of mortgage payments can default before the most junior debt tranche incurs any losses. Another important form of credit enhancement is “excess spread,” whereby the total incoming interest received from the mortgage payments exceeds the pay - ment made to senior and junior debt holders, fees to the issuer, and any other expenses. This is the first line of defense in terms of protection, as no tranche incurs losses unless total credit defaults become high enough to turn the excess spread negative. (If this does not hap - pen, the equity tranche gets whatever excess spread is left over). The repackaging of MBS into tranches does nothing to reduce the overall risk of the mortgage pool, rather it rearranges it. The senior tranches are less risky and eligible for high investment grade credit ratings, as BOx 2: The anatomy of an MBS 18. There exists much literature explaining MBS structure; for a more in-depth and very elucidating description see Ashcraft and Schuermann (2008) or Gorton (2008). 19. Senior tranches of subprime MBS were typically more subordinated and those in Alt-A or prime MBS to compensate for the higher risk of the underlying pools.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 2they are (theoretically) quite insulated from the de - fault risk. On the other hand, the lower tranches are much more risky and can face losses very quickly; the equity tranche has the potential for huge returns when defaults are low but are also the first to be wiped out when the default rate hits even a small amount above what is expected. T ranching redistributes the risk ac - cording to risk appetite of investors: senior tranches pay a lower yield but are safer bets, and the junior tranches pay a higher yield and are riskier. However, effective tranching of risk rests on the as - sumption that proper risk analysis is performed on the underlying assets. Since 2007, many previously AAA- rated securities have been downgraded, reflecting the fact that all stakeholders underestimated the true risk in these securities. As a result, many MBS holders that were previously considered relatively insulated are now getting wiped out. The idea of taking risky assets and turning them into AAA-rated securities has been received with scorn by many as the mortgage market has slumped. And with good reason, in the sense that the riskiness of these se -curities was in fact much higher than their ratings sug - gested, because the overall market slump resulted in a correlated wave of defaults. But this financial alchemy is not as strange as it seems; in fact it has been around for a long time in other markets. A public company is an asset with an uncertain stream of returns. T ypically, the claims on that income are assigned to two broad groups, the bond holders and the stock or equity hold - ers. The company’s bonds may well be of low risk and eligible for a high credit score. The bond holders get first dibs on the returns of the company and the equity holders get what is left over. Most large companies effectively tranche their liabilities into bonds with dif - ferent seniorities in terms of claims on the company’s income, and they may have different classes of equities, too. In short, the idea of different tranches of assets with differing risk levels is not at all new and there is nothing inherently wrong with it. The goal is to pro - vide investors with different risk and return options and to let investors with an appetite for risk absorb that risk. The repackaging did not stop there, however. There were second and third rounds of securitization, and the trouble that emerged there was worse.figuRE 4: anatomy of a MBS T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 26 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on Other financial institutions also issued MBS, but because of the capital advantage of the GSEs, these institutions operated in the “jumbo” market for loans that were for larger amounts than the GSEs were allowed to buy, and more recently especially in the subprime and Alt-A market. In the recent boom years since 2000, securitization through pri - vate financial institutions exploded, and the GSEs increasingly lost market share to “non-agency,” or private, MBS issuers. T o illustrate: in 2000 MBS issued by the GSEs made up 78 percent of total MBS issued in that year. By 2006 their share of MBS issuance had dropped to 44 percent.20 The list of the top subprime and Alt-A MBS issuers in 2006 includes such ill-fated names as Lehman Brothers, Bear Stearns, Countrywide, Washington Mutual, and Merrill Lynch (whose fates, among others, we will return to in a future report). As securitization became more widespread, and as the subprime 20. Inside Mortgage Finance 2008 Mortgage Market Statistical Annual; authors’ calculationsfiguRE 5: Securitization Rates by Type of Mortgage, 2001 and 2006; percentmortgage market boomed, private banks, broker dealers, and other institutions increasingly domi - nated the MBS market. Figure 5 illustrates the growing importance of securitization, showing the rates in 2006 for con - forming, prime jumbo and subprime / Alt-A loans, for which securitization rates reached 81, 46 and 81 percent, respectively Securitization was already well established among conforming loans, as the GSEs had been securitizing them for two decades; 72 percent of conforming loans were securitized in 2001. The real boom in securitization since 2001 came from subprime and Alt-A loans, as the share of these loans that were securitized had jumped 75 percent since 2001. By 2006, securitization was fund - ing most of the mortgage loans in the lower rated catego - ries — the loans that are in trouble now. Source: Inside Mortgage FinanceConforming Prime Jumbo Subprime/Alt-A0255075100 2001 2006 2001 2006 2001 2006T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 27As noted, while the GSEs dominated the se - curitization market during the 1980s and 1990s, by 2000 they began losing market share to private financial institutions as more and more subprime mortgages began to be securitized. As the securitization market came to be dominated by the financial sector, it grew more complex, and more opaque. Not only did the market become risk - ier and less transparent, but it shifted into a financial world that was unregulated and little understood. As banks, brokers, hedge funds, and other institu - tions utilized new financial innovations to maximize their exposure to these products, they fuelled the demand for risky mortgages and inflated the bubble that ultimately burst in August 2007. As discussed above, securitization has been an ex - tremely positive innovation for credit markets. By allowing banks to sell whole loans off their books, and by distributing risk according to the risk ap - petite of investors, it (presumably) has lowered the cost of lending for all and facilitated the extension of credit to new borrowers who otherwise would be shut out of credit markets.21 However, as the mar - ket became increasingly opaque and complex, new instruments based on technical computer models were wildly traded by highly leveraged institutions, many of whom did not even understand the under - lying models. In good times, these arcane instru - ments were sources of enormous profits, but their complexity and the lack of any serious infrastruc - ture and public information about them created a massive panic in the financial system that began August 2007. One of the central reasons the current crisis has been so severe (and that the bubble inflated so enor - mously) was that much of the subprime mortgage exposure has been concentrated in the leveraged fi - nancial sector. The term “leverage” typically refers to the use of borrowed funds to magnify returns on any given investment. If asset prices are rising, and the cost of borrowing is low, then banks will natu - rally try to maximize their exposure to rising asset prices by borrowing as much as they can. While borrowed funds are central to the concept of “lever - age,” its definition can expand to any instrument through which a bank can magnify its exposure to a given asset. We discuss such instruments below. collateralized Debt Obligations As the securitization of mortgages increasingly became an affair of the private financial sector, it spurred further innovation in products that in good times generated large profits, but have also been the source of some of the biggest losses since the crisis unfolded in 2007. Collateralized Debt Obligations (CDOs) represented a further step into the brave new world of securitization that really exploded af - ter 2000. CDO issuers purchased different tranch - es of MBS and pooled them together with other asset-backed securities (ABS). The other ABS were largely backed by credit card loans, auto loans, busi - ness loans and student loans. A “senior” CDO was made up predominantly of the highly rated tranches of MBS and other ABS, while “mezzanine” CDOs pooled together a higher share of junior tranches. Unlike an MBS, whose assets consisted of actual mortgage payments, a CDO’s assets were the se - curities that collected those mortgage payments; in a sense CDO’s “re-securitized” existing securities. Figure 4 would look very much the same to de - scribe a CDO rather than an MBS. Indeed, a CDO essentially “re-applied” the structure of an MBS. A CDO could thus further re-distribute the risk of its assets by re-tranching and selling off new securi - ties. In a seemingly miraculous form of “ratings arbitrage,” a mezzanine CDO could pool together low-grade junior tranches of MBS and other ABS and could convert some of them into new senior AAA-rated securities. The payment stream of an AAA-rated tranche of a mezzanine CDO was thus based on junior-rated MBS and ABS. More Securitization and More Leverage— cDOs, S iVs, and Short-Term Borrowing 21. For a more technical explanation of structured finance, see Ashcraft and Schermann (2008) or Gorton (2008).T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 28 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on The issuers worked directly with ratings agencies to structure the CDO tranches so that they could op - timize the size of highly-rated tranches in order to lower the funding costs of the CDOs; since the cou - pon rate on AAAs is lower than those on A- or BBB, it costs less to issue a highly-rated security than a lower one. Naturally, an issuer wants to maximize the size of the senior tranche so as to lower the cost of funding. However, the higher the share of se - nior tranches, the lower the subordination and thus protection of those tranches. As an additional pro - tection, CDO issuers would purchase credit default swaps (CDS) or credit insurance to raise ratings on the securities they issued and to shield the AAA tranches from the default risk (see discussion below). However, when a wave of CDO downgrades hit in 200722, many previously highly-rated tranches be - came exposed to losses. In practice, therefore, the reduced net risk exposure that CDOs appeared to embody was mostly illusory and, importantly, this second round of securitization made it even more difficult for investors to determine what risks they were actually taking. The first CDO was created in 1987 by the now- defunct Drexel Burnham Lambert, but this security structure was not widely used until the late 1990s when a banker at Canadian Imperial Bank of Com - merce first developed a formula called a Gaussian Copula that theoretically could calculate the prob - ability that a given set of loans could face correlated losses.23 Annual CDO issuances went from nearly zero in 1995 to over $500 billion in 2006. As CDO issuances grew, so did the share of them that was de - voted to mortgages: Mason and Rosner (2007) tell us that 81 percent of the collateral of CDO’s issued in 2005 were made up of MBS, or about $200 bil - lion total Thus, during the last several years of the housing bubble, CDOs increasingly funded mort - gage loans, especially subprime ones. Indeed, Mason and Rosner (2007) go even further to explain the insight that CDOs added significant liquidity to, and thus helped fuel the demand for, subprime mortgages and MBS. They estimate that in 2005, of the reported $200 billion of CDO col - lateral comprised of subprime MBS assets issued in that year, roughly $140 billion of that amount was in MBS rated below AAA (i.e. “junior” tranches). They then use figures from the Securities Indus - try and Financial Markets Association to estimate that roughly $133 billion in “junior” tranche MBS were issued in 2005. Thus, CDOs purchased more “junior” tranche MBS in 2005 than were actually issued that year! While these estimates are not pre - cise, they make the clear case that CDOs provided nearly all the demand for lower-grade subprime MBS during the later boom years, and in so doing provided a critical credit source for subprime mort - gages, fueling demand and inflating the bubble.24 Structured investment Vehicles and Off- Balance Sheet Entities One of the constraints on banks and some other institutions is that they must meet capital require - ments, that is to say, they must fund a given percent - age of their assets with shareholders’ capital rather than with some form of debt. Capital requirements for banks are mandated jointly by the FDIC, the Comptroller of the Currency, and the Federal Re - serve. As we will discuss in a forthcoming report, since 1989, when the international Basel Accord went into effect, U.S. banks have had to meet both the Basel requirement and a separate U.S. standard. Capital requirements lower the profitability of the 22. Moody’s (2008a) reports that of the CDOs it rated, a record 1,655 were downgraded in 2007 – 10 times the amount downgraded in 2006. 23. For a very interesting discussion of this formula, and its implications for the recent explosion in CDO issuances, see Mark Whitehouse, “Slices of Risk: How a Formula Ignited Market that Burned Some Big Investors; Wall Street Journal, September 12, 2005. 24. A technical fact that further illustrates the degree to which CDOs fueled demand for subprime MBS comes from the financing structure for securitized products. Mason and Rosner (2007) explain that while a typical MBS consists of 90 percent senior tranches and only 10 percent junior tranches, the junior tranches must be sold first before any of the senior tranches can be sold. Thus, presumably an MBS issuer cannot sell any AAA-rated tranches from a pool of mortgages before it gets rid of the lower-grade tranches first. By seemingly providing the sole demand for junior tranches, CDOs thus added the liquidity necessary to sell the entire MBS structure.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 2banks, since they limit the extent to which banks can leverage any initial shareholder investment (plus ac - cumulated retained earnings). Naturally, therefore, banks looked for ways to circumvent the require - ments. The favored means of getting around these mandated capital requirements became what were known as Structured Investment Vehicles (SIVs), an off-balance sheet SPV set up by banks to hold MBS, CDOs and other long-term institutional debt as their assets.25 By dodging capital require - ments, SIVs allowed banks to leverage their hold - ings of these assets more than they could on their balance sheets. T o fund these assets, the SIVs issued asset-backed commercial paper (ABCP) and medi - um term notes as their liabilities, mostly with very short-term maturity that needed to be rolled over constantly. Because they obtained the legal title of “bankruptcy remote,” SIVs could obtain cheaper funding than banks could, and thus increased the spread between their short-term liabilities and long-term assets — and for awhile they earned high profits. SIV assets reached $400 billion in July 2007 (Moody’s 2008b). Until the credit crunch hit in August 2007, this busi - ness model worked smoothly: a SIV could typically rollover its short term liabilities automatically. Li - quidity risk was not perceived as a problem, as SIVs could consistently obtain cheap and reliable fund - ing, even as they turned to shorter term borrowing (see Figure 6). T echnically, the SIVs were separate from the banks, constituting as a “clean break” from a bank’s balance sheet as defined by the Basel II Ac - cord (an international agreement on bank supervi - sion and capital reserve levels), and hence did not add to the banks’ capital or reserve requirements. Once the SIVs ran into financial trouble, however, the banks took them back onto their balance sheets for reputational reasons, to avoid alienating inves - tors and perhaps to avoid law suits.26Leverage and the Push To Short-Term Borrowing The increase in leverage over the course of the subprime bubble was widespread, spanning across many financial institutions and across many forms of instruments. This increase in leverage, as well as the growth in aggregate liquidity, was linked to the prolonged rise in house prices and asset prices across the board. Adrian and Shin (2007) illus - trate the perhaps counterintuitive, but extremely important, empirical insight that when financial institutions are forced to mark-to-market, mean - ing that they must assign a value to an asset based on its current market valuation, rising asset prices immediately show up on banks’ balance sheets, which increases the banks’ net worth and directly reduces their leverage ratio. If banks were passive, their total leverage would fall. However, financial institutions are far from passive; when asset prices are rising it is highly unprofitable for a bank to be “under-leveraged” and they will look for ways to utilize their new “surplus capital.” This search to utilize surplus capital means banks will look to fur - ther expand their balance sheet and increase their leverage. This phenomenon, for which the authors provide empirical evidence, leads to an expansion in aggregate liquidity and aggregate leverage in the financial system. As the authors put it on page 31, “Aggregate liquidity can be seen as the rate of growth of aggregate balance sheets.” In the context of the housing bubble, a feedback loop was created as the sustained rise in asset prices in mortgage-related products increased the net worth of banks, which, in turn, fueled the search for more leverage and further increased the de - mand for these assets. When the crisis hit asset prices plummeted, and the feedback loop worked in the opposite direction as leveraged institutions 25. IMF (2008) cites Standard and Poor’s to estimate that close to 30 percent of SIV assets were MBS as of October 2007, with 8.3 percent in Subprime MBS; 15.4 percent was in CDO’s. 26. The seeming contradiction that a SIV could be considered a “clean break” from a bank’s balance sheet, yet the bank could still act as the “bailout of last resort,” was made possible by a legal footnote called “implicit recourse” outlined in the Basel II Accord that says a sponsor - ing bank may provide support to a SIV that exceeds its “contractual obligations” to preserve its “moral” standing and protect its reputa - tion.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 30 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on found themselves exposed with very little capital and sharply increased leverage and were forced to shrink their balance sheets. This loop contributed to the “freezing up” of liquidity in credit markets. Investment banks were not supervised like deposit- taking commercial banks and did not have the same capital requirements, thus they were able to increase leverage to a greater extent. Nor were investment banks subject to the regulatory restrictions that ac - company the capital requirements. Institutions such as Bear Stearns and Lehman Brothers borrowed at very short term and held risky longer-term assets, with low levels of capital or reserves to cover chang - ing market conditions. Greenlaw et al (2008) cal - culate that while commercial banks are on average leveraged 9.8:1, broker/dealers and hedge funds are leveraged at nearly 32:1 (the GSEs were leveraged at 24:1 even though they were regulated). One of the favorite instruments of short-term bor - rowing for investment banks became the overnight repurchase agreement, or “repo loan” (See Morris and Shin 2008 for an insightful discussion). Over - night repos are a form of “collateralized borrowing” whereby a bank pledges its assets as collateral in an overnight loan with another bank. T o oversimplify, Bank 1 sells a portion of its assets to Bank 2, with the understanding that it will buy back the assets the next day at a slightly higher price. This process was deemed a low credit risk during the good times, but had profound systemic implications because it connected financial institutions to each other so that when one got into trouble, its problems spread to the other institutions with which it was trading. Overnight repos became an increasingly important source of funding for investment banks. Brun - nermeier (2008) shows that from 2001 to 2007, overnight repos as a share of total investment bank assets grew from roughly 12 percent to over 25 per - cent. That is, by 2007, investment banks were roll - ing over liabilities equal to one quarter of their balance sheet overnight. Figure 6 shows another example of the rapid in - creases in short-term borrowing, with maturity as low as one day that occurred as the boom peaked in 2006 and early 2007 in Asset-Backed Commer - cial Paper markets. As discussed above, ABCP is issued by off-balance sheet entities like SIVs to fund long-term assets. Like repos, ABCP was a form of “collateralized borrowing,” meaning that the issuer put up a certain value of its assets as collateral for the paper it issued. As many large banks set up off- balance sheet entities to escape regulatory scrutiny, ABCP became an important source of funding for many large institutions. Figure 6, though, illustrates the striking fact that the growth in ABCP issuance since around 2004 was nearly entirely in extremely short-term paper with maturity between 1 and 4 days. Overnight ABCP , like repos, increasingly be - came a way for banks to rely on shorter and shorter term borrowing to fund their assets. This source of funding was cheaper than longer-term borrow - ing, and until August 2007, it could be rolled over like clockwork. The drying up of these short-term funding markets has been an important element in the financial crisis since August 2007. When short-term liquidity funding like ABCP and repos suddenly dried up, financial institutions effectively faced a “run” and found themselves exposed with very little capital.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 31figuRE 6: Total Value of asset-Backed commercial Paper issuance by Date of Maturity, Daily 30-day Moving average since february 2001 in billions Source: Federal reserve In summary: the potential advantages securitization offers are that it allows loanable funds to shift easily among regions and even countries; and it distributes risk to lenders most willing to bear it, which reduces the price of risk. It was also expected to shift risk out of the heart of the payments system and reduce the risk of financial crisis. The increased use of leverage and short-term borrowing complemented the rise in securitization, as institutions sought to magnify their exposure to rising asset prices. As we discuss in future reports, while securitization was meant to spread out risk away from the center of the financial system, exactly the opposite happened. When the credit crisis hit in August 2007, risk that was meant to be dispersed throughout the system was in fact heavily concentrated among leveraged institutions at the heart of the financial system.2/13/2001 2/13/2002 2/13/2003 2/13/2004 2/14/2005 2/14/2006 2/14/2007010203040506070 With a maturity between 1 a 4 days With a maturity between 21 a 40 days With a maturity greater than 80 daysT hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 32 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on The process of securitization was further aided by the growth of credit insurers and deriva - tives called Credit Default Swaps (CDS), which in principle allowed the default risk to be taken out of mortgage-backed securities and CDOs before they were marketed to general investors. The first forms of credit insurance were developed by so-called mono-line insurers such as MBIA and Ambac, which had emerged in the early 1970s to back municipal bond issues. These insurance com - panies had very strong credit ratings and they sold default insurance to issuers of municipal bonds. By providing default insurance, the mono-lines al - lowed the municipalities to borrow at AAA rates, whereas without insurance they would have faced lower ratings and hence higher borrowing costs. The mono-lines collected fees and the municipal borrowers ended up with lower net costs even after paying the fees. This proved to be a good if not exciting line of business because defaults are rare on municipal bonds. The mono-lines were able to take advantage of “ratings arbitrage” and it worked out well because the rating agencies were overestimat - ing the chances of defaults on municipal bonds—at least until now. Having developed this line of business, the mono- line companies, along with banks, hedge funds, and financial guarantors such as AIG expanded their business model into structured products related to the housing market, selling Credit Default Swaps to insure holders of MBS, CDOs and other assets against mortgage default risk. So, just as in the case of municipal bonds, the CDS was an instrument for a ratings arbitrage, providing an outside credit enhancement to the issuers of MBS and CDOs to obtain AAA ratings for their bonds – many of which would otherwise be considered lower-grade. Like credit insurance, a CDS transaction involves a “pro - tection buyer” – a bond issuer trying to raise ratings and shield certain bonds from default risk – and a “protection seller,” a counterparty who receives a fixed income stream in return for assuming the default risk. However, these transactions were not overseen by any regulatory body. They were done in Over the Counter (OTC) markets, so that no one other than the two parties knew the terms of the contract. Thus, there exists no public knowledge as to how many CDS transaction most institutions have made. Furthermore, there are no minimum capital or as - set requirements for the protection seller, so there is no guarantee that in the case of default the seller will have adequate funds to make full payment — an issue called “counterparty risk,” which has espe - cially become a concern since Bear Stearns, a gi - ant derivatives trader, collapsed in March 2008. In good times, though, CDS were yet another way for financial institutions to leverage their exposure to the mortgage market. An AIG executive said as late as August 2007 that “It is hard for us, without being flippant, to even see a scenario within any kind of realm of reason that would see us losing one dol - lar in any of those [CDS] transactions.”27 Just over a year later, the federal government provided AIG with an $85 billion loan to cover losses it faced on its CDS contracts (then followed by an additional $38 billion). Especially since 2000, the business of insuring mort - gage-related assets, along with corporate bonds and other assets, grew exponentially. Figure 7 illustrates the exponential growth in the CDS market since 2000. The size of outstanding CDS reached a stag - gering $60 trillion in 2007. As of September 2008, AIG, a financial guarantor, had itself sold nearly $500 billion worth of CDS — most of it insuring ill-fated CDOs. As the CDS market ballooned, so did the share of CDS sold by leveraged institutions credit insurance and Tremendous growth in credit Default Swaps 27. Gretchen Morgenson, “Behind Insurer’s Crisis, Blind Eye to a Web of Risk.” The New Y ork Times; September 27, 2008.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 33figuRE 7: Value of credit Default Swaps Outstanding; in trillions Source: International Securities and Derivatives Association like hedge funds and investment banks, relative to more capital-intensive mono-line insurers. Accord - ing to Fitch (2007), hedge funds drove nearly 60 percent of CDS trading volume in 2006. As the CDS market spread further into the unregulated, opaque financial world, its enormous scale and sys - temic implications went largely unnoticed until the crisis hit in August 2007. Credit insurance and CDSs are valuable innova - tions because by assuming the default risk of a trans - action, they facilitate lower funding costs and easier access to funding liquidity for institutions that may otherwise not have access to it. In May 2006 Alan Greenspan called them the “most important instru - ment in finance,” adding, “What CDS did is lay- 28. From remarks on May 18, 2006 to the Bond Market Association in New Y ork.off all the risk of highly leveraged institutions…on stable American and international institutions.”28 However, this powerful tool became a big problem because of the enormous size of the market and be - cause participants created an instrument which fi - nancial institutions used to leverage their exposure to an asset class and put very little capital on the line. Furthermore, contrary to Greenspan’s 2006 comment, the biggest source of recent growth in the CDS market was not among “stable” institu - tions, but rather among unregulated “highly lever - aged” institutions like hedge funds and investment banks. Insurance for life and property and casualty, on the other hand, are highly regulated and very capital intensive.010203040506070 1st Half 012nd Half 011st Half 022nd Half 021st Half 032nd Half 031st Half 042nd Half 041st Half 052nd Half 051st Half 061st Half 072nd Half 062nd Half 07T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 3 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on 29. Richard T omlinson and David Evans, “The Ratings Charade,” Bloomberg , July 2007. As we discuss later, federal law requires or relies upon the use of credit ratings in many other contexts. 30. Ashcraft and Schuerman (2008) 31. Quote taken from Charles Calomiris, professor at Columbia, in Richard T omlinson and David Evans, “The Ratings Charade,” Bloomberg , July 2007.The lack of transparency of CDOs made the market reliant on the grades of ratings agen - cies as a signal of the risk of CDO assets. Regulators were not involved in these markets, so rating agencies essentially acted as proxies for regu - lators; indeed, an office as high as the U.S. Office of the Comptroller of the Currency, which regu - lates nationally chartered banks, depended on rat - ing agencies to assess CDO quality.29 Furthermore, CDOs are themselves such complex instruments that independent judgment of risk is very difficult. The principal rating agencies – Moody’s, Fitch and Standard & Poor’s -- used complex quantitative statistical models called Monte Carlo simulations to predict the likely probability of default for the mortgages underlying the CDOs and eventually to structure the CDO (or MBS) in the way described in the previous section: separating the risk into the dif - ferent tranches and calculating the required amount of subordination and credit enhancement for each tranche as computed by the model. The information fed into these models to calculate default probabili - ties consisted of the characteristics of the mortgage pool, in terms of credit scores of the borrowers, the cumulative loan-to-value (CL TV) ratio, documenta - tion of income (or lack thereof), whether the mort - gages were for the borrower’s primary residence, as well as historical default rates on similar mortgages. The credit Rating agencies At the outset, this approach was problematic in that the historical default rates used in these models were largely from the years 1992 until the early 2000s30– a period when mortgage default rates were low and home prices were rising. By basing their estimates of default probabilities of newly issued CDOs on a historical period during a housing boom when home prices increased each year in both real and nominal terms, they did not factor in correctly the possibility of a general housing bust in which many mortgages are more susceptible to go into default. The reduc - tion of risk in a pool of mortgages depends on the extent to which default probabilities within the pool are not correlated. If there is a general downturn in housing across the country (which no one at the time believed was possible), then the probabilities of de - fault go up across the board. Unlike the case of corporate bonds, where a ratings agency passively rates the risk of a company, with structured products the agencies “run the show.31” The ratings agencies advised CDO issuers on how to structure the CDO with the lowest funding pos - sible. T o do so, CDO issuers would work with the agencies to optimize the size of the tranches in order to maximize the size of highly-rated, lower yielding tranches. Since the agencies were receiving substan - tial payments for this service, it created a clear conflict of interest. If CDO issuers did not get the rating they T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 3wanted, they could try another agency, taking their fees with them – an act known as “ratings shopping.” According to the New Y ork Times , Moody’ s profits tri - pled between 2002 and 2006 to $750 million, mostly because of the fees from structured finance products.32 According to Coval et al (2008), fees from structured finance products made up 44 percent of Moody’ s rev - enue in 2006. While the rating agencies appear to have faced per - verse incentives, it was the opacity of the entire system that magnified the effect of their poor judgment and “ratings inflation.” Not only did markets in CDOs and other structured products become so complex that ratings became the only way investors could judge risk, but most institutional investors face rules that only allow them to purchase investment-grade as - sets, as judged by the rating agencies. Thus the three agencies became the effective “arbiters of risk” for the entire market in structured finance products. 32. Roger Lowenstein, “T riple-A Failure, New York Times Magazine, April 27, 2008.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 36 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on One of the culprits often cited for the financial crisis is the Federal Reserve’ s policy of keep - ing interest rates low for a long time in order to help the economy pull out of the 2001 recession. The unemployment rate was rising and inflation was falling (see the analysis in T aylor (2007)). The Federal Funds rate was moved down to 1 percent in mid 2003 and held at that level until mid 2004. With short term rates as low as 1 percent, many financial institutions struggled to earn returns they considered adequate. Money market mutual funds had trouble covering expenses and paying any return above zero to their investors, while other fund managers searched des - perately for higher yielding assets without taking on undue risks. One fund manager described the situ - ation to us as follows: he felt compelled to purchase mortgage and other asset-backed securities because they offered superior yields and were highly rated by the credit agencies. He knew that the risks might turn out to be larger than were being allowed for, but his clients would have pulled their money out of his funds had he not made the investments. Competing investment funds were advertising high returns and low risks. Because it kept short-term interest rates so low for so long, it is argued, the Fed encouraged this behavior. Should the Federal Reserve have kept interest rates at a higher level, or raised them sooner in order avert the housing price bubble? Edwin T ruman and Mi - chael Mussa of the Peterson Institute have both ar - gued that monetary policy should be adjusted when there are clear signs of developing asset price bubbles. See Mussa (2004) and T ruman (2005). For example, equity prices moved very high in the late 1990s, es - pecially technology stocks but the whole market also. The equity bubble then burst and many Americans were severely impacted. There had been overinvest - ment in technology capital stock in the 1990s, and the subsequent slump in technology investment after the tech bubble burst was instrumental in causing the 2001 recession. If interest rates had been moved up more quickly in the 1990s, perhaps this bubble could have been avoided or reduced in size. Again in the 2000s, the housing bubble resulted in a huge con - struction boom and associated spending on furniture, appliances and so on. Since the Federal Reserve is charged with keeping the economy on an even keel, there is a case that monetary policy should have be - come tighter sooner to counteract the overinvest - ment in housing. Some small amount of economic growth might have been sacrificed in 2003 to 2007, but to the benefit of economic growth later, if the slump had been avoided. Our Brookings colleague Douglas Elmendorf (2007, 2008) has made the counter argument. He concluded that monetary policy was only a little too expansionary in the early part of this decade when judged by the out - comes of unemployment and inflation. Given the oth - er forces affecting the aggregate economy, low interest rates were appropriate. He notes that countercyclical policy is a very blunt tool, and the impact on the overall economy would need to be very large to ensure that an asset price bubble was actually deflated. He also points out that it is very hard ex ante to de - termine when asset price appreciation is really part of a bubble. For example, then-Fed Chairman Alan Greenspan warned about “irrational exuberance” in the stock market in 1996 when the Dow Jones index was only at 6,000. Anyone getting out of the market at that point, as some did, would have missed out on large and sustained capital gains. More recently, an article by Jonathan McCarthy and Richard W . Peach (2004) of the New Y ork Federal Reserve Bank concluded that there was little evidence of a bubble in house prices at that time. Thus, counting on the Fed to accurately assess asset price bubbles is problematic. Although in past writing we have found this argument persuasive, we now side with those favoring adjust - ments of monetary policy to ameliorate asset price bubbles. The housing bubble and collapse has been federal Reserve Policy, foreign Borrowing and the Search for YieldT hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 37so costly to taxpayers and the economy that it would have been worth the price in terms of slower economic growth 2004-2007 in order to lessen the collapse we are now going through. While it is difficult to know for sure if there is an asset price bubble when it is hap - pening, there are situations where the probability of there being a bubble is high. Most careful observers of the housing market in 2006 knew that a collapse was very possible. Even if McCarthy and Peach were cor - rect about there being no bubble in 2004, there sure was one a couple of years later. We now conclude that monetary policy, which kept interest rates so low, was one reason for the financial crisis. Even so, it is a mistake to overstate the possible impact that might have resulted from a different path for the short-term Federal Funds rate. The Fed sets this rate but not the broad spectrum of interest rates, as can be seen back in Figure 1. The Fed tightened mon - etary policy starting in 2004 but the mortgage interest rate stayed very low compared to its past history.33 Apart from Fed policy, one very important reason interest rates have remained low in the United States and around the world is because the supply of sav - ings has been large relative to the demand for funds for investment.34 The United States is a low saving, high borrowing economy and has financed both its budget deficits and residential investment by foreign borrowing. In part, this has been direct funding by foreign institutions of U.S. companies and mortgage debt instruments. But since money is fungible, it does not matter greatly which assets foreigners were buy - ing; the key is that they were willing to finance a very large capital inflow to the United States. The inflow of capital has as its counterpart the current account deficit and Figure 8 shows the very large and growing U.S. deficit in recent years. Because of the globaliza - tion of financial markets and because of all the money from around the world looking for returns, the U.S. economy was able to finance its housing boom at low interest rates.35 33. One important factor in the crisis is that institutions were borrowing short and lending long, as we noted earlier. T o a degree, the low short-term interest rate policy of the Fed encouraged this, but importantly, this pattern persisted and even intensified even well after the Federal Funds rate was raised to 5¼ percent. The undoing of the short-term borrowers came when the risk premium increased sharply, as we describe in the following report. 34. Economists have not developed a consensus theory of the determination of interest rates and we do not intend to get into the middle of that debate. It is sufficient to note that both monetary policy and the global supply of and demand for savings are important. 35.There is another way of looking at this issue which says that it is not that the inflows allowed the U.S. to keep interest rates low; rather it is that capital inflows and the associated high dollar and weak demand for our net exports required us to keep interest rates low in order to generate enough aggregate demand to maintain full employment. If there had been no global savings glut many things would have different, with more US. net exports and the FED would have operated a different monetary policy with higher interest rates. A key issue is the composition of economic growth and whether an economic expansion is “balanced.”figuRE 8: capital inflows to the uS Economy (Equal to the current account Deficit) Reached Over 6 percent of gDP in 2006 Source: Bureau of economic Analysis2000 2001 2002 2003 2004 2005 2006 200701234567T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 38 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on We cannot know exactly the counterfactual of what the U.S. economy would have looked like if for - eigners had not been willing to lend to the U.S. on such favorable terms. But it seems highly likely that there would have been higher U.S. interest rates and less of a housing boom. In some sense, there - fore, one can assign a fraction of the “blame” for the housing bubble on those who sent capital to the U.S. economy. That is a tricky argument, however. An important policy goal for the U.S. has been to keep interest rates low on average to encourage investment and economic growth. The discipline in the federal budget developed in the 1990s was justified, correctly, on this basis. Generally, it is December 2006 • Ownit Mortgage Solutions files for bankruptcy. February 8, 2007 • HSBC Holdings, a large London-based bank, an - nounces a $10.5 billion charge for bad debt, top - ping analysts’ estimates by over $2 billion. The company claims that the 20 percent increase in the charge is due to its U.S. subprime mortgage portfolio. February 28 • Freddie Mac announces that they will no longer purchase subprime loans. March 13 • Mortgage. Banker Association data for the last three months of 2006 shows late or missed pay - ments on mortgages rose to 4.95%, rising to 13.3% in the subprime market. April 3 • New Century Financial a large subprime mort - gage lender files for Chapter 11 bankruptcy. June 10-12 • Moody’s downgrades the ratings of $5 billion worth of subprime RMBS and places 184 CDO tranches on review for downgrade. S&P places BOx 3: a Timeline of the initial Wave of the crisis $7.3 billion of 2006 vintage RMBS on downgrade watch and announces a review of CDO deals ex - posed to subprime RMBS bonds. June 12 • Bear Sterns announces trouble at two of its hedge funds, High-Grade Structured Credit Strategies Enhanced Leverage Fund and Bear Stearns High- Grade Structured Credit Strategies Master Fund, citing deterioration in the value of highly rated mortgage backed securities. June 22 • Bear Stearns attempts to bailout its hedge funds by injecting $1.6 billion in liquidity in the “Enhanced” fund, which has lost nearly all its value. July 31 • The two troubled Bear Stearns hedge funds file for bankruptcy. August 1 • French insurer AXA SA ’s money-management unit has offered to cash out investors in a billion-dollar bond fund after the fund shrank in size by about 40% in the last month. T wo of the AXA fund’s sub- funds had lost 13.5% and 12.6% of their value. better to finance investment with savings generated at home, but if those savings are not forthcoming, it is better to keep investing productively and borrow the money. Without access to foreign funds, the U.S. economy would have invested less in all kinds of capital. The problem was the diversion of too much investment into housing that was not produc - tive at the margin, a problem we should blame on ourselves more than on those who lent the money. Moreover, foreign investors have taken a big hit from their lending to us as banks all across the globe have faced heavy losses on their assets related to US mortgages.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 3August 2 • German bank IKB Deutche has to be bailed out by a German state-run bank due to troubles from exposure to U.S. Subprime loans. August 6 • American Home. Mortgage Investment Corp files for Chapter 11 bankruptcy. August 9 • French bank BNP Paribas said it was freezing three funds due to subprime-related losses. • The European Central Bank and the Federal Re - serve expanded funds for lending to banks in re - sponse to a widespread liquidity shortage. • For the first time in years, the amount of Asset- Backed Commercial Paper (ABCP) outstanding falls, signaling a seizing up of credit markets. August 16 • The Fed announced a half-percentage point cut of its discount rate to 5.75 percent. August 17 • Countrywide draws down its entire $11.5 billion line of credit and faces a run. August 23 • Countrywide receives a $2 billion liquidity injec - tion from Bank of America. September 14 • British bank and mortgage lender Northern Rock faces a run on its deposits and receives a liquidity injection from the Bank of England. September 18 • Federal Reserve lowered the Federal Funds rate by half a percentage point, to 4.75%. October 15 • Citibank announces a $6.4 billion write-down. October 24 • Merrill Lynch & Co. announces an $8.4 billion write-down. October 31 • The Fed cuts its target for the Federal Funds rate by a quarter point, to 4.50 percent. November 4 • Citigroup increases its write-down to $11 billion and its CEO resigns.November 7 • Morgan Stanley takes an additional $3.7 billion write-down. • Rating agency Fitch says it will review the ratings on CDOs insured by guarantors including Ambac and MBIA. November 14 • HSBC takes a higher-than-expected $3.4 billion charge and takes $41 billion in SIV assets onto its balance sheet. • Bear Stearns takes a $1.2 billion write-down for the fourth quarter. December 11 • Fed announces a quarter percentage-point cut in the Federal Funds rate. December 12 • In coordination with four other central banks, Fed extends up to $40 billion in special loans in the next eight days to banks. December 13 • Citigroup Inc. brings $49 billion in distressed assets onto its balance sheet. January 18, 2008 • Washington Mutual Inc reports a $1.87 billion loss in the fourth quarter. • Fitch Ratings downgrades Ambac. January 21 • While U.S. markets were closed for the Martin Luther King Jr. holiday, major worldwide indexes fell, including drops of 7.2% in Germany, 7.4% in India and 5.5% in Britain. January 22 • Fed cuts Federal Funds rate by three quarters of a percentage point. March 16 • It is announced that Bear Stearns is to be sold to J.P . Morgan Chase for $2 a share under an agreement brokered by the Federal Reserve and the T reasury, and enhanced by a $30 billion loan guarantee from the Fed. This was to forestall the impending bankruptcy of Bear Stearns. This is the first time that the Federal Reserve has pro - vided support to an investment bank.Timeline, continuedT hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 0 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on For over 30 years there has been a thrust in U.S. policy towards reduced regulation of private markets. Airlines and trucking were deregu - lated in the 1970s; President Reagan was a supporter of deregulation, as was his philosophical ally, Mrs. Thatcher, in the U.K. Financial markets have also gradually been deregulated, going back to the abil - ity of money market mutual funds to issue interest- bearing checking accounts, through the ending of Glass-Steagall prohibitions on banks. Determining how much deregulation is optimal is tricky, how - ever, as we have seen in the electric power industry. The financial sector is just as tricky, or more so. In order to prevent bank runs, there has been de - posit insurance in the United States since the 1930s that has parallels in other advanced economies. If depositors know their funds are protected, they do not have to rush to withdraw money at the first ru - mor of trouble. Recently the U.K. bank Northern Rock got into trouble and depositors were lining the streets outside Northern Rock branches be - cause the deposit insurance program in the U.K. did not provide adequate coverage. In addition, the Federal Reserve, like other central banks, stands as the lender of last resort to provide additional liquid - ity to banks in difficulty, a role that was extended to the investment bank Bear Stearns in March 2008, and since then has effectively been extended to the entire financial system.36 The Federal government has taken on the role of protecting individual bank depositors and the role of protecting the financial sector as a whole. Given that the Fed and U.S. taxpayers are on the hook to insure deposits and preserve the stability of the financial system, it is appropriate to have regu - lators that make sure the institutions are behaving Regulation and Supervision responsibly. In addition, there is a further case for supervision of the mortgage market because buying a house is usually the largest investment a family makes in their lifetime and requires a level of so - phistication in financial matters that many or most households do not possess. Markets do not work well when there are information asymmetries and this is such a market. There is a clear case, therefore, for better regulation in mortgage and financial markets. And in practice, there was still an extensive regulatory apparatus in place in financial markets. As described by a se - nior executive at one of the large U.S. banks, there were “roomfuls of regulators” going over the books. On the consumer side, anyone who has taken out a mortgage knows that there is a stack of papers to sign created by state and federal (RESPA and TILA) regulators with the goal of protecting borrowers. Why did this level of regulation not work? This is an issue that will be explored more fully in future papers in this series as we look at what should be done to avoid the same problems in the future, but a couple of points here are notable. There is no unified system of bank supervision, rather a patch - work of state and federal regulators. In researching this paper we have been struck by the complexity of SIVs and CDOs, but also astounded by the byz - antine complexity of the U.S. regulatory structure. The FED supervises all bank holding companies and banks that are members of the Federal Reserve Sys - tem. The Federal Deposit Insurance Corporation provides $100,000 (recently boosted to $250,000) of deposit insurance and is the federal regulator of about 6,500 state-chartered banks that are not in the FED system. The Office of the Comptroller of the Currency charters and regulates 1600 national 36. At the time, the Federal Reserve could not provide funds to Bear Stearns directly because it was not a deposit-taking bank, which is why it had to step in through JP Morgan, which was a deposit-taking institution.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 1banks while 50 state banking departments charter and regulate state banks. Membership in the Fed - eral Reserve System is required for national banks. The Office of Thrift Supervision supervises what is left of the S&Ls. The Financial Standards Ac - counting Board regulates accounting rules and the SEC regulates corporations, including the invest - ment banks and ratings agencies. No one has clear authority. We have developed a national mortgage market with global connections and yet we have no national, uniform regulatory authority. Along with the byzantine nature of the federal regu - latory system, a significant share of the subprime mortgages — those that are at the root of the cur - rent financial crisis — were originated by institu - tions outside the purview of federal regulation in the first place. Indeed, it is estimated that over half of subprime loans originated in 2004 and 2005 were originated by independent mortgage companies — non-depository companies unaffiliated with any bank (see Cole (2007)). These independent com - panies were not covered by the FDIC, Federal Re - serve, or federal regulation, but rather only by state regulation. Looking at the kinds of loans that are delinquent now, and thus at the root of the current financial turmoil, it is clear then that a significant share of now delinquent loans were originated “under the radar” of federal regulation. Data from the Mort -gage Bankers Association says that 6.4 percent of all mortgage loans were delinquent as of 2008Q2, and we estimate that about one quarter of those were originated outside of the federal regulatory system. The same is true of roughly 30 percent of mort - gages that entered foreclosure in 2008Q2.37 Thus, even if they were willing and apt enough to rein in on lax lending standards, federal regulators did not have the direct authority to do so for a sizeable share of the market. The fact that much of subprime lending occurred outside the purview of federal regulators does not exonerate them. Despite the limitations of its authority, the Federal Reserve should have done much more to slow or stop the erosion of mortgage lending standards. Then-Fed governor Edward M. Gramlich warned his colleagues of the decline of lending standards and the dangers that this posed as early as 2000. There is a consumer advisory board that briefs the Fed on its views and its concerns. The minutes of this group’s meeting in 2005 re - veal that they were aware of the problems emerg - ing in the mortgage markets and warned the Fed about them. The Federal Reserve had the stature to change things and to influence state regulators. Appropriate warnings given privately or publicly could have significantly reduced the amount of bad lending even in markets where the Fed had no di - rect legal power. And of course state and federal regulators should have done better also. 37. We estimate these numbers as follows. MBA data shows that as of 2008Q2, 6.41 percent of mortgages were delinquent. Subprime ARMs made up 6.8 percent of all outstanding loans, and 33 percent of them were delinquent. Subprime FRMs made up 6.3 percent of all loans, and 12 percent of them were delinquent. We calculate (.068)*(.33) + (.063)*(.12) = .03. Thus subprime loans that are delinquent make up 3 percent of all loans. We said above that roughly half of subprime loans were originated by independent mortgage companies, so delinquent subprime loans originated by independent companies make up 1.5 percent of all loans, or (.015 / .0641 = .23) 23 percent of all currently delinquent loans. The calculation for foreclosure starts in 2008Q2 is similar.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 2 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on Many financial companies have lost huge amount of money in the aftermath of the crisis and CEOs have lost their jobs. The crisis is not just or even primarily a failure of regulation; it reflects poor internal corporate gov - ernance, poor infrastructure in and oversight of opaque financial markets, and most of all, mistakes made by decision-makers in the private sector. There have been two important assessments made of the failures (and successes) of risk management practices at financial institutions in the wake of the crisis. On March 6, 2008 the Senior Supervisors Group of the Financial Stability Forum issued a report “Observations on Risk Management Prac - tices During the Recent Market T urbulence.” This report was based on a survey of eleven of the larg - est banking and securities firms (plus there was a roundtable meeting that included five additional firms). The report identifies risk management practices that helped some of these institutions avoid the worst of the losses and the practices that led to failures. On April 18, the Swiss bank UBS issued a “Share - holder Report on UBS’s Write-Downs” at the re - quest of Swiss banking authorities. It is a lengthy and extraordinary mea culpa detailing the problems that resulted in the very large losses that UBS expe - rienced in the mortgage security market. Readers are referred to the reports themselves for the detailed analysis of best and worst practices, but a couple of points emerge that are the most im - portant. The biggest problems occurred where top managers failed to monitor and control the parts of The failure of company Risk Management Practices their companies that were trading in CDOs and re - lated securities. Financial institutions had in place risk management rules, but they were not followed, largely because so much money was being gener - ated during the boom times. Without exercising adequate supervision, senior managers believed that the risky assets were simply being sold in the marketplace and not held on the balance sheets of the banks. In fact, large amounts were being held, partly because there was a lag between the issuance of the securities and their sale, and partly because holding the securities was (for a time) so profit - able. As we have noted earlier, a major problem was that the credit ratings provided by the agencies were ac - cepted without adequate knowledge of the risks of the underlying mortgage portfolios. And there was not adequate stress testing of the portfolios against a correlated shock (a broad market decline), nor did the institutions take a complete view of their risks. Different parts of the businesses were considered separately, rather than as part of larger company- wide portfolios. Faced with low interest rates and competitive pres - sures to generate high returns for investors and high profits for shareholders, several of the finan - cial institutions failed to apply the risk management practices that they already had in place. They have now learned a lesson and doubtless will behave dif - ferently in the future, at least for a while. This is a discouraging story, however, because the Sarbanes- Oxley Act of 2002 was intended to beef up risk man - agement practices and make senior managers take full responsibility for avoiding this kind of crisis.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 3One of the changes in accounting rules that has been introduced gradually into cor - porate reporting standards is the prac - tice of marking assets on the balance sheet to the value they would have if sold in the marketplace. For many economists this is a common sense move in which companies can no longer carry on their books assets that no longer hold the value being calculated through historical cost and depreciation. Assets that have become more valuable can also be recognized by this approach. Mark to Market was introduced in 1993 after the S&L crisis, when then backward-looking GAAP ac - counting standards prolonged the crisis by allowing many thrifts to appear solvent on their books, even though their equity had effectively been wiped out. Because book value of assets was calculated based on the cost at which the thrift bought it, rather than its current market price, bank balance sheets appeared unaffected as asset values plummeted. This opacity allowed banks that should have been shut down to linger on and prolong the crisis. Mark to Market was seen as the solution to bring greater transpar - ency to balance sheets and prevent an S&L type debacle from happening again. There are two problems that can arise with mark to market accounting, however. The first is that many assets are unique and are not traded regularly in markets. There really is no market price read - ily available at which to value them. And second, market prices are very volatile and can overshoot on the upside when a bubble is forming, but may The impact of Mark to Market overshoot on the downside when the bubble bursts. Both kinds of overshooting can be problematic. On the upside, company asset values become overstat - ed when marked to market and, as we described earlier, this rise in asset prices can tempt companies to expand lending and over-leverage. When the bubble bursts, asset prices fall too much and banks are forced to contract lending sharply and they may become insolvent if liabilities exceed the value of assets when marked to market. If the institutions were allowed to hold the assets to maturity, in this case, they would realize a greater present value of their cash flow than the short term market valua - tion would indicate. An important rationale for the TARP program pro - posed by T reasury as part of its response to the crisis was to try and restore more accurate and transpar - ent pricing to bank assets. The reverse auction that they proposed was designed to work as a “price dis - covery” mechanism, allowing markets to find out what some of the distressed assets were really worth when evaluated in terms of their “hold to maturity” value, rather than at the fire sale prices that had prevailed in the atmosphere of the crisis. There are observers who conclude that mark to market accounting bears a great responsibility for the crisis, forcing institutions into bankruptcy when they had positive net worth, if evaluated on a long- term value basis. At this point in the crisis, the jury is still out on that issue. We just do not know yet what the true value of some of the assets will be.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS  The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on This concludes the discussion of the origins of the crisis. We include a timeline of the initial wave of the crisis in Box 3, starting in late 2006 and up to the collapse of Bear Stearns in March 2008. This final section presents a few les - sons from the story so far. The next piece in the Fixing Finance series will discuss the spread of the crisis to global markets, the response by regula - tors, and how it has played out in the Main Street economy. • Some factors that contributed to the crisis are ones that are not amenable to change, except at unacceptable cost. For example, the housing boom would surely not have continued as it did if funds had not been available on a large scale from foreign lenders. But closing off the U.S. borders to foreign capital is not acceptable. The price would be too high and, given the integra - tion of U.S. companies with the rest of the world, it would be infeasible. • A more aggressive tightening of monetary policy earlier in the cycle might have constrained the housing boom. We think this would have been a good idea, but there are limits to how much could have been done this way. There were better ways to avoid crisis than pushing the economy into stagnation in 2004 or 2005. Lessons from Studying the Origins of the crisis • The erosion of mortgage lending standards stands out as something that could and should have been stopped, especially when there were fears of a housing bubble. The challenge going forward is either to create an incentive structure within the “originate to distribute” model that leads to the outcome we want, or to provide a better and more integrated force of regulators to make sure that there is not too much bad behav - ior. Or to use a combination of the two. • Securitization of mortgage assets went beyond the point of value and created assets that were not transparent. We know from economic the - ory that markets with information asymmetries are trouble and the compounding layers of secu - ritization seem to have been designed to exacer - bate this problem. • The infrastructure of the financial system needs to be overhauled. While complex derivatives like CDS have grown exponentially, no one knows how exposed any one institution is to these prod - ucts because each CDS transaction is done Over the Counter (OTC) rather than on an exchange. This lack of transparency further exacerbates the problem of asymmetric information and magni - fied the potential for systemic risk.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 • Credit Rating Agencies failed to accurately as - sess the risk of the securitized assets they grad - ed. They faced a conflict of interest in their fee structure. A big part of the credit ratings prob - lem is that the system got so opaque that rating agencies became the de facto “arbiters of risk,” as everyone — even regulators — came to utterly rely on their opinion of risk assessment. There should be reforms in the credit ratings structure and perhaps less reliance on agency ratings. • Financial institutions did not follow their own best practices for risk management. In the short run, they will surely make internal changes, but experience suggests that some years from now there will be another problem. Is this problem amenable to policy change or not? Sarbanes-Ox - ley is already creating competitiveness problems for U.S. financial markets and did not work to forestall this crisis. One important issue in this area is determining whether banks were over lev - eraged and had inadequate capital. Apparently, Basel II rules did not work either. • The “pro-cyclicality” of liquidity and leverage in the financial system must be addressed in future regulatory discussions.• The general public was not given adequate warning of the emerging dangers in the mort - gage market. We cannot expect policymakers to second guess markets or to know when assets are overvalued. But we can and should expect policymakers to warn of the growing riskiness of certain assets that might generate large rewards but that could also lead to large losses. House - holds should have been warned that continuing large increases in house prices were not a sure thing. • Why did Federal and state regulators not do more? An important issue is that they believed that less regulation was better and that the mar - ket would take care of any problems. The push to deregulate of the past thirty years has led to a lack of discrimination in policy. We need to get rid of bad regulation that stifles competition and inhibits innovation, but we need to improve regulation where it can make markets work bet - ter and avoid crises.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS 6 The Initiative on Business and Public Policy | The Brook IngS InSTITUTI on Adrian, T obias and Hyun Song Shin. 2007. “Liquidity and Leverage.” Unpublished paper, Federal Reserve Bank of New Y ork and Princeton University. (September). Ashcraft, Adam B. and Til Schuerman. 2008. “Understanding the Securitization of Subrpime Mortgage Credit.” Staff Report no. 318, Federal Reserve Bank of New Y ork. (March). Barth, James R. and Glenn Yago. 2008. Demystifying the Mortgage Meltdown: What it Means for Main Street, Wall Street and the US Financial System. Presentation at The Milken Institute; October 2. Basel Committee on Banking Supervision. 2001. The New Basel Capital Accord. Bank for International Settlements (May). Bikhchandani, Sushil; David Hirshleifer and Ivo Welch. 1992. A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades. The Journal of Political Economy, Vol. 100, No. 5 (Oct., 1992), pp. 992-1026. Brunnermeier, Markus. 2008. Thoughts on a New Financial Architecture. Remarks from “Crisis on Wall Street” Panel held at Princeton University; September 23. Case, Karl E., Katharine Coman and A. Barton Hepburn. 2008. “The Central Role of House Prices in the Current Crisis: How Will the Market Clear?” Brookings Papers on Economic Activity 2:2008. Case, Karl E. and Robert J. Shiller. 2003. “Is There A Bubble in the Housing Market?” Brookings Papers on Economic Activity 2:2003: 299-342. Cole, Roger T . 2007. “Subprime Mortgage Market.” T estimony before the U.S. Senate Committee on Banking, Housing, and Urban Affairs (March 22). Coval, Joshua D., Jakub Jurek, and Erik Staffurd. 2008. The Economics of Structured Finance. Harvard Business School Working Paper; Cambridge, MA. Credit Suisse. 2007. “Mortgage Liquidity du Jour: Underestimated No More.” Equity Research. (March). Demyanyk, Yuliya and Otto Van Hemert (2008). “Understanding the Subprime Mortgage Crisis.” Unpublished paper, Federal Reserve Bank of St. Louis and New Y ork University. (February). Elmendorf, Douglas W. 2008. “Financial Innovation and Housing: Implications for Monetary Policy.” Unpublished paper; The Brookings Institution (April 21). Available at <http://www.brookings.edu/papers/2008/0421_monetary_ policy_elmendorf.aspx> Elmendorf, Douglas W. 2007. “Was the Fed T oo Easy for T oo Long?” Unpublished paper, the Brookings Institution. (November 9). Available at <http://www.brookings.edu/ opinions/2007/11_fed_elmendorf.aspx> Fitch Ratings. 2007. “Special Report: CDx Survey — Market Volumes Continue Growing While New Concerns Emerge.” (July 16). Available at <www.fitchratings.com> . Gallin, Joshua. 2004. “The Long-Run Relationship between House Prices and Rents.” Finance and Economics Discussion Paper no. 2004-50, Washington, DC: Board of Governors of the Federal Reserve System. (September). Gerardi, Kristopher, Adam Hale Shapiro, and Paul Willen. 2007. “Subprime Outcomes: Risky Mortgages, Homeownership Experiences, and Foreclosures.” Working Paper no. 07-15, Federal Reserve Bank of Boston. (December).reFeren CeS Gorton, Gary B. 2008. The Subprime Panic. NBER Working Paper 14398. National Bureau of Economic Research; Cambridge, MA. Gramlich, Edward. 2007a. “Booms and Busts: The Case of Subprime Mortgages.” Paper presented at the Federal Reserve Bank of Kansas City Symposium, Jackson Hole, Wyo., August 31-September 1. Gramlich, Edward M. 2007b. Subprime Mortgages: America’s Latest Boom and Bust. Washington, D.C.: The Urban Institute. Green, Richard K. and Susan M. Wachter. 2007. “The Housing Finance Revolution.” Paper presented at the Federal Reserve Bank of Kansas City Symposium, Jackson Hole, Wyo. (August 31-September 1.) Greenlaw, David, Jan Hatzius, Anil K Kashyap, and Hyun Song Shin. 2008.“Leveraged Losses: Lessons from the Mortgage Market Meltdown.” Paper prepared for the U.S. Monetary Policy Forum, University of Chicago and the Rosenberg Institute for Global Finance at Brandeis University. (February 29). Greenspan, Alan and James Kennedy. 2007. “Sources and Uses of Equity Extracted From Homes.” Finance and Economics Discussion Paper no. 2007-20, Board of Governors of the Federal Reserve System. (March.) International Monetary Fund. 2008. Global Financial Stability Report: Containing Systemic Risks and Restoring Financial Soundness. Washington DC. (April). JP Morgan. 2005. “ABCP Market Dynamics and T rends.” Global Structured Finance Research. (June 6). Mason, Joseph R. and Joshua Rosner (2007). “How Resilient are Mortgage Backed Securities to Collateralized Debt Obligation Disruptions?” (February 15). Available at SSRN: <http://ssrn.com/abstract=1027472> McCarthy, Jonathan and Richard W. Peach. 2004. “Are Home Prices the Next Bubble?” Federal Reserve Bank of New Y ork Economic Policy Review. 10(3): 1-17. Mian, Atif and Amir Sufi. 2008. “The Consequences of Mortgage Credit Expansion: Evidence from the 2007 Mortgage Default Crisis.” Unpublished paper, University of Chicago. (January). Minsky, Hyman. 1992. The Financial Instability Hypothesis. The Jerome Levy Economic Institute of Bard College; Working Paper No. 74 (May). Moody’s Investor Services. 2008a. “2008 U.S. CDO Outlook and 2007 Review.” Structured Finance Special Report. (March.) Moody’s Investor Services/ 2008b. “Moody’s Update on Structured Investment Vehicles.” International Structured Finance Special Report. (January.) Mussa, Michael. 2004. “Global Economic Prospects: Bright for 2004 but with Questions Thereafter.” Presentation at The Institute for International Economics; Washington DC (April 1). Office of Federal Housing Enterprise Oversight. 2008. Report to Congress. Passmore, Wayne; Shane M. Sherlund, and Gillian Burgess. 2005. The Effect of Housing Government-Sponsored Enterprises on Mortgage Rates. Real Estate Economics 33 (Fall 2005), 427-63.T hE ORi g iN S Of ThE fiNaNc i a L c RiSiS noVeMBer 2008 7ABoUT T he AUT hor S MARTIN NEIL BAILY is a Senior Fellow in Economic Studies at the Brookings Institution and the director of the Initiative on Business and Public Policy. ROBERT E. L ITAN is a Senior Fellow in Economic Studies at the Brookings Institution. MATTHEW S. JOHNSON is a Senior Research Assistant in the Economic Studies Program at the Brookings Institution.Shiller, Robert. 2008. The Subprime Solution: How T oday’s Global Financial Crisis Happened, and What to Do About It. Princeton University Press; Princeton, NJ. T aylor, John. 2007. “Housing and Monetary Policy.” Paper presented at the Federal Reserve Bank of Kansas City Symposium, Jackson Hole, Wyo., August 31-September 1. T ruman, Edwin. 2004. Postponing Global Adjustment: An Analysis of the Pending Adjustment of Global Imbalances. Working Paper: The Institute for International Economics; Washington DC.The Brookings Institution 1775 Massachusetts Ave., NW , Washington, DC 20036 (202) 797-6000 The Initiative on Business and Public Policy provides analytical research and constructive recommendations on public policy issues affecting the business sector in the United States and around the world.Note: Data refer to FDIC-insured commercial and savings banks that were closed or received FDIC assistance.Figure 1.1 Number of Bank Failures, 1934–1995 Number 1935 1945 1955 1965 1975 1985 1995050100150200250300Chapter 1 The Banking Crises of theThe Banking Crises of the 1980s and Early 1990s:1980s and Early 1990s: SummarSummar y and Implicationsy and Implications Introduction The distinguishing feature of the history of banking in the 1980s was the extraordi- nary upsurge in the number of bank failures. Between 1980 and 1994 more than 1,600banks insured by the Federal Deposit Insurance Corporation (FDIC) were closed or re-ceived FDIC financial assistance—far more than in any other period since the advent offederal deposit insurance in the 1930s (see figure 1.1). The magnitude of bank failures dur-An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 4 Histor y of the Eighties—Lessons for the F utureing the 1980s put severe, though temporary , strains on the FDIC insurance fund; raised ba - sic questions about the ef fectiveness of the bank regulatory and deposit insurance systems; and led to far -reaching legislative and regulatory actions.1 This chapter summarizes the findings and implications of Histor y of the Eighties— Lessons for the Futur e: An Examination of the Banking Crises of the 1980s and Early 1990s , a study conducted by the FDIC’ s Division of Research and S tatistics to analyze var - ious aspects of the 1980–94 experience. The four sections of this summary deal with (1) the factors underlying the rapid rise in the number of bank failures; (2) the regulatory issuesraised by this experience; (3) questions that remain open despite the legislative and regula - tory remedies adopted between 1980 and 1994; and (4) concluding comments. The Rise in the Number of Bank F ailures in the 1980s: The Economic, L egislative, and R egulator y Backgr ound The rise in the number of bank failures in the 1980s had no single cause or short list of causes. Rather , it resulted from a concurrence of various forces working together to pro - duce a decade of banking crises. First, broad national forces—economic, financial, legisla - tive, and regulatory—established the preconditions for the increased number of bankfailures. Second, a series of severe regional and sectoral recessions hit banks in a number ofbanking markets and led to a majority of the failures. Third, some of the banks in these mar - kets assumed excessive risks and were insuf ficiently restrained by supervisory authorities, with the result that they failed in disproportionate numbers. Economic and F inancial Market Environment During m ost of the 1980s, the performance of the national economy , as measured by broad economic aggregates, seemed favorable for banking. After the 1980–82 recession the national economy continued to grow , the rate of inflation slowed, and unemployment and interest rates declined. However , in the 1970s a number of factors, both national and inter - national, had injected greater instability into the environment for banking, and these earlierdevelopments were directly or indirectly generating challenges to which not all bankswould be able to adapt successfully . In the 1970s, exchange rates am ong the world’ s major currencies became volatile after they were allowed to float; price levels underwent majorincreases in response to oil embar goes and other external shocks; and interest rates varied widely in response to inflation, inflationary expectations, and anti-inflationary Federal Re - serve m onetary policy actions. 1Although this study is devoted to banking, it is appropriate to recall that the thrift industry suf fered an even greater ca - tastrophe. In 1980 there were 4,039 savings institutions; approximately 1,300 savings institutions failed during the 1980–94period. This high proportion of failures led to the demise of the fund that insured savings institution deposits, and imposed heavy costs on surviving institutions and on taxpayers.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 5Developments in the financial markets in the late 1970s and 1980s also tested the banking industry . Intrastate banking restrictions were lifted, allowing new players to enter once-sheltered markets; regional banking compacts were established; and direct credit mar - kets expanded.2In an environment of high market rates, the development of m oney market funds and the deregulation of deposit interest rates exerted upward pressures on interest ex - penses—particularly for smaller institutions that were heavily dependent on deposit fund - ing. Competition increased from several directions: within the U.S. banking industry itselfand from thrift institutions, foreign banks, and the commercial paper and junk bond mar - kets. The banking industry’ s share of the market for loans to lar ge business borrowers de - clined, partly because of technological innovations and innovations in financial products. 3 As a result, many banks shifted funds to commercial real estate lending—an area involvinggreater risk. Some lar ge banks also shifted funds to less-developed countries and leveraged buyouts, and increased their of f-balance-sheet activities. Condition of Banking on the Eve of the 1980s Yet on the eve of the 1980s m ost banks gave few obvious signs that the competitive environment was becoming m ore demanding or that serious troubles lay ahead. At banks with less than $100 million in assets (the vast majority of banks), net returns on assets(ROA) rose during the late 1970s and averaged approximately 1.1 percent in 1980—a levelthat would not be reached again until 1993, after the wave of bank failures had receded (seefigure 1. 2). 4For this group of banks, net returns on equity (ROE) in 1980 were also high by historical standards, equity/asset ratios were m oving gradually upward, and char ge-of fs on loans averaged approximately what they would again in the early 1990s. The fact that key performance ratios in 1980 compared favorably with those in 1993–94—a period of extra - ordinary health and profitability in banking that has continued to the present (mid-1997)—emphasizes the absence of obvious problems at m ost banks at the beginning of the eighties. Large banks, however , showed clearer signs of weakness. In 1980 ROA and equity/as - sets ratios were much lower for banks with m ore than $1 billion in assets than for small 2Many of these developments are discussed in Allen N. Ber ger, Anil K. Kashyap, and Joseph M. Scalise, “The Transforma - tion of the U.S. Banking Industry: What a Long, S trange Trip It’ s Been ,”Brookings Papers on Economic Activity 2 (1995). 3Between 1980 and 1990, commercial paper outstanding increased from 7 percent of bank commercial and industrial loans (C&I) to 19 percent. 4Data in figure 1.2 are unweighted averages of individual bank ratios. Use of median values or averages weighted by assetsreveals broadly similar trends, except that medians are less af fected by extreme values and tend to be less volatile than un - weighted averages, while weighted averages are dominated by lar ger banks in each size group. The data in figure 1.2 are for banks with assets greater than $1 billion (lar ge banks) or less than $100 million (small banks) in each year; thus, the num - ber of banks included in the two size groups varies from year to year . In 1980, there were 192 banks with assets greater than $1 billion and 12,735 banks with assets less than $100 million. In 1994, the comparable figures were 392 banks and 7,259banks. Asset data are not adjusted for inflation.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 6 Histor y of the Eighties—Lessons for the F utureLoans and Leases/AssetsNet Loan Charge-Offs/Loans Note: Data areunweighted averages ofindividual FDIC-insured commercial andsavings bank ratios. Largebanks arethose with assets greater than$1billion inanygiven year.Small banks arethose with assets lessthan$100 million inanygiven year.Figur e 1.2 Bank Performance Ratios, 1973 94–19 ROE ROA All Banks Large Banks Small BanksEquity/Assets 1974 1978 1982 1986 1990 199446810121974 1978 1982 1986 1990 19940.20.61.01.4 1974 1978 1982 1986 1990 1994-505101520 1974 1978 1982 1986 1990 199400.40.81.21.6 1974 1978 1982 1986 1990 1994455055606570Percent PercentPercent Percent PercentChapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 7Source: Bank Annual Salomon Brothers, , 1992 and 1996 editions.Figur e 1.3 Bank Price-Earnings Ratios as a Per centage of S&P 500 Price-Earnings Ratios, 1964–1995Percent 1964 Note: Data for superregional bank price-earnings ratios begin in 1982.1970 1975 1980 1985 1990 1995405060708090100 Money-Center RegionalBanks Superr egionalbanks and were also well below the lar ge-bank levels they would reach in the early 1990s. Market data for lar ge, publicly traded banking or ganizations suggest that investors were valuing these institutions with reduced favor . During the 1960s and 1970s price-earnings ratios for m oney-center banks trended generally downward relative to S&P 500 price-earn - ings ratios, although for regional banks the decline was much less pronounced (see figure1.3). For the 25 lar gest bank holding companies in the late 1970s and early 1980s, the mar - ket value of capital decreased relative to—and fell below—its book value, suggesting thatto investors, the franchise value of lar ge banks was declining. 5 Differences in performance between lar ge and small banks in 1980 are not surprising. At that time, because of branching restrictions and deposit interest-rate controls, manysmall institutions operated in still-protected markets. Accordingly , they were af fected m ore slowly by external forces such as increased competition and increased market volatility . 5Michael C. Keeley , “Deposit Insurance, Risk, and Market Power in Banking,” American Economic Review (December 1990): 1 185. Data are for the 25 lar gest bank holding companies as of 1985.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 8 Histor y of the Eighties—Lessons for the F utureDuring the 1980s, of course, performance ratios of banks of all sizes weakened and exhib - ited increased risk. Profitability declined and became m ore volatile, while loan char ge-of fs rose dramatically .6Large banks assumed greater risk in order to boost profits, as is indicated by the sharp rise in the ratio of loans and leases to total assets for these banks. In contrast,equity ratios increased over the period, particularly for lar ge banks, in line with increased regulatory capital requirements and perhaps also in response to market concerns about dis - tress in the banking system. Then in the 1990s the performance of banking improved markedly . This is apparent not only from the accounting data presented in figure 1.2 but also from the market data pre - sented in figures 1.3 and 1.4, which suggest that to investors, the value of publicly tradedbanks improved greatly in the 1990s. From 1993 to 1995, bank price-earnings ratios roserelative to S&P 500 price-earnings ratios, although the m ovements in this measure were ex - tremely volatile. After the early 1980s market prices per share of m oney-center and regional banks increased from below book value per share to well above book value, except for asharp and temporary drop in 1990 (figure 1. 4). The major improvement in the performance and investor perceptions of banking in the 1990s, albeit of limited duration so far , does not support earlier concerns that banking was a declining industry or the view that banking wascharacterized by widespread and persistent overcapacity that would lead to increased fail - ures. 7 Although the overall performance of the banking industry varied greatly during the 1980–94 period, in its structure the industry showed a strong trend in one direction—towardconsolidation into fewer banking or ganizations. This trend was partly due to the relaxation of branching restrictions. 8From the end of 1983 through the end of 1994, the number of in - sured commercial banks declined by 28 percent, from 14,461 to 10,451. The number of separate corporate units—bank holding companies plus independent commercial banks— 6The 1986 peak in net loan char ge-of fs for small banks was associated with the agricultural, ener gy, and real estate problems of the Southwest; the 1991 peak for lar ge banks was associated with the real estate problems in the Northeast. 7The issue of whether banking is a declining industry and the related question of overcapacity in banking are explored in Fed - eral Reserve Bank of Chicago, The (Declining?) Role of Banking, Pr oceedings of the 30th Annual Confer ence on Bank Structur e and Competition (May 1994). In the Proceedings , see particularly Alan Greenspan, “Optimal Bank Supervision in a Changing World,” 1–8; John H. Boyd and Mark Gertler , “Are Banks Dead? Or , Are the Reports Greatly Exaggerated?” 85–1 17; and Sherrill Shaf fer, “Inferring Viability of the U.S. Banking Industry from Shifts in Conduct and Excess Capac - ity,” 130–144. Shaf fer concludes that a small am ount of excess capacity in bank loans was eliminated in the mid-1980s. 8Some observers have ar gued that bank failures in the 1980s were partly due to restrictions on bank ownership (geographic restrictions within the banking industry , and prohibition of acquisitions by nonbank or ganizations), which prevented weak or inef ficient banks from being taken over before they failed. Although such restrictions on ownership probably contributed to the rise in the number of bank failures, particularly in the early 1980s, the lar ge number of voluntary mer gers and con - solidations within the industry may have averted some other failures by eliminating weaker institutions while they still hadsome value.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 9Figur e 1.4 Price-to-Book Value per Shar e, 1982–1995 Note: Values are industry composite medians. Data for superregional bank price-to-book ratios begin in 1987.Source: Bank Annual Salomon Brothers, , 1992 and 1996 editions.Money-Center RegionalBanks Superr egionalPercent 1982 1984 1986 1988 1990 1992 19955075100125150175 decreased somewhat m ore, by 31 percent. The 4,010 reduction in the number of insured commercial banks was due primarily to the consolidation of bank af filiates of multibank holding companies and to unassisted mer gers of unaf filiated banks (4,803). The net ef fect of failures, new charters, conversions, and other changes was an addition of 793 banks. Legislative Developments Banking legislation also played a lar ge role in the bank-failure experience of the 1980s and early 1990s.9This legislation was lar gely shaped by two broad factors: widespread recognition that banking statutes should be m odernized and adapted to new marketplace re - alities, and the need to respond to the outbreak of bank and thrift failures. In the early 1980sthe focus was on the attempt to m odernize, and congressional activity was dominated by ac - tions to deregulate the product and service powers of thrifts and to a lesser extent of banks. 9See Chapter 2, “Banking Legislation and Regulation.” Tax legislation was also a significant influence. After -tax yields on real estate investment were enhanced by the Economic Recovery Act of 1981 and then reduced by the Tax Reform Act of 1986 (see the appendix to Chapter 3).An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 10 Histor y of the Eighties—Lessons for the F utureThese deregulatory actions were generally unaccompanied by actions to restrict the in - creased risk taking they made possible, and so they contributed to bank and thrift failures.As the number of failures m ounted, the legislative emphasis then shifted to recapitalizing the depleted deposit insurance funds and providing regulators with stronger tools, while atthe same time restricting their discretion. As a group, the various legislative actions ad - dressed a variety of issues, but only the provisions m ost relevant to the increased number of bank failures are discussed here. The Depositor y Institutions Der egulation and Monetar y Contr ol Act of 1980 (DIDMCA) phased out deposit interest-rate ceilings, broadened the powers of thrift insti - tutions, and raised the deposit insurance limit from $40,000 to $100,000. Two years later the m ost pressing problem was the crisis of thrift institutions in an environment of high in - terest rates. Accordingly , the Garn–S t Germain Depositor y Institutions Act of 1982 (1) authorized m oney market deposit accounts for banks and thrifts to stem disintermediation, (2) authorized net worth certificates to implement capital forbearance for thrifts facing in - solvency in the short term, and (3) increased the authority of thrifts to invest in commercialloans to strengthen the institutions’ viability over the long term. In the case of national banks, Garn–S t Germain rem oved statutory restrictions on real estate lending, and relaxed loans-to-one-borrower limits. With respect to commercial m ortgage markets, this legisla - tion set the stage for a rapid expansion of lending, an increase in competition betweenthrifts and banks, overbuilding, and the subsequent commercial real estate market collapsein many regions. As the thrift crisis deepened and commercial bank problems were developing, Con - gress passed the Competitive Equality Banking Act of 1987 (CEBA) . It provided for re - capitalizing the fund of the Federal Savings and Loan Insurance Corporation (FSLIC)through the Financing Corporation (FICO), authorized a forbearance program for farmbanks, extended the full-faith-and-credit protection of the U.S. government to federally in - sured deposits, and authorized bridge banks. Two years later , again grappling with the thrift debacle, Congress passed the Financial Institutions Reform, Recover y, and Enfor ce- ment Act of 1989 (FIRREA), which authorized the use of taxpayer funds to resolve failed thrifts. Other provisions reflected congressional dissatisfaction with the regulation ofthrifts: the act abolished the existing thrift regulatory structure, m oved thrift deposit insur - ance to the FDIC, and mandated that bank and thrift insurance fund reserves be increasedto 1.25 percent of insured deposits. The belief that regulators had not acted promptly to head of f problems was again evident in the Federal Deposit Insurance Corporation Impr ovement Act of 1991 (FDICIA) . This act was aimed lar gely at limiting regulatory discretion in m onitoring and resolving industry problems. It prescribed a series of specific “prompt corrective actions”to be taken as capital ratios of banks and thrifts declined to certain levels; mandated annualChapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 1110Passage of the Deposit Insurance Funds Act was helped along by (1) the possibility of a FICO default if deposits were to shift from the Savings Association Insurance Fund, with higher assessment rates, to the Bank Insurance Fund, with lower assessment rates, and (2) the budgetary treatment of deposit insurance assessments, $3 billion of which was to be countedas revenue to “pay” for nonbanking spending programs. 11See Chapter 2, “Banking Legislation and Regulation.”examinations and audits; prohibited the use of brokered deposits by undercapitalized insti - tutions; restricted state bank activities; tightened least-cost standards for failure resolutions;and mandated a risk-based deposit insurance assessment system. Two years after the enactment of FDICIA, the Omnibus Budget Reconciliation Act of 1993 included a national depositor preference provision, which provided that a failed bank’ s depositors (and the FDIC standing in the place of insured depositors it has already paid) have priority over nondepositors’ claims. It was believed that national depositor pref - erence would make failure transactions simpler and less expensive to the insurance fundand would encourage nondeposit creditors to m onitor bank risk m ore closely . The final chapter of the savings and loan emer gency legislation was completed in Oc - tober 1996 with the enactment of the Deposit Insurance Funds Act, which provided for the capitalization of the Savings Association Insurance Fund, phased in pro rata bank and thrift payments of interest on FICO bonds, and required mer ger of the bank and thrift insurance funds in 1999 if no savings associations are in existence at that time. Given Congress’ s past reluctance to address promptly the need to fund thrift deposit insurance, enactment of thislegislation at a time when no major thrift failure was on the horizon suggests the extent towhich safety-and-soundness considerations had come to dominate banking legislation. 10 Legislation addressed not only the thrift and banking crises of the 1980s but also, af - ter those crises had ended, the question of interstate banking. By the end of the 1980s therisks posed by geographic lending concentrations were well understood, so attempts weremade to eliminate the remaining legal impediments to full interstate banking. Already state action had enabled many banking firms to use bank holding company af filiations to cir - cumvent geographic restrictions. Interstate banking was enacted in the Riegle-Neal Inter - state Banking and Branching Ef ficiency Act of 1994 , which enables banks to diversify loan portfolios m ore ef fectively . (As noted below , it also requires existing regulatory risk- monitoring systems to adapt to the changing nature of individual bank loan portfolios.) Regulation The tension between the two objectives of deregulating depository institutions and preventing or containing failures was manifest not only in legislative activity but also inpolicy dif ferences am ong the federal bank regulators. 11Of course, all three agencies were sensitive to issues of safety and soundness as well as to the importance of m odernizing bank powers. On specific issues, however , the Of fice of the Comptroller of the Currency (OCC)An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 12 Histor y of the Eighties—Lessons for the F uture12With respect to the potential short-term conflict between pro-competitive and safety-and-soundness objectives, the fol - lowing statement on S&L deregulation, made by the National Commission on Financial Institution Reform, Recovery and Enforcement, is instructive: “[C]omm on sense and prudence should have dictated that the industry be required to wait out the high interest rates, regain net worth, and then gradually shift into new activities. This is what well-managed and re - sponsible S&Ls did on their own, and they were lar gely successful” ( Origins and Causes of the S&L Debacle: ABlueprint for Reform [1993], 32). 13In 1984, 356 new commercial banks were chartered. By 1994 the number had declined to 47, but it then increased to 97 in 1995 and 140 in 1996.tended to emphasize the need to allow banks m ore freedom to compete and seek profit op - portunities, the FDIC leaned toward protecting the deposit insurance fund, and the FederalReserve often took a middle-of-the-road position. Differences between the FDIC and the OCC reflected the dif ferent responsibilities of an insurer and a chartering agency . They also reflected a problem that may potentially arise in bank regulation regardless of the agency involved: how to strike the correct balance be - tween encouraging increased competition and preserving stability and safety . To be sure, no such conflict is likely to exist in the long run: depository institutions must be able to com - pete and to participate in market innovations if they are to be viable in the long term. At any particular time, however , a short-term conflict may arise. The classic case is that of the sav - ings and loan industry . Broadened nonm ortgage powers were deemed essential to the long- term viability of thrift institutions, but the very act of providing these powers (withoutappropriate safeguards and at a time when thrifts were undercapitalized) contributed to thecollapse of many thrift institutions and the weakening of many banks in the 1980s. 12 In varying degree, dif ferences am ong regulators were evident in the development of policies relating to chartering new banks, the use of brokered funds, and capital require - ments. With respect to the entry of new banks, both the OCC and the states sharply in - creased chartering in the 1980s. (T exas—where branching was restricted—accounted for particularly lar ge shares of total new state and national bank charters.) In 1980, when the OCC sought to foster increased competition by allowing new entrants into banking mar - kets, the agency revised its requirements for approving new charters. But when a dispro - portionate number of new banks became troubled and failed, the FDIC expressed itsconcern about the OCC’ s policy . Abasic issue was the FDIC’ s ability to deny insurance coverage to newly chartered institutions. FDIC approval of insurance was, for all practicalpurposes, necessary before a state would grant a new charter , but national banks and Fed - eral Reserve member banks received insurance upon being chartered as a matter of law . Congress settled this issue in FDICIA by requiring that all institutions seeking insurance formally apply to the FDIC, thereby assuring the deposit insurer a role in new bank char - tering. Meanwhile, the number of new commercial bank charters reached a peak in 1984,then gradually declined until 1994. 13Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 13The regulators also dif fered on the appropriate treatment of brokered deposits. (Bro - kered deposits had a lar gely indirect influence on bank failures in that many weak savings institutions used them to fund rapid loan expansion in competition with healthier banks andthrift institutions.) In 1984, the FDIC and the Federal Home Loan Bank Board proposedthat brokered deposits be insured only up to $100,000 per broker per bank, whereas theOCC favored a less-stringent approach. Safety-and-soundness considerations seemed to bepitted against the objective of permitting evolution to proceed in the financial markets. Inthe end Congress stepped in, and both FIRREA and FDICIA limited the use of brokered de - posits by troubled institutions. Athird instance of regulatory disagreement concerned the adoption of formal capital requirements with uniform standards for minimum capital levels. In view of the relativelylow capital ratios at many lar ge banks and the rise in the number of failures, all of the agen - cies favored the objective of explicit capital standards, but initially they dif fered on the specifics; the FDIC generally favored higher capital requirements than the OCC, and theFederal Reserve of fered a compromise in at least one instance. In 1985, with congressional encouragement, the regulators agreed on a uniform system covering all banks. In 1990 afurther , major change came with the adoption of interim risk-based capital requirements, supplemented by leverage requirements. Capital standards became part of the triggeringmechanism for the Prompt Corrective Action (PCA) prescribed by FDICIA in 1991. Final risk-based requirements took ef fect in 1992. Geographic P atter n of Bank F ailur es The national economic, legislative, and regulatory factors discussed above af fected potentially all banks. Avariety of other factors af fected banks dif ferently in particular regions of the country , as indicated by the geographic pattern of bank failures. During the 1980–94 period, 1,617 FDIC-insured commercial and savings banks were closed orreceived FDIC financial assistance (see table 1.1). This number was 9.14 percent of the sum of all banks existing at the end of 1979 plus all banks chartered during the subsequent 15 years. The comparable figure for the preceding 15-year period (1965–79) was 0.3 percent. The geographic pattern of bank failures can be expressed in a number of ways—by number of failed banks, am ount of failed-bank assets, proportion of failed banks and failed- bank assets relative to all banks in individual states, or particular states’ shares in national totals for bank failures and failed-bank assets. But by any of these measures, it is evidentthat bank failures during the 1980–94 period were highly concentrated in relatively few re - gions—which, however , included some of the country’ s largest banking markets in terms of number of institutions and dollar resources. Thus, geographically confined crises were translated into a national problem. At one end of the scale, in 7 states the number of bankAn Examination of the Banking Crises of the 1980s and Early 1990s Volume I 14 Histor y of the Eighties—Lessons for the F utureTable 1.1 Bank Failur es by S tate, 1980–1994 Number of Bank Percent of Total Assets of Failed Banks Percent of Total Failur es Number of Banks ($Thousands) Bank Assets Alabama 9 2.47 $ 215,589 1.18 Alaska 8 44.44 1,083,417 41.58 Arizona 17 26.15 331,059 1.66 Arkansas 11 4.03 160,797 1.47 California 87 15.26 4,222,302 1.69 Colorado 59 12.39 1,035,553 5.24 Connecticut 32 18.39 6,818,223 22.17 Delaware 1 1.61 582,350 0.74 District of Columbia 5 17.86 1,135,066 13.39 Florida 39 4.56 4,524,461 4.30 Geor gia 3 0.53 60,922 0.17 Hawaii 2 20.00 13,941 0.29 Idaho 1 3.13 42,931 0.84 Illinois 33 2.52 35,031,196 25.75 Indiana 10 2.40 241,463 0.76 Iowa 40 6.07 652,681 3.25 Kansas 69 10.71 1,233,874 7.26 Kentucky 7 1.91 97,742 0.48 Louisiana 70 22.44 4,105,621 17.39 Maine 2 2.63 875,303 13.51 Maryland 2 1.45 43,827 0.06 Massachusetts 44 10.63 10,240,719 12.90 Michigan 3 0.75 159,917 0.29 Minnesota 38 4.87 1,491,250 4.95 Mississippi 3 1.63 338,680 3.18 Missouri 41 5.24 1,043,379 2.25 Montana 10 5.75 172,739 3.32 Nebraska 33 6.88 323,646 2.91 Nevada 1 4.17 18,036 0.10 New Hampshire 16 12.60 3,320,916 31.98 New Jersey 14 5.71 4,695,156 9.49 New Mexico 11 11.00 568,326 9.47 New York 34 8.79 31,701,442 6.22 North Carolina 2 1.59 74,553 0.27 North Dakota 9 5.00 77,565 1.76 Ohio 5 1.14 171,765 0.29 Oklahoma 122 22.02 5,838,273 23.85 Oregon 17 17.00 599,703 4.34 Pennsylvania 5 1.19 17,454,150 16.99 Puerto Rico 5 33.33 527,375 8.94Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 15Table 1.1 (continue d) Bank Failur es by S tate, 1980–1994 Number of Bank Percent of Total Assets of Failed Banks Percent of Total Failur es Number of Banks ($Thousands) Bank Assets Rhode Island 2 8.33 323,861 3.29 South Carolina 1 0.87 64,629 0.67 South Dakota 8 4.73 711,345 4.04 Tennessee 36 9.05 1,730,076 6.34 Texas 599 29.41 60,192,424 43.84 Utah 11 11.58 339,237 4.04 Vermont 2 5.41 93,802 2.94 Virginia 7 2.45 133,529 0.47 Washington 4 2.63 713,803 2.42 West Virginia 5 1.98 123,829 1.25 Wisconsin 2 0.30 50,882 0.19 Wyoming 20 16.67 375,332 10.30 U.S. 1,617 9.14% $206,178,657 8.98% Note: Data refer to FDIC-insured commercial and savings banks that were closed or received FDIC assistance. Total num - ber of banks is the number of banks on December 31, 1979, plus banks newly chartered in 1980–94. Asset data are assets of banks existing on December 31, 1979, plus assets of newly chartered banks as of date of failure, mer ger, or December 31, 1994, whichever is applicable, and first available assets for Massachusetts banks that became FDIC-insured in the mid-1980s.Data exclude 13 newly chartered banks that reported no asset figures and 4 banks in U.S. territories. 14The 8.98 percent figure refers to the failed-bank portion of the following: assets of all banks existing as of December 31, 1979, plus assets of banks chartered in 1980–94 as of the date of mer ger, failure, or December 31, 1994, whichever is ap - plicable, and first available assets for Massachusetts banks that became FDIC-insured in the mid-1980s. Data are not ad - justed for inflation.failures constituted at least 20 percent of the total number of existing and new banks (Alaska, Arizona, Hawaii, Louisiana, Oklahoma, Puerto Rico, and Texas). At the other end of the scale, in 24 states bank failures represented less than 5 percent of the total number ofexisting and new banks. Of the total 1,617 failures during the entire 1980–94 period, nearly60 percent were in only 5 states: California, Kansas, Louisiana, Oklahoma, and Texas. Included in these numbers are failures of bank holding company subsidiaries; in Texas and other states with branching restrictions, these were m ore like branches than independent institutions. An alternative measure of the severity of bank failures is based on assets. Assets of banks failing in 1980–94 constituted 8.98 percent of the sum of total bank assets at the endof 1979 plus the assets of banks chartered during the 1980–94 period. 14In 6 states (Alaska, Connecticut, Illinois, New Hampshire, Oklahoma, and Texas), failed-bank assets consti -An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 16 Histor y of the Eighties—Lessons for the F uturetuted at least 20 percent of total assets at year -end 1979 plus new-bank assets. On the other hand, in 33 states the failed-bank share was less than 5 percent. Of all banks that failed dur - ing the 1980–94 period, 59 percent of assets at the quarter before failure were accounted forby 3 states: Illinois, New York, and Texas. (See table 1.2.) 15 Although widespread bank failures were limited to a few areas of the country , even a relatively “small” number of failures could cause serious strains on the deposit insurancefund. In 1988, for example, the number of failures and the am ount of failed-bank assets reached post-Depression records of 279 and $54 billion (nominal dollars), respectively , but still represented in each case less than 2 percent of the total number of banks and total bankassets at the beginning of the year . Nevertheless, in that year the FDIC sustained the first operating loss in its history , and operating losses continued through 1991, after which, pro - visions for insurance losses were sharply reduced. And even the smaller number of failures before 1988 had an evident ef fect on the FDIC’ s income and expense position. Beginning in 1984, provisions for insurance losses exceeded annual deposit insurance assessments,and this shortfall continued through 1990. 16 The figures by state illustrate some of the factors associated with bank failures. The incidence of failure was particularly high in states characterized by •severe economic downturns related to the collapse in ener gy prices (Alaska, Louisiana, Oklahoma, Texas, and Wyoming); •real estate–related downturns (California, the Northeast, and the Southwest); •the agricultural recession of the early 1980s (Iowa, Kansas, Nebraska, Oklahoma, and Texas); •an influx of banks chartered in the 1980s (California and Texas) and the parallel phe - nomenon of mutual-to-stock conversions (Massachusetts); •prohibitions against branching that limited banks’ ability to diversify their loan portfo - lios geographically and to fund growth through core deposits (Colorado, Illinois,Kansas, Texas, and Wyoming); 17 •the failure of a single lar ge bank (Illinois) or of a small number of relatively lar ge banks (New York and Pennsylvania). 15Comparisons based on assets of failed banks are subject to distortion because of the ef fect of inflation, dif ferences in the timing of failures am ong the states, and dif ferences in asset dates between new banks and banks existing at year -end 1979. 16Beginning in 1989, data refer to the Bank Insurance Fund (FDIC, Annual Repor t, various years). 17Information on state branching provisions is as of September 30, 1985, as compiled by the Conference of S tate Bank Su - pervisors. CSBS listed 7 states as having unit banking as of September 30, 1985, 6 as a result of legal prohibitions (Col - orado, Illinois, Kansas, Montana, North Dakota, and Texas). One (W yoming) had no statute, but unit banking was prevalent. Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 17Table 1.2 Assets of Failed Banks at the Quar terbefor e Failur e, by S tate, 1980–1994 Assets of Failed Banks State ($Thousands) Percent Distribution Alabama $ 266,443 0.08 Alaska 3,049,573 0.96 Arizona 453,522 0.14 Arkansas 229,700 0.07 California 6,018,036 1.90 Colorado 1,072,556 0.34 Connecticut 17,717,959 5.59 Delaware 582,350 0.18 District of Columbia 2,189,658 0.69 Florida 15,471,515 4.88 Geor gia 104,607 0.03 Hawaii 11,486 0.00 Idaho 55,867 0.02 Illinois 40,765,430 12.87 Indiana 311,825 0.10 Iowa 809,089 0.26 Kansas 1,697,588 0.54 Kentucky 114,931 0.04 Louisiana 4,616,370 1.46 Maine 2,228,177 0.70 Maryland 57,000 0.02 Massachusetts 26,632,401 8.41 Michigan 160,300 0.05 Minnesota 1,669,974 0.53 Mississippi 288,949 0.09 Missouri 3,096,719 0.98 Montana 212,896 0.07 Nebraska 402,185 0.13 Nevada 18,036 0.01 New Hampshire 5,393,842 1.70 New Jersey 6,919,198 2.18 New Mexico 723,576 0.23 New York 51,577,291 16.28 North Carolina 74,553 0.02 North Dakota 120,109 0.04 Ohio 152,254 0.05 Oklahoma 6,712,651 2.12 Oregon 622,091 0.20 Pennsylvania 14,265,742 4.50 (continued)An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 18 Histor y of the Eighties—Lessons for the F utureTable 1.2 (continue d) Assets of Failed Banks at the Quar terbefor e Failur e, by S tate, 1980–1994 Assets of Failed Banks State ($Thousands) Percent Distribution Puerto Rico 543,748 0.17 Rhode Island 600,706 0.19 South Carolina 64,629 0.02 South Dakota 743,698 0.23 Tennessee 2,446,083 0.77 Texas 93,061,510 29.37 Utah 469,637 0.15 Vermont 329,478 0.10 Virginia 296,368 0.09 Washington 769,109 0.24 West Virginia 123,139 0.04 Wisconsin 70,757 0.02 Wyoming 428,606 0.14 U.S. $316,813,917 100.00% Note: Failed-bank assets are assets as of the quarter before failure or assistance, or assets as of the last available Call Report before failure or assistance. In some states bank failures were af fected by m ore than one of these factors. For ex - ample, the particularly high incidence of failures in Texas reflected the rapid rise and sub - sequent collapse in oil prices, the commercial real estate boom and bust, the ef fects of the agricultural recession, the lar ge number of new banks chartered in the state during the 1980s, and state prohibitions against branching. (The high proportion of bank failures inTexas also reflected supervisory developments. As noted below , declines in the number and frequency of on-site examinations in the 1983–86 period were particularly pronounced inTexas; earlier identification of troubled banks might have prevented some failures.) 18By the same token, some states that exhibited only one or two of the factors associated withbank failures had relatively few failures. Montana and North Dakota, for example, had pro - hibitions against branching, but their failure rates were below the national average, whethermeasured by number of institutions or by assets. Dif ferences am ong the states in failure rates and in the presence or absence of factors associated with failures illustrate the conclu - sion that the rise in the number of bank failures cannot be ascribed to any single cause. 18Texas was also a leading state for S&L failures. Texas S&Ls accounted for 18 percent of all of the failures resolved by the Resolution Trust Corporation (R TC), 14 percent of S&L assets at time of takeover , and 29 percent of total estimated R TC resolution costs. See R TC, Statistical Abstract (August 1989/September 1995).Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 19Regional and Sectoral R ecessions Although the interplay of broad economic, legislative, and regulatory forces helped make the environment for banking increasingly demanding, the m ore immediate cause of bank failures was a series of regional and sectoral recessions. Because m ost U.S. banks served relatively narrow geographic markets, these regional and sectoral recessions had asevere impact on local banks. It should be noted, however , that not all regional recessions of the magnitude experienced during the 1980–94 period resulted in a major increase in thenumber of bank failures. Rather , bank failures were generally associated with regional re - cessions that had been preceded by rapid regional expansions—that is, they were associatedwith “boom-and-bust” patterns of economic activity . Bank loans helped to fuel the boom phase of the cycle, and when economic activity turned down, some of these loans went sour , with the result that banks holding these loans were weakened. By contrast, recessions thatwere preceded by relatively slow economic activity , such as those in the Rust Belt, gener - ally did not lead to widespread bank failures. This relationship between the number of bank failures and regional boom-and-bust patterns of economic activity is illustrated by the data in tables 1.3 and 1. 4, which show that bank failure rates were generally high in states where, in the five years preceding state re - cessions, real personal income grew faster than it did for the nation as a whole. Conversely , bank failure rates were relatively low in states where, in the five years preceding state re - cessions, real personal income grew m ore slowly than it did for the nation as a whole. 19 There were four major regional and sectoral economic recessions that were associated with widespread bank failures during the 1980–94 period. The first accompanied the down - turn in farm prices in the early and middle 1980s after years of rapid increases during thelate 1970s (see figure 1. 5). The downturn in prices led to reductions in net farm income and farm real estate values and a rise in the number of failures of banks with heavy concentra - tions of agricultural loans. The second recession occurred in Texas and other ener gy- producing southwestern states, where gross state product dropped after oil prices turneddown in 1981 and again in 1985 (see figure 1. 6). The 1981 oil price reduction was followed by a regional boom and bust in commercial real estate activity . The third recession was in the northeastern states, which experienced negative growth in gross state product in1990–91. The final episode was a recession in California, as growth in gross state product turned negative in 1991–92. Of the 1,617 bank failure and assistance cases from 1980 to 1994, 78 percent were lo - cated in the regions suf fering these economic downturns—the Southwest, the Northeast, 19In some high-growth states the number of bank failures rose sharply after the states’ recessions, but the increase fell out - side the three-year periods shown in table 1.3. For example, Arizona experienced especially rapid growth before the state’ s 1982 recession and also saw a high rate of bank failures (tables 1.1 and 1.2), but m ost of them occurred in 1989–90.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 20 Histor y of the Eighties—Lessons for the F utureTable 1.3 Bank Failur es and Gr owth Rates of Real Personal Income, by S tate, 1980–1994 (Per cent) Growth Rates of Real Personal Income Five Years befor e Recession State Minus Percent of Banks Failing Recession State Growth Rate, State Growth U.S. Growth in Recession and Next 2 State* Years† Recession Years Rate Rate Years‡ Wyoming 1982–87 −3.03 8.26 5.05 18.52 Nevada 1982 −0.17 7.83 4.62 8.33 Oklahoma 1983–87 −1.42 6.05 3.78 20.83 Alaska 1986–87 −5.46 6.63 3.75 50.00 Arizona 1982 −0.18 6.69 3.49 0.00 New Hampshire 1990–91 −0.43 5.69 2.50 19.51 Louisiana 1983–87 −0.75 4.69 2.41 21.22 Washington 1982 −0.24 4.97 1.76 0.93 Maryland 1991 −0.33 4.49 1.61 1.92 Texas 1986–87 −0.98 4.43 1.55 20.45 Maine 1991 −2.15 4.42 1.54 5.13 Vermont 1991 −1.45 4.32 1.44 6.25 Connecticut 1991 −1.94 4.30 1.42 22.05 California 1991 −1.04 4.20 1.32 7.26 Oregon 1981–82 −2.40 5.03 1.21 14.63 New Jersey 1991 −1.13 3.89 1.01 6.00 Rhode Island 1991 −1.82 3.79 0.91 13.33 Massachusetts 1991 −1.87 3.79 0.91 9.77 New York 1991 −0.88 3.71 0.83 3.86 Mississippi 1980 −1.09 4.15 0.42 0.00 Arkansas 1980–82 0.27 4.14 0.42 2.33 Kentucky 1980–83 0.17 4.08 0.36 0.58 Tennessee 1982 −0.05 3.12 −0.09 7.41 West Virginia 1981–83 −0.73 3.63 −0.19 0.84 Illinois 1991 −0.09 2.64 −0.24 0.55 Missouri 1980–82 0.55 3.41 −0.32 0.69 Wisconsin 1981–82 −0.22 3.49 −0.33 0.00 North Dakota 1985–88 −3.54 2.28 −0.38 4.52 Kansas 1980 −0.30 3.32 −0.41 0.49 Idaho 1982 −1.91 2.79 −0.41 0.00 Michigan 1991 −0.58 2.41 −0.47 0.00 Alabama 1982 −0.24 2.72 −0.48 0.97 Michigan 1980–82 −2.73 3.12 −0.60 0.54 Hawaii 1981 −0.63 3.20 −0.62 0.00 Indiana 1980–82 −1.39 3.03 −0.69 0.49 Iowa 1979–85 −0.31 1.83 −0.79 4.92 Iowa 1991 −0.39 2.04 −0.84 0.18Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 21Table 1.3 (continue d) Bank Failur es and Gr owth Rates of Real Personal Income, by S tate, 1980–1994 (Per cent) Growth Rates of Real Personal Income Five Years befor e Recession State Minus Percent of Banks Failing Recession State Growth Rate, State Growth U.S. Growth in Recession and Next 2 State* Years† Recession Years Rate Rate Years‡ Montana 1980–82 1.21 2.87 −0.86 0.62 Nebraska 1979–83 0.24 1.67 −0.96 4.20 Montana 1985–88 −0.17 1.39 −1.28 4.79 Ohio 1980–82 −0.73 2.41 −1.31 0.00 Illinois 1980–82 −0.28 2.34 −1.38 1.60 South Dakota 1980–82 −1.38 2.09 −1.63 1.30 West Virginia 1987 −1.33 0.51 −2.65 0.47 North Dakota 1991 −2.50 0.08 −2.80 0.00 Iowa 1988 −1.11 1.01 −3.09 1.17 District of Columbia 1980 −2.94 −0.08 −3.80 0.00 North Dakota 1979–80 −3.54 −1.59 −4.21 0.58 Note: Data refer to all states that experienced a decrease in real personal income in any year from 1980 to 1992. *States are ranked according to the magnitude of the dif ference between state growth rates and the U.S. growth rate in real personal income during the five years before state recessions. †Recessions are defined as years in which personal income deflated by GDP deflator decreased. Recoveries are counted as having at least two consecutive years of growth in real personal income. In some states, therefore, personal income increasedduring a single year suf ficiently to produce positive growth for the recession as a whole. ‡Percent of banks failing is based on the number of banks existing as of December of the year preceding the recession. Table 1.4 Bank Failur es and Gr owth Rates of Real Personal Income, by State Recession Quar tile (Percent) Average Dif ference between S tate Average S tate Bank Failur e State Recession Growth Rate and U.S. Gr owth Rate, Rate in Recession Quar tile* 5 Years befor e Recession† and Next 2 Years 1 2.79 14.42 2 0.71 7.34 3 −0.48 1.06 4 −2.07 1.28 *State recessions are grouped in quartiles according to the magnitude of the dif ference between state growth rate and U.S. growth rate in real personal income from table 1.3. †Data are unweighted averages of individual state data.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 22 Histor y of the Eighties—Lessons for the F uture$Billions $BillionsIndex $Billions Average Farm Real Estate Value per AcreFarm Debt Source:Economic Report ofthePresident ,1986, 1996.Figur e 1.5 Farm Prices, Exports, Income, Debt, and Real Estate Value, 1975 94–19 Farm ExportsPrices Received by Farmers Net Farm Income Dollars1990 1992 =100 – 1975 1980 1985 1990 1994253545 1975 1980 1985 1990 1994203040 1975 1980 1985 1990 1994901201501801975 1980 1985 1990 19948090100 1975 1980 1985 1990 1994400600800Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 23Figur e 1.6 Changes in Gr oss S tate Pr oduct and Gr oss Domestic Pr oduct, 1980–1994 Source:U.S. Department of Commerce, Bureau of Economic Analysis.CaliforniaNortheast Southwest Peak Number of Failur esPeak Number of Failur es Peak Number of Failur esPercent Percent Percent1980 1982 1984 1986 1988 1990 1992 1994-30369 1980 1982 1984 1986 1988 1990 1992 1994-303691980 1982 1984 1986 1988 1990 1992 1994-30369 U.S. U.S. U.S.Southwest Northea st Califo rniaand California—or were agricultural banks outside of these three regions.20These failures accounted for 71 percent of the assets of failed banks over the period. Although all four of 20Agricultural banks are defined as banks with 25 percent or m ore of total loans in agricultural loans. Data on assets of failed banks are as of the quarter before the date of failure. The Southwest includes Arkansas, Louisiana, New Mexico, Okla - homa, and Texas. The Northeast includes New Jersey , New York, and the six New England states (Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont). The bulk of the agricultural bank failures, other than those in the two southwestern states of Oklahoma and Texas, were in Iowa, Kansas, Minnesota, Missouri, and Nebraska.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 24 Histor y of the Eighties—Lessons for the F uture21See Chapter 8, “Banking and the Agricultural Problems of the 1980s.” 22See John O’Keefe, “The Texas Banking Crisis: Causes and Consequences, 1980–1989,” FDIC Banking Review 3, no. 2 (1990); and Chapter 9, “Banking Problems in the Southwest.” 23See Chapter 1 1, “Banking Problems in California.” 24See Chapter 10, “Banking Problems in the Northeast.”the recessions associated with bank failures were partly shaped by their own distinct cir - cumstances, certain comm on elements were present: 1.Each followed a period of rapid expansion; in most cases, cyclical for ces wer e ac- centuated by external factors. 2.In all four recessions, speculative activity was evident. “Exper t” opinion often gave suppor t to overly optimistic expectations. 3.In all four cases ther e wer e wide swings in r eal estate activity , and these con - tributed to the severity of the r egional r ecessions. 4.Commer cial r eal estate markets in par ticular deserve attention because boom and bust activity in these markets was one of the main causes of losses at both failedand surviving banks. Rapid expansion . In the agricultural belt, increased farm production and purchases of farmland were stimulated by rapid inflation during the 1970s in the prices of farm products,a sharp run-up in farm exports, and widespread expectations of strong worldwide demandin the 1980s. But as farm exports declined and higher interest rates increased farm costs, theexpansion gave way to a downturn. 21Similarly , in the Southwest (as well as other oil-pro - ducing areas around the world) strong worldwide demand for oil plus OPEC restrictions onsupply led to a major rise in oil prices and strong economic expansion—but the weakeningin oil prices after 1981 and their rapid drop in 1985 (brought on partly by the collapse of dis - cipline in the international oil cartel) resulted in two economic downturns during the 1980sin the Southwest. 22California enjoyed a rate of economic growth above the national aver - age during the 1980s but was hit particularly hard during the 1991–92 national recession,partly because of cutbacks in defense spending. 23In the Northeast, growth rates in overall production were above the national average during 1982–88; the subsequent decline cameabout mainly because a local economic slowdown was followed—and aggravated—by the1991–92 national economic recession and by a boom and bust in northeastern residentialand commercial real estate activity . 24 Speculative activity with “exper t” suppor t. Speculative activity was reflected in a number of developments. Farm real estate values showed an uninterrupted rise in the late1970s and early 1980s, even though gross returns per acre for major crops were tracing aChapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 2525In 1982, when land values reached their zenith, gross rates of return for corn and soybeans were less than two-thirds their 1970 levels and approximately one-third their 1973 levels. See Chapter 8, “Banking and the Agricultural Problems of the 1980s.” 26See Chapter 3, “Commercial Real Estate and the Banking Crises”; and O’Keefe, “The Texas Banking Crisis.” 27“Speculative activity” in this context is synonym ous with economic “bubbles” defined as follows: “if the reason that the price is high today is only because investors believe that the selling price will be high tom orrow—when “fundamental” fac - tors do not seem to justify such a price—then a bubble exists.” See Joseph E. S tiglitz, “Symposium on Bubbles,” Journal of Economic Perspectives 4, no. 2 (spring 1990): 13. 28Robert Ber gland, secretary of agriculture in 1980, said, “The era of chronic overproduction...is over .” In 1972, then-Secre - tary of Agriculture Earl Butz is said to have advised farmers to plant “from fencerow to fencerow .” (Both quotations are from Gregg Easterbrook, “Making Sense of Agriculture: ARevisionist Look at Farm Policy ,” The Atlantic 256 (July 1985): 63. See Chapter 8, “Banking and the Agricultural Problems of the 1980s.” 29Interviews with regulators and bankers. See Chapter 10, “Banking Problems in the Northeast.” 30See citations in Chapter 1 1, “Banking Problems in California.”highly variable and generally downward trend.25In the Southwest, commercial construc - tion and lending activity continued in major markets after vacancy rates began to soar . In many commercial real estate m ortgage markets, underwriting standards were relaxed.26The presence of speculative activity was frequently mentioned in interviews conducted in 1995by staf f of the FDIC’ s Division of Research and S tatistics as part of the research for this study . 27(In all, approximately 150 bankers and regulators were interviewed in Atlanta, Boston, Dallas, Kansas City , New York, San Francisco, and Washington). Numerous inter - viewees cited a belief comm on in the 1980s that the boom economies of this period had un - limited viability . They also noted that in many cases bankers were engaged in asset-based lending, relying on collateral values supported by inflationary expectations rather than bycash flows. Examples of “expert” opinion that supported optimism included statements attributed to two secretaries of agriculture 28and comments by many observers in the Northeast that the area’ s economy was diversified, mature, and lar gely immune to Texas-style real estate problems.29Another example is provided by economists and other analysts, who as late as 1990 and 1991 were discounting the prospect of a bust in California home prices.30 Wide swings in r eal estate activity . In the agricultural belt, prices of farmland were bid up during the 1970s by farmers and investors, who were responding to increases in theprices of farm products as well as expectations of continued strong foreign demand. Farm - land values continued to rise until 1982, remained at high levels until 1984, and then col - lapsed (figure 1.5). In the Southwest, both residential and nonresidential construction rosesharply during the early 1980s before falling precipitously later in the decade; these widereal estate swings followed the earlier oil-generated cycle and contributed to the secondSouthwest recession in the 1980s. In both the northeastern states and California, boom-and-bust real estate activity aggravated general state recessions in the early 1990s.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 26 Histor y of the Eighties—Lessons for the F utureCommer cial r eal estate markets and bank losses . Commercial real estate development is inherently risky , partly because of the long gestation period of many commercial con - struction projects. When completed projects finally come to market, demand conditions may have changed considerably from what they were at the time of conception. Another cause of risk is that many firms seeking commercial floor space are geographically m obile, so developers are af fected by economic events not only in the project’ s proximity but in far - distant areas as well. In addition, commercial real estate projects tend to be highly lever - aged, a condition that increases the volatility of returns. Relevant data on commercial realestate are often dif ficult to obtain because these markets are not highly or ganized and be - cause transactions are often “private deals” whose crucial elements may not be publiclyavailable. Finally , commercial loan contracts usually have nonrecourse provisions prohibit - ing lenders from satisfying losses from other borrower assets. In the early 1980s, booming activity in commercial construction was supported by rapidly increased bank and thrift commercial m ortgage lending. Amajor stimulus for this activity was provided by public policy actions: tax breaks enacted as part of the EconomicRecovery Act of 1981 greatly enhanced the after -tax returns on real estate investment, and the Garn–S t Germain Act expanded the nonresidential lending powers of savings associa - tions. Competitive pressures, including those reflected in the reduced bank share of themarket for business loans to lar ge companies, also provided an important stimulus. Many banks and thrifts m oved aggressively into commercial real estate lending. Dur - ing the 1980s, when total real estate loans of banks m ore than tripled, commercial real es - tate loans nearly quadrupled. As a percentage of total bank assets, total real estate loans rose from 18 to 27 percent between 1980 and 1990, while the ratio for nonresidential and con - struction loans nearly doubled, from 6 to 1 1 percent. Apervasive relaxation of underwriting standards took place, unchecked either by the real estate appraisal system or by supervisoryrestraints. Overly optimistic appraisals, together with the relaxation of debt coverage, ofmaximum loan-to-value ratios, and of other underwriting constraints, meant that borrowersfrequently had no equity at stake, and lenders bore all of the risk. 31 Overbuilding occurred in many markets, and when the bubble burst, real estate values collapsed. (The downturn was aggravated by the Tax Reform Act of 1986, which rem oved tax breaks for real estate investment and caused a reduction in after -tax returns on such in - vestment.) At many financial institutions loan quality deteriorated significantly , and the de - terioration caused serious problems. As discussed in detail below , banks that failed in the 1980s had higher ratios of commercial real estate loans to total assets than surviving banks. 31These observations on underwriting practices, taken from Chapter 3, reflect the comments of, and have been reviewed by , a number of FDIC examiners and supervisory personnel who were actively engaged in bank examination and supervisionduring the 1980s.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 27Failing banks also had higher ratios of commercial real estate loans to total real estate loans, of real estate char ge-of fs to total char ge-of fs, and of nonperforming real estate assets to to - tal nonperforming assets. Bank P erformance in R egional and Sectoral R ecessions The behavior of banks in the regions and sectors that suf fered recessions during the 1980s also exhibited some comm on elements: 1.In the economic expansions that preceded these recessions, banks generally responded aggressively to rising credit demands. 2.Banks that failed during the regional recessions generally had assumed greater risks, on average, than those that survived, as measured by ratios of total loans and commercialreal estate loans to total assets. Banks that failed had generally not been in a seriouslyweak condition (as measured by equity-to-assets ratios) in the years preceding the re - gional recessions. 3.Banks chartered in the 1980s and mutual institutions converting to the stock form of ownership failed with greater frequency than comparable banks. Aggr essive r esponse . In the case of agricultural banks, aggressive response is evident in the growth of farm loans, which increased rapidly and reached a peak in 1984, after the1981 highs in prices received by farmers and net farm income and the 1982 high in farm - land values. In Texas, banks responded to the rise in oil prices by rapidly increasing not only their commercial and industrial loans (including loans to oil and gas producers) but also theshare of commercial and industrial loans in total bank assets. In m ost of the regions that un - derwent recessions, the aggressiveness of bank lending is evident as well in the rapid ex - pansion in nonresidential m ortgage lending and in the increased share of commercial mortgages in total bank assets. Risk taking and failur e. Banks that would fail during the 1980–94 period generally had higher ratios of total loans to assets and commercial real estate loans to assets throughoutmost of the period (see figures 1. 7 and 1. 8). (In this context, commercial real estate loans in - clude construction loans, nonfarm nonresidential loans, and multifamily m ortgages.) This was true for banks in the agricultural belt, the Southwest, the Northeast, California, and thetotal United S tates. In the agricultural belt, the Southwest, and the Northeast, banks that would fail during the regional recessions had significantly higher loans-to-assets ratios inthe year before the recessions began (see table 1.5). 32In the Northeast and Southwest, com - 32Regional recessions are considered to have begun in the agricultural belt in 1982 (following the 1981 high in prices re - ceived by farmers), in the Southwest in 1982 (after oil prices reached a peak in 1981), and in the Northeast and Californiain the first year of negative gross state product (figure 1.6).An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 28 Histor y of the Eighties—Lessons for the F uturePercentPercent PercentPercent Total U.S.CaliforniaFigur e 1.7 Ratio of Gr oss Loans to TotalAssets, Failed and Nonfailed Banks, 1980 94–19 Southwest Northeast PercentAgricultural Banks* Note: Data areunweighted averages ofindividual bank ratios. Data forbanks thatsubsequently failed arenotshown foryears when there were fewer thantenbanks thatwould failinsubsequent years. Open-bank assistance cases arenotcounted asfailures.Banks That Subsequently Failed Banks That Did Not FailAgricultural banks are banks where agricultural loans are at least 25% of total loans.1980 1982 1984 1986 1988 1990 1992 199445 *505560 1980 1982 1984 1986 1988 1990 1992 1994455565 1980 1982 1984 1986 1988 1990 1992 1994556575 1980 1982 1984 1986 1988 1990 1992 1994556575 1980 1982 1984 1986 1988 1990 1992 1994506070Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 29Percent 1980 1982 1984 1986 1988 1990 1992 19940102030PercentPercent PercentPercent Total U.S.CaliforniaFigur e 1.8 Ratio of Commer cial Real Estate Loans to TotalAssets, Failed and Nonfailed Banks, 1980 94–19 Southwest NortheastAgricultural Banks* Note: Commercial realestate loans =construction loans +multifamily loans +nonfarm, nonresidential loans. Data are unweighted averages ofindividual bank ratios ofcommercial realestate loans tototal assets. Data forbanks thatsubsequently failed arenotshown foryears when there were fewer than tenbanks thatwould failinsubsequent years. Open-bank assistance cases arenotcounted asfailures.Banks That Subsequently Failed Banks That Did Not FailAgricultural banks are banks where agricultural loans are at least 25% of total loans.1980 1982 1984 1986 1988 1990 1992 1994 1980 1982 1984 1986 1988 1990 1992 1994 1980 1982 1984 1986 1988 1990 1992 1994 1980 1982 1984 1986 1988 1990 1992 1994102030 10152025345 *61014An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 30 Histor y of the Eighties—Lessons for the F utureTable 1.5 Selected Financial Ratios A. Failed and Nonfailed Banks 1 Yearbefor e Regional Recession 1981 1989 1990 Agricultural Banks Southwest Banks Northeast Banks California Banks Ratio Failed Nonfailed Failed Nonfailed Failed Nonfailed Failed Nonfailed Equity/Assets 7.91% 8.30%* 7.00% 7.63%* 6.67% 9.21%* 5.71% 10.47%* Eq.+Loss Res./Assets 9.11 9.77* 8.64 9.25* 8.34 9.93 7.20 11.46* Nonprfm Lns/T ot Lns NA NA NA NA 8.60 2.95* 6.23 2.39* ROA 1.26 1.33 1.22 1.38* -1.68 0.67* -0.63 0.36 ROE 16.90 16.44 18.98 18.99 -23.65 6.73* -7.78 9.88* Loans/Assets 56.30 48.48* 53.94 47.72* 75.16 68.05* 73.12 69.63 Comm. Mtgs/Assets 2.08 2.19 3.92 3.42* 13.91 9.44* 10.79 11.91 B. Failed and Nonfailed Banks 3 Years befor e Regional Recession 1979 1987 1988 Agricultural Banks Southwest Banks Northeast Banks California Banks Ratio Failed Nonfailed Failed Nonfailed Failed Nonfailed Failed Nonfailed Equity/Assets 7.39% 7.87%* 6.94% 7.45%* 7.96% 8.86%* 6.95% 9.58% Eq.+Loss Res./Assets 8.85 9.45* 8.45 9.08* 8.53 9.37 8.02 10.52 Nonprfm Lns/T ot Lns NA NA NA NA 1.70 1.14* 4.86 2.28* ROA 1.15 1.28* 1.00 1.28* 0.62 1.04* 0.08 0.78* ROE 16.10 16.64 15.55 17.80* 11.66 14.32 2.29 10.85 Loans/Assets 58.40 55.56* 53.42 50.02* 74.31 66.33* 68.72 63.01* Comm. Mtgs/Assets 2.13 2.42* 3.99 3.71 13.08 8.25* 7.78 8.76 Note: Data are unweighted averages of individual bank ratios. Asset and loan figures are year -end values of the given year , and equity figures are year -end of the previous year . Excluded were banks chartered within the specified year , banks that failed before the recession, and banks participating in the Net Worth Certificate Program. Nonperforming loans were not re - ported before 1982. *Significant at 95 percent level 33The comparison in California is between failing and surviving banks with assets below $300 million. All but one of the state’ s bank failures were in that asset-size group, while the total state data are dominated by California’ s four megabanks (see Chapter 1 1). mercial m ortgages were higher relative to total assets for failed banks. Banks that would fail also had lower equity-to-assets ratios than survivors in the year before the recession.33Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 3134The 1980–90 period was selected in this comparison to compensate roughly for the fact that banks chartered between 1991 and 1994 did not have as much chance to fail during the period through 1994.Three years before the onset of the regional recessions, banks that would fail likewise had significantly higher ratios of loans to assets, but these banks’ equity-to-assets ratios—al - though somewhat lower than those of banks that would survive—were in the generallyhealthy range of nearly 7 percent to nearly 8 percent (table 1. 5). These results are generally consistent with the findings on measures of risk and con - dition summarized below in the section on of f-site surveillance. As noted in that section, five years before their failure, banks that would subsequently fail dif fered little from banks that would survive in terms of equity-to-assets ratios and other measures of current condi - tion. On the other hand, banks that would fail had higher loans-to-assets ratios than sur - vivors, and high loans-to-assets ratios were the risk factor with the strongest statisticalrelationship to incidence of failure five years later . Although high loan volumes were a prominent feature of failing banks from 1980 to 1994, they obviously were not an automatic route to failure. Banks earn income by manag - ing risk, including risk of loan defaults. The averages of individual bank ratios discussed above obscure the fact that some banks that survived also had high concentrations of assetsin total loans and/or commercial m ortgages. Similarly , as noted below in the section on of f- site surveillance, only a fraction of the banks with high loans-to-assets ratios would fail fiveyears later . The conditions enabling many banks with high-risk financial characteristics to survive the recessions and avoid failure may include the following, am ong others: strong equity and reserve positions to absorb losses, m ore-favorable risk/return trade-of fs, superior lending and risk-management skills, changes in policies before high risk was translated intosevere losses, improvements in local economic conditions, and timely supervisory actions.High lending volumes may lead to trouble if a bank achieves them by relaxing credit stan - dards, entering markets where management lacks expertise, or making lar ge loans to single borrowers, or if loan growth strains the bank’ s internal control systems or back-of fice oper - ations. That such factors were present at many banks that failed from 1980 to 1994 has been suggested by numerous observers, including those interviewed during the research for thisstudy . New and conver ted banks . Approximately 2,800 new banks were chartered in the pe - riod covered by this study , 39 percent of them in the Southwest (notably Texas) and Cali - fornia. Of all the institutions chartered in 1980–90, 3416.2 percent failed through 1994, compared with a 7.6 percent failure rate for banks that were already in existence on De -An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 32 Histor y of the Eighties—Lessons for the F utureTable 1.6 Failur e Rates, Newly Char tered and Existing Banks Banks Char tered, 1980–1990 Number Failed Percent Failed Region 1980–1994 1980–1994 Southwest 248 33.3 Southeast 26 4.3 Northeast 38 19.3 California 41 13.1 U.S. 420 16.2 Banks Existing on December 31, 1979 Number Failed Percent Failed Region 1980–1994 1980–1994 Southwest 538 21.4 Southeast 77 3.1 Northeast 89 8.5 California 31 12.8 U.S. 1,114 7.6 35Astudy of the Texas experience concluded that “the relatively high failure rate for newly established Texas banks can be explained by high-risk financial policies” (Jef fery W. Gunther , “Financial S trategies and Performance of Newly Estab - lished Texas Banks,” Federal Reserve Bank of Dallas Financial Industr y Studies [December 1990]: 13). 36In the Southwest and Northeast, newly chartered banks failed with greater frequency than preexisting banks, whether “newly chartered” includes all banks chartered during the 1980–90 period or only those that were in existence for five yearsor less. In Southern California, however , failure rates for banks in existence for five years or less were lower than those for preexisting banks, whereas failure rates for all banks chartered in the entire 1980–90 period were higher . 37Jennifer L. Eccles and John P . O’Keefe, “Understanding the Experience of Converted New England Savings Banks,” FDIC Banking Review 8, no. 1 (1995): 1–18.cember 31, 1979 (see table 1. 6).35Although the data are dominated by the Texas experience, in most areas banks chartered in the 1980s generally had a higher failure rate than banks ex - isting at the beginning of the 1980s.36 In the Northeast, mutual savings banks that converted to the stock form of ownership represented a somewhat comparable phenomenon.37Of the mutuals that converted in the middle and late 1980s after state legislation permitted such action, 21 percent of the insti - tutions existing at the end of 1989 failed in 1990–94. This compared with 8 percent of theChapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 33Table 1.7 Failur e Rates of Conver ted Mutual Savings Banks and Other Banks, Northeastern S tates Commer cialSavings BanksCooperative Banks Stock Mutual Banks* Total Number Existing 12/31/89 588 149 211 101 1,049 Number of Failures, 1990–94 65 32 16 5 118 Percent Failed 11 21 8 5 11 Note: Data are for Connecticut, Maine, Massachusetts, New Hampshire, New Jersey , New York, Rhode Island, and Vermont. * “Cooperative banks” is the term used for state-chartered savings and loan associations in Massachusetts.mutuals that existed as of the end of 1989 and had not converted, and 1 1 percent of the re - gion’ s commercial banks (see table 1. 7). New banks and converted mutuals highlighted in extreme fashion the problems confronting many other banks in the 1980s. These institu - tions had strong incentives to expand loan portfolios rapidly in order to leverage high ini - tial capital positions, increase earnings per share, and meet stockholder expectations.38In so doing, these institutions rapidly increased their lending in markets already experiencingvigorous competition and deteriorating credit standards. They combined powerful compet - itive pressures to assume greater risk with relative inexperience in a demanding new envi - ronment. Newly chartered banks began operations at a time when inexperience was adistinct liability , while many converted mutuals responded to internal and external pres - sures by entering unfamiliar markets or geographic areas. As a result, a disproportionate number of new and converted banks failed. 38Managers of savings banks that converted may also have been willing to take greater risks with their personal compensa - tion than managers of banks that retained the mutual form.Fraud and F inancial Misconduct The precise role of fraud and financial misconduct as a cause of bank failures is dif fi- cult to assess. The consensus of a number of studies is that fraud and financial misconduct (1) were present in a lar ge proportion of bank and thrift failures in the 1980–94 period, (2) contributed significantly to some of these failures, and (3) were able to take root because ofthe same managerial deficiencies and inadequate internal controls that contributed to the fi - nancial problems of many failed and problem institutions (apparently internal weaknessesleft some institutions vulnerable not only to adverse economic developments but also toAn Examination of the Banking Crises of the 1980s and Early 1990s Volume I 34 Histor y of the Eighties—Lessons for the F utureabuse and fraud). The studies also agree that the dollar impact of such activity is extremely difficult to estimate. A1988 OCC study of 162 national bank failures between 1979 and 1987 concluded that insider abuse was a significant contributing factor in 35 percent of the failures, andfraud in 1 1 percent. 39As for problem banks that recovered and survived, the OCC found that 24 percent of these banks had suf fered from significant insider abuse, while none had significant problems with fraud. Another study , which drew on a number of analyses and re - ports prepared by Congress and the regulators, concluded that fraud and insider abuse con - tributed to between 33 and 50 percent of commercial bank failures and from 25 to 75percent of thrift failures in 1980–88. 40A1993 U.S. General Accounting Of fice (GAO) re - port pointed to the dif ficulties of quantifying the ef fects of fraud and to the wide variations in estimates of its impact.41Whereas the OCC study found that fraud played a significant role in 1 1 percent of national bank failures, the FDIC found that fraud and insider abuse were present in 25 percent of 1989 bank failures; and the Resolution Trust Corporation (RTC) reported in 1992 that potential criminal abuses by insiders contributed to 33 percent of RTC failed thrift cases. Finally , a 1994 GAO report indicated that FDIC investigators had found insider fraud to be a major cause of failure in 26 percent of a sample of 286 banks thatfailed in 1990–91 and insider “problems” (fraud, noncriminal abuses, and loan losses on in - sider loans) to be present in 61 percent. 42 Anumber of factors make it dif ficult to measure the ef fect of fraud and abuse. First, some cases of fraud go undetected. Second, sometimes the line between poor businessjudgment and fraud is dif ficult to draw , as is the line between criminal and noncriminal ac - tivities. Third, the regulators and the Federal Bureau of Investigation do not maintain com - plete or consistent records on fraud convictions, reported incidents of fraud, and financialmisconduct. Fourth, new legislation had ef fects that make comparisons over time dif ficult to draw: FIRREA and the Crime Control Act of 1990 increased the resources for detecting and reporting fraud and broadened the agencies’ powers to deal with bank and thrift fraud. For all of these reasons, any attempt at precision would be unwarranted. However , it seems reasonable to infer that fraud and abuse not only were present in a lar ge number of bank and thrift failures in the 1980–94 period but also contributed to some of them. 39Office of the Comptroller of the Currency , Bank Failur e: An Evaluation of the Factors Contributing to the Failur e of Na - tional Banks (1988). 40Benton E. Gup, Bank Fraud: Exposing the Hidden Thr eat to Financial Institutions (1990). 41U.S. General Accounting Of fice, Bank and Thrift Criminal Fraud: The Federal Commitment Could Be Br oadened (GAO/GGD-93-48, January 1993). 42U.S. General Accounting Of fice, Bank Insider Activities: Insider Pr oblems and V iolations Indicate Br oader Management Deficiencies (GAO/GGD-94-88, March 1994).Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 35Factors Associated with Bank F ailur es: Conclusion The preceding discussion points to a variety of factors—economic, financial, legisla - tive, regulatory , supervisory , managerial—that contributed to bank failures during the 1980s. Not all observers subscribe to a multiple-cause interpretation of bank-failure historyor to the particular set of multiple causes described in this study . Some place particular em - phasis on one or two specific causes that they believe were especially influential. For ex - ample, bank regulators tend to place heavy weight on deficiencies in bank management. 43 Bankers tend to blame government policy and adverse changes in the economy . Journalists point to cases of malfeasance. Academic writers have placed special emphasis on the fi - nancial incentives facing bank owners and managers. With respect to these last, a considerable body of academic literature has stressed the effect that flat-rate deposit insurance (whose cost is unrelated to the level of risk assumed by individual institutions) had in encouraging m oral-hazard risk taking and leading to depository-institution failures.44There seems little question that excessive risk taking by then-solvent banks contributed to bank failures and that flat-rate deposit insurance con - tributed to risk taking. However , singling out deposit insurance pricing as the principal ex - planation of bank failures seems unwarranted. Deposit insurance was available at fixed ratesthroughout m ost of the FDIC’ s history , but before the 1980s bank failures were few in num - ber and were often caused by fraud rather than by financial risk taking. It was changes in themarketplace (increased competition, downward pressure on profits, lifting of legal re - straints, and so forth) that created the environment in which increased risk taking (includingexploitation of flat-rate deposit insurance) became advantageous or necessary for manybanks. Furtherm ore, as mentioned above, although banks that failed had generally assumed greater risk before their failure, many other banks with similar risk profiles did not fail. Inthe case of these surviving banks, the ef fects of risk taking, including risk taking stimulated by underpriced deposit insurance, were apparently of fset by other factors, including supe - rior risk-management skills. The absence of these of fsetting factors should therefore be considered m ore important causes of bank failures. Moral-hazard risks appear to have had greater significance in the savings and loan industry than in the banking industry; this wasmainly because thrift regulators permitted (or were forced by a depleted insurance fund topermit) a lar ge number of thrifts to operate for lengthy periods with little or no equity , a sit - uation that produced extraordinary incentives for risk taking. 43See OCC, Bank Failur e,5, 10: “The study showed that deficiencies within boards of directors and management were the primary internal problems of problem and failed banks . . . The evidence from healthy and rehabilitated banks also supports our hypothesis that economic conditions are rarely the primary factor in determining a bank’ s condition.” See also Richard Duwe and James Harvey , “Problem Banks: Their Characteristics and Possible Causes of Deterioration,” Federal Reserve Bank of Kansas City Banking S tudies (1988): 3–1 1. 44See discussion below of m oral hazard (“Role of Deposit Insurance”).An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 36 Histor y of the Eighties—Lessons for the F utureThe academic literature has also produced a second, alternative explanation of the in - centives facing solvent banks, focusing on issues related to the control of banks exercisedby owners and managers. 45According to this view , managers rather than owners make lend - ing decisions. If managers are entrenched (imperfectly controlled by owners), they maymake decisions that are at odds with the interests of stockholders. According to this view , in periods (such as the 1980s) when the lending opportunities for banks were reduced as a re - sult of the loss of market share in financing lar ge businesses, some managers sought to pre - serve their perquisites by shifting lending to risky loans—a shift that led to loan losses andreductions in capital. Focusing on the sometimes dif ferent incentives of managers and own - ers is useful for understanding variations in the behavior of dif ferent institutions. However , it is not clear that such dif ferences played a leading role in the increased number of bank failures. Many managers may have believed that maintaining their reputations and futureemployment prospects would best be served by risk-averse policies that avoided the failureof their institution. Furtherm ore, some “entrenched” managers of solvent institutions (for example, managers of savings banks that retained the mutual form) seem to have operatedtheir institutions relatively conservatively in the late 1980s. 46 Athird view of the role of incentives in explaining risk taking by banks draws an anal - ogy between federal deposit insurance and a trilateral performance bond in which the in - surance agency provides a bond that protects depositors against poor performance by thebank. 47This view emphasizes incentive conflicts between various parties: for example, bank owners and managers, stakeholders in insured institutions and managers of the in - surance agency , insurance agency managers and elected government of ficials, elected government of ficials and taxpayers. In this setting, regulators lack the incentives to enforce effective loss-control measures (capital requirements, m onitoring, etc.) that are opposed by the regulated industry or “threaten a regulator ’s ability to mask poor performance.”48 45Gary Gorton and Richard Rosen, “Corporate Control, Portfolio Choice, and the Decline of Banking,” Journal of Finance (December 1995): 1377–410. Gorton and Rosen conclude that issues of corporate control are m ore important than m oral hazard in determining the behavior of solvent institutions. In the case of insolvent institutions, managers and owners haveidentical interests and behave in the manner suggested by the m oral-hazard principle. Another study , based on experience in the 1990s, concluded that the relationship between corporate structure and risk is significant only at low-franchise-valuebanks where m oral hazard problems are m ost severe and conflicts between owner and manager risk preferences are there - fore the strongest. See Rebecca S. Demsetz, Marc R. Saidenber g, and Philip E. S trahan, “Agency Problems and Risk Tak- ing at Banks,” S taff Report, Federal Reserve Bank of New York, September 1997. 46As noted before, mutual savings banks that converted to the stock form failed with greater frequency in the late 1980s and early 1990s than mutuals that retained the mutual form. Mutual savings banks had no stockholders and were governed byself-perpetuating boards of trustees or directors, which in some cases were dominated by their chief executive of ficers; managers of such institutions might reasonably be considered entrenched in the sense of being imperfectly controlled. 47Edward J. Kane, “Three Paradigms for the Role of Capitalization Requirements in Insured Financial Institutions,” Journal of Banking and Finance 19 (1995): 431–59. 48Ibid., 447. Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 37These academic views share an emphasis on the sometimes conflicting incentives of bank owners, managers, regulators, and others as the principal explanation of insuf ficiently restrained bank risk taking. They also share the view that bank risk is essentially endoge - nous, arising from factors internal to the banking and regulatory systems, including mis - priced deposit insurance, inadequate owner -control of bank managers, or m ore general principal-agent problems am ong various parties involved in or af fected by deposit insur - ance and bank regulation. The importance of exogenous factors (the economy , financial markets, etc.) is correspondingly diminished in explaining bank risk taking and failures. Ultimately , the role of financial incentives in bank failures is inseparable from the role of broader economic, financial, legislative, and regulatory factors; the extent to which flat-rate deposit insurance pricing, for example, led to excessive risk taking and widespread fail - ures apparently depended on the circumstances. The multiple-cause explanation appears to be a m ore plausible reading of the history of the 1980s. According to this view , the rise in the number of bank failures was caused by a variety of factors internal and external to theindustry . 49This is not to say that failures were due merely to “bad luck,” with everything going wrong at the same time. More realistically , the preconditions for a rise in the number of bank failures were present well before the 1980s. These preconditions included, am ong others, a structure of banking laws that inhibited competition, geographic diversification ofrisks, and consolidation of units. They also included managerial attitudes and regulatory provisions that reflected the relatively benign pre-1980 environment for banking when fail - ures were rare, and a system of flat-rate deposit insurance premiums that was tenable whenother incentives and opportunities for risk taking were weak. The localized nature of many banks and a lack of experience with hard times left them vulnerable to external shocks andregional and sectoral recessions. Under the pressure of increased competition, many banksassumed greater risks, and as long as they remained solvent and profitable they were insuf - ficiently restrained by the supervisory authorities. When the economic, financial-market, and competitive environment turned markedly less favorable for banks and some govern - ment policy actions (principally ill-timed deregulation and tax changes) exacerbated the sit - uation, the preconditions were translated into increased numbers of bank failures. Which banks failed and which banks survived in an increasingly demanding environment waslargely determined by an individual bank’ s circumstances, particularly variations in the lev - els of risk it assumed, its success (or lack thereof) in operating with high risk levels, the 49Astudy by the FDIC Of fice of Inspector General (OIG) of the 13 bank failures in 1994, when conditions for banking were much dif ferent from in the 1980s, concluded that in a majority of cases problems were evident in loan underwriting, credit concentrations, high overhead, imprudent management, and external economic factors. Less comm on or critical factors were financial derivatives, volatile deposits, cross-guarantee assessments, and newly chartered banks. See FDIC Of fice of Inspector General, 1994 Failed Banks T rend Analysis (1995), 2. Similar results were found for 6 failures in the OIG’ s 1995 Failed Banks T rend Analysis (1996).An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 38 Histor y of the Eighties—Lessons for the F utureoverall strength of its management, good or bad fortune, and (in some cases) the presence or absence of fraud and misconduct. Regulator y and Super visor y Issues Raised by the Experience of the 1980s The principal regulatory and supervisory issues arising from the experience of the 1980s include the role of deposit insurance, the treatment of lar ge-bank failures, the use of forbearance, the impact of Prompt Corrective Action, and the ef fectiveness of supervisory tools—examination, enforcement, and of f-site surveillance. Role of Deposit Insurance Deposit insurance has often been described as involving a trade-of f between stability and m oral hazard.50On the one hand, by protecting depositors against loss, deposit insur - ance virtually eliminates the risk of bank runs and disruptive breakdowns in bank lending.On the other hand, by assuming the risk of losses that would otherwise be borne by depos - itors, deposit insurance eliminates any incentive for insured depositors to m onitor bank risk and permits bank managements to take increased risks. Because of deposit insurance, banksare able to raise funds for risky projects at costs that are not commensurate with the risk ofthe projects, a situation that may lead to the misallocation of resources and to failures. 51 Moral hazard is a particularly serious concern if the bank is insolvent or close to insolvency , in which case the owners have strong incentives to make risky investments because profitsaccrue to the owners, whereas losses fall on the insurer . (On the other hand, risk taking may be restricted if the bank has suf ficient franchise value, defined as the present value of future income expected to be earned by the bank as a going concern.) In principle, the insuringagency can protect itself by requiring deductibles (equity positions) so that owners havetheir own funds at risk and by char ging premiums commensurate with the risk assumed by the various banks. However , because it is dif ficult to identify indicators that give accurate advance warning of future distress, m oral-hazard problems are inherent in deposit insur - ance, as in other types of insurance. 52Deposit insurance suf fers from the additional prob - 50Arthur J. Murton, “Bank Intermediation, Bank Runs, and Deposit Insurance,” FDIC Banking Review 2, no. 1 (1989): 1–10. The term “m oral hazard” has been defined as “a description of the incentive created by insurance that induces those insured to undertake greater risk than if they were uninsured because the negative consequences are passed through to the insurer”(Congressional Budget Of fice, Reforming Federal Deposit Insurance [September 1990], 163). 51In principle, owners of mar ginally solvent nonbank firms may also have incentives to take greater risk, but they are gener - ally constrained by uninsured creditors. 52The unreliability of ex ante risk measures has been attributed to information asymmetries between the insured and the in - surer , whereby the former is seen to be better informed about his or her risky behavior .Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 39lem that it insures against losses that are not independent but are interrelated through the ef - fects of cyclical economic activity and the possibility of contagious bank runs. During the 1980s, the balance in this trade-of f was generally tipped in favor of stabil - ity. In this respect, regulatory policy was eminently successful; despite an unprecedented number of bank and thrift failures, there was no evidence of serious runs or credit-flowdisruption at federally insured institutions. S tability was achieved, it should be noted, at substantial cost to surviving institutions and to their customers (assuming the institu-tions passed on at least part of the burden of increased assessments). In the case of thrift-institution failures, some of the costs were borne by taxpayers as well. The estimated total cost of FDIC failed-bank resolutions in 1980–94 is $36.3 billion. The estimated cost of the savings and loan debacle is $160.1 billion, of which an estimated $132.1 billion was borneby taxpayers. 53 In contrast, the record of regulators with respect to controlling risk taking was mixed—and in the case of still-profitable and solvent banks, often unfavorable. Here a dis - tinction must be made between controlling the risky behavior of profitable, solvent banksand controlling risk taking by problem banks that already face the near -term prospect of in - solvency and failure. The record of the 1980s seems clear on this point. The regulators were reasonably successful in m odifying the behavior of of ficially designated problem banks so as to reduce the prospects of their failure or the cost to the insurance fund if failure oc - curred. The regulators were less successful in constraining risk taking by still-profitable and healthy banks, partly because there were no reliable, generally accepted, forward-lookingmeasures of risk. There are three traditional means of controlling m oral hazard: (1) examination and su - pervision; (2) regulatory capital requirements and risk-based deposit insurance premiums;and (3) uninsured depositor and creditor discipline. 54In varying degrees and at various times, all three of these means were operating imperfectly in the 1980s. As discussed below , examination of many banks was infrequent in the early and middle 1980s, with the resultthat the consequences of risky behavior and other problems were not always identified on a 53The savings and loan cost figure includes the costs of the FSLIC and the R TC plus tax benefits under FSLIC assistance agreements, but excludes potential costs from supervisory goodwill claims. See U.S. General Accounting Of fice, Resolu - tion T rust Corporation’ s 1995 and 1994 Financial S tatements (July 1996), 13. 54See Murton, “Bank Intermediation.” Astudy of Texas banks concluded that “the propensity to engage in risky activities de - pends on m ore than just changes in capital. Abank’ s current risk influences the response of bank lending to changes in cap - ital. As long as banks possessed the ability to expand their lending, lower growth rates of capital were associated with lar ger increases in lending, as m oral hazard would suggest. However , once banks were m ore exposed to risk, those institutions with lower capital growth recorded statistically insignificant dif ferences in lending compared to those banks with greater increases in capital. While this latter finding is inconsistent with m oral hazard, it points out the potential importance of both regulatory and liquidity constraints at work” (Jef fery W. Gunther and Kenneth J. Robinson, “Moral Hazard and Texas Banking in the 1980s,” Federal Reserve Bank of Dallas Financial Industr y Studies [December 1990]: 6).An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 40 Histor y of the Eighties—Lessons for the F uturetimely basis. Although for some time regulators had been using capital standards to assess the condition of banks, uniform minimum capital requirements covering all banks were notadopted until 1985, and risk-based capital requirements not until 1990. Most bank failureswere resolved through purchase-and-assumption transactions or open-bank assistanceagreements that protected uninsured depositors and nondeposit creditors and therefore fos - tered the belief that all deposits of lar ge banks were 100 percent insured. This belief se - verely limited the discipline that depositors might otherwise have exerted on the behaviorof banks. More specifically , supervisory restraints did not prevent the speculative binge of com - mercial real estate and other risky lending by solvent banks in many regions of the countryin the 1980s. Regulators apparently believed that as long as risky behavior was profitable,they had limited leverage to restrain such behavior . Examiners interviewed for this study stated that as long as the banks were profitable, it was dif ficult to persuade bank manage - ments or their own superiors in the regulatory agencies that problems could lie ahead. When risky behavior resulted in actual losses, regulators were m ore ef fective, but often by that time the damage had been done. Part of the problem was the absence of explicit penalties or costs to make risky be - havior less attractive—penalties and costs such as risk-based premiums and capital re - quirements that, as stated, were not adopted until late in the period. Earlier adoption ofuniform capital requirements and risk-based premiums would have improved the positionof the bank regulators but might still have been insuf ficient to curb the excessive risk tak - ing in the 1980s. As noted, capital regulation is a principal means of restraining risky be - havior , but equity-to-assets positions are lagging indicators of a bank’ s risk profile and therefore poor indicators of the risk of failure several years before the fact. Current risk-based capital standards, which dif ferentiate am ong broad asset categories, permit consider - able shifting toward riskier lending within categories without requiring additional capital,while higher risk-based premiums are char ged to banks whose condition has already dete - riorated. 55In short, regulators’ ability to restrain the risky behavior of currently profitable banks was limited by the absence of penalties or costs based on reliable and generally ac - cepted early-warning signals.56 55The shift in bank lending from business loans to commercial m ortgages during the 1980s would not have required in - creased capital under present risk-based capital standards. Risk-based premiums vary according to capital positions and su - pervisory ratings. 56Some would ar gue that problems of controlling risky behavior would be solved by the adoption of market value account - ing. This ar gument assumes that market participants, utilizing publicly available data, would be better able than regulators to correctly recognize advance warning signs of risk, even though regulators have access to information developed throughon-site examinations. This assumption remains unproven. See section below on “T reatment of Big Banks: Systemic Risk and Market Discipline.”Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 41The problems faced by regulators in controlling the risky behavior of profitable banks as compared with troubled banks illustrate dif ferences between ex ante and ex post mea - sures of risk. Comm on measures of ex ante risk (for example, loans-to-assets and other as - set-composition ratios) measure risk taking independent of the current condition of thebank. They tend to be limited in their reliability—for example, many banks with high-risk profiles were able to avoid failure in the 1980s. Thus, regulators may be reluctant to apply stringent restraints and penalties on the basis of ex ante risk measures. On the other hand,ex post measures of risk (for example, capital-to-assets ratios) are the m ost proximate mea - sures of risk to the insurance fund and measure the consequences of risk taking after it hasmaterially weakened the condition of the bank. Supervisory restraints and penalties can bemore confidently applied on the basis of ex post risk measures, but they may be less ef fec- tive than those based on reliable ex ante measures in curbing risk taking before it weakensthe condition of the bank. Moreover , the weakened condition of banks identified on the ba - sis of ex post risk limits the magnitude of penalties that can actually be applied. 57 Whereas bank regulators may have lacked the tools to restrain solvent banks from ex - cessive risk taking, thrift regulators were in a far dif ferent position. The Federal Home Loan Bank Board was not confronted by the problem of limiting risk taking by healthy institu - tions but by a lar ge number of savings and loan associations that were insolvent or barely solvent in the early 1980s. The course that thrift regulators followed may in retrospect be termed high risk, featuring reduced capital standards, liberalized ownership restrictions forstockholder -owned thrifts, and capital and accounting forbearance that allowed savings and loan associations to operate with minimal or no equity while their true condition was ob - scured. 58This course was followed partly because the financial resources of the FSLIC fund were inadequate. It was apparently m otivated by the belief (or hope) that thrifts could grow out of their problems by acquiring new assets, that external capital could be attracted toshore up the industry , and that thrift institutions should be permitted to operate with mini - mal capital until they were able to improve earnings by using new asset powers. In contrastto banks, in the first half of the 1980s undercapitalized thrifts were allowed and even en - couraged to grow . 59Apart from dif ferences in regulatory philosophy , FHLBB policies re - flected the depleted state of the FSLIC insurance fund. The closure of all thrifts as they reached or approached insolvency was not a viable option. One obvious conclusion fromthe experience of the 1980s is that an adequate insurance fund is a prerequisite for anyattempt to control m oral hazard. 57For discussions of this topic, see two FDIC studies: Deposit Insurance for the Nineties: Meeting the Challenge (1989) and AStudy of the Desirability and Feasibility of a Risk-Based Deposit Insurance System (1990). 58See Chapter 4, “The Savings and Loan Crisis and Its Relationship to Banking.” 59See “Use of Forbearance” below .An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 42 Histor y of the Eighties—Lessons for the F utureTreatment of L arge Banks Regulators’ preference for solutions that prom oted stability rather than market disci - pline is apparent in the treatment of lar ge banks (mutual savings banks, m oney-center banks, and Continental Illinois). At various times and for various reasons, regulators gener - ally concluded that good public policy required that big banks in trouble be shielded fromthe full impact of market forces and that their uninsured depositors be protected. This pol - icy contributed to the overall record of stability achieved by the deposit insurance system inthe 1980s. At the same time, however , it weakened any incentive for uninsured depositors to monitor and restrain risk taking by the banks. The first big bank to fail in the 1980s was First Pennsylvania Bank, N.A., of Philadelphia, with $8 billion in assets in early 1980. 60In this case the FDIC provided open-bank assistance, and the agency’ s determination of the bank’ s “essentiality” was based mainly on First Pennsylvania’ s size as the city’ s largest bank and on the possibility that its failure would have local and national repercussions. Large mutual savings banks. The issue of systemic risk was raised m ore explicitly by the threatened insolvency of mutual savings banks. Located mainly in New York and other northeastern states, these institutions suf fered a severe earnings squeeze because of the rapid rise in interest rates in the late 1970s and early 1980s, pushing interest costs on short-term deposits above interest rates on the institutions’ long-term, fixed-rate m ortgage loans and bond holdings. Earnings were also held down by usury ceilings applicable to residen - tial m ortgage loans in New York. Although asset quality was not generally a problem at this time, the net worth shortfall at market values was so lar ge, according to one estimate, that if the banks had failed, the liability facing the FDIC would have exceeded the size of the in - surance fund. 61 The first savings bank to fail was the Greenwich Savings Bank with $2.5 billion in as - sets—at the time, the third-lar gest bank failure in the FDIC’ s history . The initial estimated cost of the Greenwich failure was m ore than the recorded total cost of all previous failures of insured banks. Federal action was precipitated by the bank’ s inability to roll over foreign borrowings. Among the FDIC’ s first acts was to announce that no depositors, insured or uninsured, would lose any principal or interest, a m ove designed to preserve confidence in other savings banks that were also suf fering severe interest-rate pressures. The bank was re - solved through an FDIC-assisted mer ger transaction with another savings bank, a transac - tion assisted through an Income Maintenance Agreement, and this became the prototype for 60FDIC, The First Fifty Y ears: AHistor y of the FDIC, 1933–1983 (1984), 95. 61See Chapter 6, “The Mutual Savings Bank Crisis.”Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 43other savings bank transactions. In all, 17 mutual savings banks with $24 billion in assets were resolved through assisted mer gers during 1981–85.62 Money-center banks with LDC (less-developed-countr y) loans. The case of m oney- center banks with lar ge concentrations of loans to developing countries also illustrates the regulators’ preference for stability (as well as other public policy objectives) over market discipline.63Between year -end 1978 and year -end 1982, total LDC debt held by the eight largest m oney-center banks expanded from $36 billion to $55 billion. Total LDC portfolios held by these banks averaged m ore than double the banks’ aggregate capital and reserves at the end of 1982, a ratio that put some of the lar gest banks at risk. Bank regulators made some attempt to curtail LDC lending activity and ensure diversification of foreign lendingrisk, doing this partly through the Interagency Country Exposure Review Committee, com - posed of of ficials of the OCC, the FDIC, and the Federal Reserve. These ef forts apparently had little ef fect on the growth of LDC loans. Conversely , LDC lending may have been en - couraged by the OCC’ s 1979 interpretation of the loans-to-one-borrower rule, an interpre - tation according to which public sector borrowers that met certain conditions did not haveto be counted as parts of a single entity . On balance, it may be said that government policy supported LDC lending activity by the banks. In August 1982, the government of Mexico announced it could no longer meet inter - est payments, and by the end of the year 40 nations were in arrears. By the end of 1983, 27countries were in negotiations to restructure their existing loans. Following the Mexican de - fault, U.S. banking of ficials did not require that lar ge reserves be immediately set aside for the restructured LDC loans, apparently believing that some lar ge banks might have been deemed insolvent and that an economic and political crisis might have been precipitated. 64 Although loss reserves did increase, at the end of 1986 they still averaged only approxi - mately 13 percent of the total LDC exposure of the m oney-center banks. S tarting in 1987, however , the m oney-center banks began to recognize massive losses on LDC loans that in some instances had been carried on the banks’ books at par for m ore than a decade. By the 62The assisted mer ger transaction was chosen over a purchase and assumption or a deposit payof f so that the FDIC could avoid the immediate outlays necessary to of fset the full am ount of asset depreciation and because these institutions had no stockholders to benefit from the transactions (and, in m ost cases, few uninsured depositors to share the cost with the FDIC). Most of the transactions were accomplished before the Net Worth Certificate Program was adopted as part of the Garn–S t Germain Act (see the section below on forbearance). 63See Chapter 5, “The LDC Debt Crisis.” 64L. William Seidman, Full Faith and Cr edit: The Gr eat S&L Debacle and Other W ashington Sagas (1993), 127. According to former FDIC Chairman Seidman, “U.S. bank regulators, given the choice between creating panic in the banking systemor going easy on requiring our banks to set aside reserves for Latin American debt, had chosen the latter course. It would appear that the regulators made the right choice.”An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 44 Histor y of the Eighties—Lessons for the F utureend of 1989, total reserves at the m oney-center banks were nearly 50 percent of total LDC loans. The LDC experience illustrates the high priority given to maintaining financial mar - ket stability in the treatment of lar ge banks. It also represents a case of regulatory forbear - ance. The OCC’ s 1979 interpretation of the loans-to-one-borrower rule permitted banks to continue lending in the face of signs that Latin American nations were having increasing difficulty meeting their obligations. Regulatory forbearance also enabled m oney-center banks to delay recognizing the losses and thereby avoid repercussions that might havethreatened their solvency . In time, loss reserves and char ge-of fs were greatly increased, and no m oney-center bank failed because of LDC loans. 65The creation of the Brady Plan in 1989 reflected recognition that banks would not recover the full principal value of existingloans and turned international ef forts from debt rescheduling to debt relief. As part of the process, substantial funds were raised from the International Monetary Fund and the World Bank to facilitate debt reduction. Ultimately , the shareholders of the world’ s largest banks assumed the losses under the Brady Plan, which brought the crisis to an end. Continental Illinois. The failure of Continental Illinois—a bank with $45 billion in as - sets in 1981 and one of the ten lar gest in the nation—was the lar ge-bank transaction that set the terms for the ensuing “too-big-to-fail” debate. 66The $4.5 billion rescue package devised by the regulators in May 1984 was prompted by a high-speed electronic bank run that fol - lowed a period of deteriorating performance. Problems in Continental’ s loan portfolio had been highlighted in July 1982, when Penn Square Bank failed; Continental Illinois had hada heavy concentration of loan participations with Penn Square. The rescue package in - cluded the promise to protect uninsured depositors fully , and it brought to an end the FDIC’ s modified payof f program, in which only a portion of the am ount owed to uninsured depositors was paid; that portion was based on the estimated recovery value of the failed in - stitution’ s assets. The reversal in FDIC policy reflected concerns that other lar ge banks might be subject to bank runs and that Continental’ s correspondent banks would suf fer losses if the FDIC resolved the bank through a deposit payof f or otherwise failed to protect uninsured deposits. The justification for the Continental Illinois transaction has been debated at length. For example, a 1993 article criticizing the transaction and its rationale concluded that in 65One analysis concluded that “had these institutions been required to mark their sometimes substantial holdings of under - water debt to market or to increase loan-loss reserves to levels close to the expected losses on this debt (as measured bysecondary market prices), then institutions such as Manufacturers Hanover , Bank of America, and perhaps Citicorp would have been insolvent” (Robert A. Eisenbeis and Paul M. Horvitz, “The Role of Forbearance and Its Costs in Handling Trou- bled and Failed Depository Institutions,” in Reforming Financial Institutions in the United S tates, ed. Geor ge G. Kaufman [1993], 60). 66See Chapter 7, “Continental Illinois and ‘T oo Big to Fail.’ ”Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 45most cases losses on deposits held by correspondent banks at Continental would have been relatively small and that these banks probably would have been able to meet any liquiditystrains through the Federal Reserve’ s discount window . 67As for the possibility that prob - lems at Continental Illinois might have caused contagious runs on otherwise viable banks,the essential question is whether the market would have been able to distinguish betweenviable and nonviable banks (so that it would be able to end quickly any run on the former).Uncertainties on this point have made decisions on the resolution of lar ge-bank failures dif - ficult and will continue to make them dif ficult in the future. (See “Open Questions” below .) ** * These transactions in the early 1980s involving mutual savings banks, m oney-center banks, and Continental Illinois generally set the pattern for the treatment of lar ge banks throughout the rest of the decade. In lar ge-bank resolutions in the Southwest and Northeast as well as in other regions, the FDIC used purchase-and-assumption transactions, bridgebanks, and open-bank assistance agreements that provided full protection for uninsured de - positors. These methods eliminated the need for uninsured depositors to m onitor the per - formance of lar ge banks and raised questions of fairness, since numerous small-bank failures were resolved through deposit payof fs, in which uninsured depositors suf fered losses. 68 The treatment of some lar ge-bank failures has also been criticized on the ground that regulators were not assertive or prompt enough in curbing the risky behavior that led to thefailures. It is clear that some years before its failure in May 1984, Continental Illinois hadembarked on a rapid-growth strategy built on decentralized loan management that was un - constrained by an adequate system of internal controls and was heavily reliant on volatilefunds. It is also clear that supervisory restraints were insuf ficient to m odify the bank’ s be- havior . AHouse subcommittee report in 1985 criticized a lack of “decisive action” on the part of the OCC and also found fault with the Federal Reserve’ s supervision of the parent holding company . Some of the regulators who participated in the Continental Illinois trans - action have indicated that while the bank was profitable, regulators were reluctant to takeearly action in opposition to the bank’ s board of directors. 67Larry D. Wall, “T oo-Big-to-Fail after FDICIA,” Federal Reserve Bank of Atlanta Economic Review (January/February 1993): 1–14. 68It is likely that even without the too-big-to-fail policy , large banks would have been resolved less frequently through de - posit payof fs because they tended to have greater franchise value and marketability . The greater marketability of lar ge banks may have been due to their greater flexibility in seeking new markets and of fering new product lines, their location in states where the absence of restrictions on geographic expansion meant a greater number of qualified bidders, and theearlier resolution action (to the extent that disclosure requirements applicable to publicly traded companies alerted regula - tors to problems at an earlier stage).An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 46 Histor y of the Eighties—Lessons for the F utureCriticism has also been leveled against the supervisory treatment of the Bank of New England in the years before its failure in January 1991.69According to the General Ac- counting Of fice, problems in the bank’ s operations were identified through the examination process several years before its failure. The firm grew rapidly from 1985 to 1989, primarily through acquisitions and aggressive real estate lending. During this high-growth period,OCC examiners repeatedly identified and reported problems with the bank’ s controls over lending operations and strategies. However , not until 1989 were any enforcement actions taken against the bank to compel corrective measures. The GAO concluded that the OCC relied on management’ s assurances that it would address the problems; it also concluded that m ore vigilant supervision could have reduced losses. Use of F orbearance Forbearance has taken on such pejorative connotations that various uses of the term need to be distinguished. 70At one extreme, forbearance may be said to occur when super - visory authorities permit an insured depository institution to operate without meeting es - tablished safety-and-soundness standards for a limited period of time while taking remedialactions to reduce risk exposure and correct other weaknesses. Forbearance in this sense hasoften been applied by bank regulators on a case-by-case basis. As an example, problem banks that face near -term insolvency and closure frequently attempt, under pressure from regulators, to acquire additional capital. The success or failure of such ef forts often deter - mines whether the bank survives or is closed. Decisions as to whether , and for how long, to allow these ef forts to continue are in fact decisions as to whether , and for how long, for - bearance of this limited type should be granted. Whether regulators make the correct deci - sions in these situations cannot be tested with any precision. However , such limited, case-by-case forbearance seems to be an integral part of the overall supervision of problembanks, and its usefulness is best judged by the degree of success of such supervisory ef - forts. 71 At the other extreme is the type of forbearance practiced by the FSLIC, as a result of which a lar ge number of insolvent or mar ginally solvent savings and loan associations were permitted to operate as open institutions for lengthy periods.72The dif ference between the extremes is m ore than a dif ference of degree. Limited, case-by-case forbearance is designed to provide an opportunity to reduce risk exposure and correct weaknesses. Longer -term, 69See Chapter 10, “Banking Problems in the Northeast.” 70Bank forbearance programs are discussed by Dean Forrester Cobos, “Forbearance: Practices and Proposed S tandards,” FDIC Banking Review 2, no. 1 (1989): 20–28. 71See “Ef fectiveness of Supervisory Tools: Examination and Enforcement” below . 72In 1984, 687 FSLIC-insured thrifts with $358 billion in assets, constituting 22 percent of the number of thrifts and 37 per - cent of total industry assets, were insolvent on the basis of tangible net worth. See Lawrence J. White, The S&L Debacle: Public Policy Lessons for Bank and Thrift Regulation (1991), 1 14. Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 47wholesale forbearance as practiced by the FSLIC was a high-risk regulatory policy whose main chances of success were that the economic environment for thrifts would improve be - fore their condition deteriorated beyond repair or that the new , riskier investment powers they had been granted would pay of f. The latter type of forbearance, which the FSLIC adopted against the background of a depleted insurance fund, is widely judged to have in - creased the cost of thrift failures.73Because of the state of the FSLIC fund, forbearance be - came a necessity for the thrift regulators rather than a matter of choice74and continued to be widely granted after interest-rate reductions in the early and middle 1980s had alleviatedmaturity mismatches in thrift portfolios, and poor -quality assets had become the chief prob - lem of S&Ls. Generally , the bank regulators did not practice such wholesale, protracted, and risky forbearance. The bank regulators did, however , allow several lar ge banks that subsequently failed to operate for long periods with minimal capital (see “Impact of Prompt Corrective Action” below). As noted above, bank regulators also eased the problems of m oney-center banks with lar ge holdings of LDC loans by not requiring prompt establishment of reserves against such loans. This was a form of temporary forbearance; eventually m oney-center banks sub - stantially increased their reserves. 75Finally , bank regulators administered three forbearance programs that were applied to classes of banks rather than to individual institutions (seetable 1. 8). These programs were initiated or inspired by Congress rather than by the bank regulators. The first such program was the Net Worth Certificate Program for thrifts that was adopted, despite FDIC reservations, as part of the Garn–S t Germain Act. 76This program was applied mainly to FDIC-insured mutual savings banks in New York and other north - eastern states that were suf fering extreme earnings pressures in a period of high and rising 73See, for example, Edward J. Kane, The S&L Insurance Mess: How Did It Happen? (1989); Eisenbeis and Horvitz, “For - bearance and Its Costs,” 49–68; Edward J. Kane and Min-T eh Yu, “Opportunity Cost of Capital Forbearance during the Fi - nal Years of the FSLIC Mess,” Quar terly Review of Economics and Finance 36, no. 3 (fall 1996): 271–90; and Ram on P. DeGennaro and James B. Thompson, “Capital Forbearance and Thrifts: An Ex Post Examination of Regulatory Gam - bling,” in Proceedings of the 29th Confer ence on Bank S tructur e and Competition , Federal Reserve Bank of Chicago, May 1993, 406–20. However , one analysis concluded that “[F]orbearance was not a major culprit in the taxpayer bill for the thrift crisis.” See Geor ge J. Benston and Mike Carhill, “FSLIC Forbearance and the Thrift Debacle” in Credit Markets in Transition, Pr oceedings of the 28th Annual Confer ence on Bank S tructur e and Competition, Federal Reserve Bank of Chicago, 1992: 131. 74One analysis concluded that the FSLIC’ s ability to dispose of insolvent thrifts was constrained by S&L industry pressures, by the extent of past cover -ups of thrift insolvencies, and by the actions of elected of ficials (Kane, The S&L Mess, 97, 98). 75According to some authors, the case for forbearance rests on the existence of market imperfections (such as legal impedi - ments to diversification), deadweight bankruptcy costs, inef ficient markets for bank assets, information asymmetries whereby assets have greater value when managed by the banks that originated them than when managed by FDIC liquida - tors, and macroeconomic considerations (Eisenbeis and Horvitz, “Forbearance and Its Costs,” 52, 64, 65). 76FDIC, The First Fifty Y ears, 102.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 48 Histor y of the Eighties—Lessons for the F utureTable 1.8 Results of Bank Forbearance Pr ograms Mutual Savings Banks, Agricultural and Net Worth Cer tificates Energy Sector Banks Number of banks in program 29 301 Assets of banks in program ($billions) 40 13 Number of banks that survived 22 236 Number of banks that failed 7 65 Losses as per cent of assets at failur e Banks in forbearance program 4 21* Comparable banks not in program 12 22* * Data refer to banks with less than $100 million in assets. interest rates. Between 1982 and 1986, 29 mutual savings banks with approximately $40 billion in assets participated. Of these, 22 banks were restored to profitability as falling in - terest rates in the early and middle 1980s enabled these institutions to improve equity posi - tions and retire their net worth certificates. Seven savings banks that participated in theprogram failed as a result of interest-rate pressures and were resolved at a cost of $420 mil - lion, or approximately 4 percent of total assets at the time they entered the program. 77This loss rate was substantially less than the average loss rate of 12 percent for savings banks re - solved before the Net Worth Certificate Program was adopted.78 The ef fectiveness of the Net Worth Certificate Program was due lar gely to the drop in interest rates after 1981. In ef fect, Congress required that action against insolvent savings banks be deferred until after interest rates had come down, by which time, it was thought,profitability and equity positions would be restored, and in fact in m ost cases they were. 79 Also important was the fact that the FDIC was generally able to contain m oral-hazard risks associated with the continued operation of banks having little or no equity . Most of the sav - 77See Chapter 6, “The Mutual Savings Bank Crisis.” Two of the 22 savings banks failed subsequently , four to six years after having retired their net worth certificates. These failures were probably the result of actions taken after the two banks left the program. 78The lower loss rate of banks that failed while in the Net Worth Certificate Program was probably due in part to the fact that after the program was introduced, interest rates were generally declining. In addition, the first savings banks to fail mighthave been in a m ore serious condition than those that failed later . 79By comparison, many insolvent savings and loan associations did not recover as a result of the drop in interest rates. At the end of 1982, there were 222 GAAP-insolvent FSLIC-insured thrifts. In September 1986, despite a nearly 500 basis-pointdrop in 90-day Treasury bill rates from 1982 to 1986, only 29 percent of these institutions were now GAAP-solvent, whereas 36 percent were still GAAP-insolvent and 35 percent had ceased to exist. See U.S. General Accounting Of fice, Thrift Industr y: Forbearance for T roubled Institutions , 1982–1986 (GAO/GGD-87-78BR, May 1987), appendix 1.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 49ings banks were free of serious credit-quality problems (as mutual institutions, they might have had less incentive than stockholder -owned institutions to make risky investments), and the relatively small number of savings banks in the program simplified supervision andfacilitated control of risky behavior . The second instance of class-of-bank forbearance was the 1986 temporary capital for - bearance program for banks that were weakened as a result of lending to the troubled agri - cultural and ener gy sectors; this program was later extended to all banks that were experiencing dif ficulties because of economic factors beyond their control. Bank regulators developed the program at a time when support for forbearance was building in Congress.By developing their own program, bank regulators sought to include a strong safety-and-soundness focus and to avoid being required to use measures like the Net Worth Certificate Program or those the thrift regulators employed. 80Of the 301 banks in the capital forbear - ance program, 201 were operating as independent institutions one year after leaving theprogram, another 35 had been mer ged without FDIC assistance, while 65 had failed. As these results indicate, after a period of forbearance a lar ge majority of the institutions in the program either were able to recover as independent institutions or had suf ficient value to be acquired by mer ger partners without FDIC assistance. Losses of the 65 banks that failed were similar to those of comparable failed banks, a fact suggesting that the period of for - bearance did not result in serious deterioration. Of the 65 failed banks in the program, 59were under $100 million in assets and had losses of 21 percent of assets. In comparison, 965banks with assets less than $100 million that were not in the forbearance program and failedduring 1986–94 had a 22 percent loss rate. As in the case of the Net Worth Certificate Pro - gram, the ef fectiveness of the 1986 regulators’ program was lar gely due to its temporary na - ture and to cyclical economic forces, in this case, a recovery in the agricultural sector . Athird instance of class-of-bank forbearance was the Agricultural Loan-Loss Amor- tization Program adopted by Congress in 1987 as part of CEBA, apparently because Con - gress concluded that the regulators’ program was inadequate. Of 33 banks in the program, 27 survived as independent institutions one year after leaving it, another 2 had mer ged, while 4 had failed. 81Essentially the same conclusions apply to this program as to the 1986 regulators’ agricultural and ener gy forbearance program. In assessing the ef fectiveness of class-of-bank forbearance programs, one needs to consider how banks are chosen to participate when the regulators are allowed to exercisediscretion. Ideally , the regulators must be able to distinguish between institutions that will recover after a period of forbearance and those that will not recover and should therefore 80See Chapter 2, “Banking Legislation and Regulation”; and Cobos, “Forbearance: Practices and Proposed S tandards,” 23. 81Data exclude banks that were in both the CEBA and the 1986 regulators’ programs. These banks are included only in data for the latter program.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 50 Histor y of the Eighties—Lessons for the F uturenot be granted forbearance. The ability to make such distinctions accurately is important for reasons of fairness and because of m oral hazard. In making such distinctions, the regulators have the benefit of information derived from examination reports—information that is notavailable in financial reports or other public records. Nevertheless, picking winners andlosers is dif ficult, and some writers have concluded that regulators were unsuccessful in their attempts. 82 Furtherm ore, applying forbearance to a group of banks may have adverse competitive effects on institutions outside the program. Unless restrained by the supervisory authorities, insolvent banks may of fer above-market deposit rates and submarket loan rates, thereby weakening healthy competitors. Such behavior by many thrift institutions during the 1980sgenerated frequent complaints, but it was apparently less of a problem in the bank forbear - ance programs because a smaller number of institutions were involved and the participantswere closely m onitored and supervised. In other words, while forbearance may provide an opportunity to correct weaknesses, without ef fective oversight it may also permit further deterioration. As noted below (“The Impact of Prompt Corrective Action”), allowing un - profitable banks to continue operating can increase resolution costs as operating losses ac - cumulate. Even if it is successfully applied to some banks, forbearance may haveundesirable ef fects if it encourages other banks to expect similar treatment. Moreover , if forbearance is granted to a lar ge number of institutions, it may have adverse ef fects on the economy . 83 Thus, forbearance programs may have a number of disadvantages—and, when prac - ticed on the scale and with the purposes of the FSLIC program, they can be a disaster . While survival of the institution is not the only criterion for the success of forbearance programs,it remains significant that m ost of the banks in class-of-bank forbearance programs sur - vived, 84and the minority that failed had losses comparable to, or lower than, those of failed banks not included in the programs. The m ore favorable results of bank forbearance pro - grams as compared with the FSLIC strategy reflect the smaller number of banks involved,the closer m onitoring of banks, the fact that the problems addressed by bank forbearance programs were temporary and cyclical in nature, 85and (m ost important) the fact that bank 82See, for example, Emile J. Brinkmann, Paul M. Horvitz, and Ying-Lin Huang, “Forbearance: An Empirical Analysis,” Journal of Financial Ser vices Resear ch(1996): 39–40. 83See the discussion in Congressional Budget Of fice, The Economic Effects of the Savings and Loan Crisis (1992). 84One analysis states that “the cost to taxpayers of FDIC gambling lies in of fering the equivalent of dividend-free equity cap - ital to undercapitalized banks. The success of these gambles must not be measured by whether assisted banks recovered, but by whether societal returns on taxpayer funding proved high enough to justify the waiver of dividends.” See Kane,“Three Paradigms,” 444. 85One view of S&L forbearance programs is that as a result of deregulation, these institutions were under going a permanent change that could not be addressed by an essentially temporary measure. See Congressional Budget Of fice, Reforming Federal Deposit Insurance , xiv. Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 51regulators sought to control risk taking by participating institutions rather than encourage it.86In the absence of the class-of-bank forbearance programs, m ore of the banks that actu - ally survived might have been closed: for example, as shown in the next section, if the pro - visions of Prompt Corrective Action had been in ef fect throughout the 1980s, 12 of the 22 mutual savings banks that participated in the Net Worth Certificate Program and recovered would have faced the prospect of closure, while 50 of the 236 surviving farm and ener gy banks in the regulators’ 1986 temporary program might also have been closed. Impact of P rompt Cor rective A ction The Prompt Corrective Action (PCA) provisions of FDICIA were designed to limit regulatory forbearance by requiring m ore-timely and less-discretionary intervention, with the objective of reducing failure costs. FDICIA mandated that the regulatory authorities adopt five capitalization categories, ranging from “well capitalized” to “critically under - capitalized,” to serve as the basis for Prompt Corrective Action. As an institution’ s capital position declines, the appropriate regulator is required to increase the severity of its actions.These actions range from restricting asset growth (for undercapitalized institutions) to clos - ing banks (those that are critically undercapitalized for a prescribed period). The top four capital categories are defined in terms of risk-based capital and leverage ratios. Criticallyundercapitalized institutions are those with tangible capital ratios of 2 percent or less. Ingeneral, a receiver must be appointed for any institution that is critically undercapitalizedfor up to 270 days. 87 It is dif ficult to judge what would have happened if PCA had been in ef fect during the 1980s, for the behavior of both banks and bank regulators would have been altered. How - ever, it appears that some banks that failed might have been closed earlier than they actu - ally were, whereas some banks that survived might have faced the prospect of being 86The dif ference between the FDIC forbearance program for mutual savings banks and the FSLIC program for savings and loan associations has been described as follows: “[A]ccounting gimmicks were limited—and the mutual savings bankswere not allowed to grow . With a conservative policy of temporary forbearance in place, many mutual savings banks re - covered, and those ultimately shut down or mer ged did not put an intolerable burden on the FDIC . . . the S&Ls that fol - lowed the incentives and implicit advice of government policy to enter new areas rapidly and grow their way out of theproblems became part of the S&L debacle” (National Commission on Financial Institution Reform, Recovery and En - forcement, Origins and Causes, 32–33). 87Under FDICIA, when an institution is critically undercapitalized for 90 days a receiver or conservator must be appointed or some other action must be taken to achieve the purpose of the provision. The 90-day delay may be extended, provided that the regulator and the FDIC concur and document why extension would better serve the purposes of the provision. Af- ter the institution has been critically undercapitalized for 270 days, a receiver or conservator must be appointed unless theregulator and the FDIC certify that the institution is viable and not expected to fail. Under the conditions existing in the1980s when failures were bunched and the market for failed institutions was often saturated, it seems reasonable to sup - pose that taking m ore than 90 days to spread out marketing ef forts for failed banks would have been an acceptable reason for delay up to the 270-day limit.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 52 Histor y of the Eighties—Lessons for the F utureunnecessarily or erroneously closed. Alternatively , banks in the latter group might have been compelled to try either to recapitalize earlier than they actually did or to mer ge with healthier banks. Alarge majority of banks that failed were closed within the time frame specified by FDICIA for critically undercapitalized banks. However , 343 banks that failed (21 percent of all failures from 1980 to 1992) with $88 billion in assets would have facedearlier closure because they were critically undercapitalized for m ore than 270 days. 88For the same reason, 143 problem banks (those with CAMEL ratings of 4 or 5) with $1 1 billion in assets that did not fail would have faced the possibility of unnecessary closure becauseof the 2 percent rule. Of the 343 failed banks that would have been closed earlier under the PCA rule, 201 (59 percent) were national banks, 131 (38 percent) were state nonmember banks, and 1 1 (3 percent) were state member banks. In the case of national banks, closure is the responsibil - ity of the OCC; in the case of state-chartered institutions, of state banking departments. Inthe states that had the m ost closings and the m ost late closings, the state authorities closed problem banks m ore quickly than the OCC did. 89The dif ference was especially apparent in Texas and Oklahoma, which accounted for a disproportionate number of bank failures. Part of the dif ference was due to the fact that state banking authorities had greater flexibility un - der applicable law . The OCC had statutory authority to close a national bank “whenever the Comptroller shall become satisfied of the insolvency of the bank” (12 U.S.C. 191). Thus, the OCC had to wait until the bank was insolvent before being able to close it. On the otherhand, the six states had the authority to close banks when capital was “impaired,” when thebank faced “imminent insolvency” or was in an “unsafe” or “unsound” condition. These more flexible standards made it possible for the states to close banks earlier . 90However , al- though the OCC’ s closing policy was constrained by the statutory-insolvency rule, the agency had wide latitude to define insolvency and presumably could have adopted a m ore flexible standard than was actually in ef fect during m ost of the 1980s. Until December 1989, the OCC’ s definition of insolvency was the exhaustion of primary capital (equity plus loan-loss reserves). In December 1989, after approximately a year of study , the OCC shifted to equity capital alone, without loss reserves, and the new definition permitted m ore expe - ditious closing of national banks.91This change was made after m ost of the failures of the 1980s had already been resolved. 88Excluded from this analysis are banks that participated in forbearance programs mandated or inspired by Congress. 89In six states (California, Colorado, Louisiana, New York, Oklahoma, and Texas) the OCC closed 473 banks during the 1980–92 period, 38 percent of which were closed later than would have been required under PCA. The state authorities closed 459 banks, 17 percent of which were closed later than would have been required under PCA. 90Information on the statutory authority of the six state banking departments is based on conversations with representativesof each of the six departments. 91OCC, Bulletin BB-89-39 (December 13, 1989).Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 53Estimates of the cost savings that would have resulted from the earlier closure of failed institutions are necessarily very rough.92For m ost of the 343 banks that would have faced earlier closure if PCA had been in ef fect, the interval between the date that closure would have been required by PCA and the actual closure date was approximately two quar - ters.93 During this interval, these banks experienced a reduction in equity from $220 million to a negative $1.6 billion. But a lar ge part of this reduction was due to the recognition of losses that were already embedded in loan portfolios and would not have been af fected by more-timely closure. Another portion—chiefly operating losses associated with higher pri - vate-sector funding costs and the cost of operating retail bank branch systems—could havebeen avoided by earlier closure. This avoidable cost for the 343 banks is estimated to be on the order of $825 million and constituted 8 percent of the actual estimated resolution costsof the 343 banks and approximately 2 percent of the cost of all bank failures during the1980–92 period. 94Approximately 60 percent of the $825 million estimated cost savings is attributable to six lar ge banks. An alternative estimate of the avoidable cost, based on net operating losses, produced essentially the same aggregate result. For the 343 banks, net operating losses before loan-loss provisions, gains/losses on asset transactions, taxes, and extraordinary items totaled$815 million for the intervals between closure dates required by PCA and actual closure dates. As with the previous estimate, these losses were concentrated in a few lar ge banks. Anumber of caveats are in order when one considers these estimates. Regulators’ bank closure policies would have been dif ferent if PCA had been in ef fect in the 1980s, and such policy changes might have reduced projected cost savings. For example, for the lar ge number of banks that were allowed to operate with tangible capital below the 2 percentlevel for only a few m onths beyond the interval allowed by PCA, earlier closure might have 92The calculations are described in note 94. R. Alton Gilbert concluded, contrary to the implications of this study , that FDIC resolution costs were not positively related to the length of time that banks operated with relatively low capital ratios be - fore their failure. See “The Ef fects of Legislating Prompt Corrective Action on the Bank Insurance Fund,” Federal Reserve Bank of S t. Louis Review 74, no. 4 (July/August 1992): 3–22. 93The unweighted average interval was two quarters. Weighted by assets, the average interval was three quarters, reflecting the especially long intervals for a few lar ge banks. 94The avoidable cost is estimated as the sum of (1) the actual funding costs of these banks minus the one-year Treasury rate and (2) the operating expenses of transactions and nontransactions deposit accounts as estimated by the 1990 FunctionalCost Analysis of the Federal Reserve Board. The avoidable cost was computed for the period of time beyond 270 days that the bank’ s tangible capital ratio was below 2 percent. In cases where the tangible capital ratio fluctuated below and above 2 percent, the bank was considered to be critically undercapitalized for the entire period after the ratio first fell below 2 p er- cent, except when the ratio subsequently rose above 3 percent. In the latter case, that bank was counted as critically under - capitalized only for the period it was below 2 percent subsequent to having reached the 3 percent level. Two lar ge savings banks that had entered into Income Maintenance Agreements with the FDIC in connection with the acquisition of other failed institutions were counted as critically undercapitalized from the time the bank’ s agreement was terminated (in one case) and (in the other case) from the date the FDIC formally permitted the bank to miss capital tar gets prescribed in its agreement.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 54 Histor y of the Eighties—Lessons for the F uturemeant that, because of insuf ficient time to market the institutions am ong potential acquir - ers, m ore institutions would have been resolved through insured-deposit payof fs.95This likelihood would have been greatest in periods when failures were bunched, temporarilysaturating the market for failed bank and thrift deposit franchises and assets. Spreadingclosings over a longer period of time might have attracted better bids and of fset some of the additional costs resulting from delayed closings. Thus for many of the 343 banks, the cost savings resulting from earlier implementation of PCA might have been smaller than the es - timates set forth above suggest. For the six lar ge banks that operated for extended periods of time with minimal capital, earlier closure would probably have achieved cost savings.For some of these banks, fairly lengthy marketing periods might have been needed and, be - cause of PCA, regulators might have had to start the marketing process while the banks hadcapital above the 2 percent tangible level. In any event, whatever savings might have beenachieved through earlier closure would apparently have been concentrated lar gely in a few large banks that were permitted to operate with little or no equity for relatively long periods of time. During the interval between the actual and the PCA-required closure dates, problem institutions were generally under close supervision and many of them were subjects of en - forcement actions aimed at reducing losses to the insurance fund. Of the 343 failed banksthat would have faced earlier closure under PCA, 127 were FDIC-supervised state non - member banks for which enforcement data are available. Of the 127 banks, 101 (approxi - mately 80 percent) had been issued formal enforcement actions before the closure daterequired by PCA—in fact, an average of 14 m onths before—and the remaining 26 banks might have had informal enforcement actions. 96 The consequences of unnecessarily closing some of the 143 problem banks that were below the 2 percent level but did not fail must be weighed against the cost savings of clos - ing failed banks earlier . As noted above, some of these banks would have recapitalized or would have mer ged sooner to avoid closure.97However , any unnecessary or erroneous clo - sure of these institutions would be dif ficult to justify and might have involved unnecessary 95This possibility was pointed out by R. Alton Gilbert. See his comments in volume 2 of this study . 96Data on formal enforcement actions (such as cease-and-desist orders and terminations of insurance) are presented here only for FDIC-supervised state nonmember banks. Comparable data are not available for OCC-supervised banks; relatively fewbanks were supervised by the Federal Reserve. Systematic data on informal enforcement actions are unavailable for theFDIC and the OCC. 97The cost of unnecessary or erroneous closure of banks that would otherwise have survived is likely to be lar ge if bank - ruptcy costs are high and if investors undervalue the assets of the banks. As noted by S tanley C. Silverber g in volume 2 of this study , “Early resolution works very well when the market places reasonable or high valuations on bank franchises. However , in, say , 1990, the stock prices of several of the m ost conservatively run banks were well below book value. In - vestors and other banks were reluctant to pay positive prices for troubled banks without FDIC assistance. That has changed considerably during the past several years.”Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 55deadweight bankruptcy costs. In any future period of widespread failures, balancing the benefits of earlier closure against the consequences of closing some banks that otherwisewould have survived may be dif ficult. Presumably banks will strive to avoid becoming sub - ject to the 2 percent rule, or to any other similarly binding rule, by maintaining capital lev - els higher than they otherwise would or by seeking mer ger partners while they still have value. However , the history of the 1980s shows that capital levels may decline quickly in the face of external shocks or other unforeseen events. Or at times the market may tem - porarily undervalue a bank franchise, making it dif ficult for some banks to secure external capital when they are in danger of failing the 2 percent rule. Thus, in some future period of widespread depository-institution failures, the issue of erroneously closing salvageable in - stitutions may be unavoidable and critical in implementing statutory closure rules. The computations that produced the estimates that 343 failing banks would have been closed earlier and 143 banks might have been unnecessarily closed as a result of the appli - cation of PCA in the 1980s did not include banks in the class-of-bank forbearance pro - grams, because the assumptions underlying these programs were obviously at variance withthe later views of Congress as expressed by the PCA provision of FDICIA. However , for the sake of completeness, separate calculations using the same methodology were made forthe banks that participated in these forbearance programs. The results show that (1) 48 banks with $1 1 billion in assets that actually failed would have been closed earlier as a re - sult of PCA, and (2) 66 banks with $16 billion in assets that actually survived would havebeen closed. In addition to the closure of critically undercapitalized banks, FDICIA requires spe - cific regulatory intervention geared to capital positions of open banks. For example, in thecase of undercapitalized banks, FDICIA requires regulators to have the bank submit a cap - ital restoration plan, restrict asset growth, and get prior approval for expansion. For signif - icantly undercapitalized banks, m ore-stringent actions are prescribed. In this regard, a study of the New England banking crisis, which occurred before the adoption of FDICIA in 1991, found that regulators were already imposing formal actions on banks before they becameundercapitalized as defined by PCA. Moreover , according to the study , the regulators im - posed restrictions m ore comprehensive than those prescribed in the PCA legislation. 98The reason given for this result is that capital ratios prescribed in PCA are lagging indicators of the health of the institution and will trigger enforcement action well after problems are iden - tified in examinations. Examiners analyze considerably m ore information than capital ra - tios to determine a bank’ s likelihood of failure. Therefore, m ore-timely intervention would 98Joe Peek and Eric S. Rosengren, “The Use of Capital Ratios to Trigger Intervention in Problem Banks: Too Little, Too Late,” Federal Reserve Bank of Boston New England Economic Review (September/October 1996). See also Peek and Rosengren, “W ill Legislated Early Intervention Prevent the Next Banking Crisis?” working paper series no. 96-5, Federal Reserve Bank of Boston, September 1996.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 56 Histor y of the Eighties—Lessons for the F utureresult from triggers that mimic the timing of problem-bank identification by examiners. This view of the lagging nature of capital ratios is consistent with the findings summarizedbelow in the section on of f-site m onitoring. Effectiveness of Super visor y Tools: Examination and Enforcement The increased number of bank failures in the 1980s raised questions about the ef fec- tiveness of bank regulators’ systems of identifying problem banks and then influencing their behavior in order to prevent failures and reduce insurance losses. The evidence sug - gests that bank examination ratings provided a reasonably accurate indication of theprospect of failure if the ratings were based on recent examinations. But in the early and middle 1980s many banks were not examined frequently , and the ratings available for them at any point tended to be obsolete. Troubled banks that were properly identified, however , were generally subject to enforcement actions that appear to have been ef fective in reduc - ing insurance losses. The critical issues, therefore, are the frequency and use of examina - tions, the ef fectiveness and limitations of CAMEL ratings, and the ef fectiveness of follow-up enforcement actions. 99 Evolution in the fr equency and use of examinations. In the late 1970s and early 1980s, the bank examination process was af fected by two key policy changes embraced particu - larly by the OCC and the FDIC: (1) relatively m ore reliance was placed on of f-site m oni- toring and relatively less on on-site examination, and (2) examination resources wereconcentrated on those institutions that posed the greatest threat to the insurance fund and tothe stability of the financial system. These changes were made partly because it was be - lieved that comprehensive Call Report data and the use of computer technology would en - hance of f-site surveillance and enable the agencies to reduce the examination burdens on banks and on their own staf fs. Further , the decision to concentrate resources on the lar ger and the m ore-troubled banks was seen as an ef ficient allocation of resources. (Both the FDIC and the Federal Reserve also made increasing use of state bank examinations for non - problem institutions.) Another important change took place at the OCC, where the tradi - tional emphasis on a detailed audit and verification system was replaced by a focus on thequality of management and internal policies. The OCC also placed increased weight on tar - geted examinations, which focused on a particular aspect of a bank’ s operations, rather than full-scope examinations. These policy changes implied that fewer examiners would be needed. In addition, both the Carter and Reagan administrations restricted federal hiring in an attempt to reducethe size of the federal government. In this climate, the FDIC and the OCC froze examinerstaffing levels in 1981. As a result, between 1979 and 1984 the total number of examiners 99This section is based on Chapter 12, “Bank Examination and Enforcement.”Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 57Table 1.9 Number of Bank Examiners, Federal and S tate Banking Agencies, 1979–1994 Year FDIC FRS OCC States Total 1979 1,713 805 2,151 2,496 7,165 1984* 1,389 820 1,722 2,201 6,132 1990 2,645 1,025 1,907 2,470 8,047 1994 2,547 1,529 2,376 2,564 9,016 Source: Compiled by FDIC on the basis of information from FRS, OCC, and Conference of S tate Bank Supervisors. *Trough in total number of examiners.in federal and state banking agencies declined by 14 percent (see table 1. 9). Among the agencies, the reductions varied in size: examiner staf fing at the FDIC declined by 19 per - cent, at the OCC by 20 percent, and at state agencies by 12 percent. At the Federal Reserve, examiner staf fing was lar gely unchanged. While examination forces were being reduce d, the total number of troubled banks was increasing from 217 in 1980 to 1,140 in 1985. In themid-1980s, therefore, the FDIC and the OCC began to rebuild examiner staf fs—but several years of training are required to produce qualified examiners, so it was not until the late1980s that the examiner forces at those two agencies were restored to 1980 levels in num - ber and experience. 100 100The demands on the shrunken examiner staf fs extended to training new hires and taking on duties related to settlement and asset liquidation for failed banks.These trends in examiner staf fing contributed to marked changes in the number and frequency of examinations. Between 1981 and the low point of 1985, the number of exam - inations declined from approximately 12,300 to 8,300. The decline was particularly sharp for state nonmember banks; for national banks and state member banks it was less severe.In 1979, the average length of time between examinations was 379 days, or 13 m onths (see table 1. 10). By 1986, the average interval had increased to 609 days, or 20 m onths. The greatest change was for CAMEL 1-rated banks, whose average interval increased from 392 days to 845, or from 13 to 28 m onths. The increase in examination intervals was greatest at the OCC and the FDIC and smallest at the Federal Reserve. As the agencies built up their examination staf fs in the late 1980s, intervals between examinations shortened once again, and by 1990, the average interval was 41 1 days (14 m onths) for all banks; for all banks with CAMEL ratings below 2, it was one year or less. In 1991 FDICIA reinforced the return to greater frequency of examinations by requiring annual full-scope examinations for allAn Examination of the Banking Crises of the 1980s and Early 1990s Volume I 58 Histor y of the Eighties—Lessons for the F utureTable 1.10 Mean Examination Interval, by Initial Composite CAMEL Rating (in days) Year 1 2 3 4 5 All Banks 1979 392 396 338 285 257 379 1986* 845 656 407 363 313 609 1990 463 436 331 303 270 411 1994 380 357 296 279 245 354 Source: FDIC, FRS, and OCC. *Peak of mean intervals. banks, except that for small banks with satisfactory ratings an 18-m onth interval could be substituted.101 For some banks during the mid-1980s, these changes meant that CAMEL ratings and other information derived from examinations were sometimes obsolete and unrepresenta - tive. CAMEL ratings are a measure of the condition of a bank essentially at the time it is ex - amined; as a bank’ s condition changes, old ratings become increasingly inaccurate as indicators of its current health.102Problems developing at some banks in the 1980s were not identified on a timely basis; this view is supported by examiners interviewed for this study , who indicated that extended examination intervals and increased demands on staf f re- sources meant that some banks received insuf ficient attention. For example, banks that were well rated but deteriorating might not receive attention until it was too late to preventserious losses. In Texas, which had the lar gest concentration of bank failures and losses to the insurance fund, the problem of extended examination intervals was particularly acute.The severe problems of some Texas banks might have been recognized sooner if examina - tions had been m ore frequent. 103 The reduced frequency of examinations limited the usefulness not only of information derived from the examinations but also of the financial reports used in of f-site m onitoring. On-site examiners are able to evaluate the quality of the loan portfolio and verify the data 101John O’Keefe and Drew Dahl, “The Scheduling and Reliability of Bank Examinations: The Ef fects of FDICIA” (unpub - lished paper presented at the Financial Management Association conference, October 1995). 102Rebel A. Cole and Jef fery W. Gunther , “A CAMEL Rating’ s Shelf Life,” Federal Reserve Bank of Dallas Financial In - dustr y Studies (December 1995). Cole and Gunther concluded that the information content of CAMEL ratings decays rapidly; examination ratings indicate bank survivability m ore accurately than of f-site m onitoring does for two quarters af - ter examinations; for periods longer than two quarters, examinations are less accurate than of f-site m onitoring. 103O’Keefe, “The Texas Banking Crisis.”Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 591 and 2 3 4 5 0 10 20 30 4036% 25% 8%26% 28% 10%CAMEL Rating As a Per centage of Failing Banks31% 36%Compo site CAMEL Rating sTwoYears befor e Failur eforBanks Failing between 19 80and1994Figur e 1.9 AllAvailable Ratings Ratings Less Than One YearOld* Ratings thatwere lessthanoneyearoldasofthetwo-years-before-failure date; that is,ratings based onexaminations dated between two andthree years before failure.*on nonperforming loans and loan char ge-of fs that banks report in Call Reports.104In other words, on-site examinations are needed to ensure the accuracy of bank financial reports. Ifexaminations are less frequent, the accuracy of of f-site m onitoring systems using Call Re - port data suf fers. Effectiveness of CAMEL ratings. When examination ratings were up-to-date, they generally identified m ost of the banks that required increased supervisory attention well be - fore the banks actually failed. As shown in figure 1. 9, of the m ore than 1,600 banks that failed in 1980–94, 36 percent had CAMEL 1 and 2 ratings two years before failure; 25 per - cent had ratings of 3, 31 percent had ratings of 4, and 8 percent had ratings of 5. But these 104R. Alton Gilbert, “Implications of Annual Examinations for the Bank Insurance Fund,” Federal Reserve Bank of S t. Louis Review 75, no. 1 (January/February 1993); and Drew Dahl, Gerald A. Hanweck, and John O’Keefe, “The Influence of Au- ditors and Examiners on Accounting Discretion in the Banking Industry ,” unpublished paper presented at Academy of Financial Services conference (October 1995).Table 1.1 1 Failing Banks with CAMEL Ratings of 1 or 2 Two Years befor e Failur e, 1980–1994 Number Percent of Total Failur es Total 1- and 2-rated futur e failur es 565 35% Specific types: Cross-guarantee cases 25 Failures associated with fraud 24 First City Bancorporation af filiates 36 First RepublicBank Corporation af filiates 26 CAMEL ratings m ore than one year old* 194 Total of above 305 19 Remaining 1- and 2-rated futur e failur es 260 16 * Failures of banks with ratings m ore than one year old (two years before failure) do not include cross-guarantee cases, fail - ures associated with fraud, First City Bancorporation af filiates, or First RepublicBank Corporation af filiates.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 60 Histor y of the Eighties—Lessons for the F uturedata refer to examination ratings available two years before failure, whereas some of the examinations had actually been conducted considerably m ore than two years before failure. Also included in these data are banks that failed for types of reasons that cannot be antici - pated well in advance by safety-and-soundness examinations: cross-guarantee failures; fail - ures due to fraud; and failures of af filiates of certain Texas holding companies that were essentially operating as branches of the parent institution, were tracked outside the CAMELsystem, and were resolved through procedures that had much the same ef fect as cross-guar - antees. 105If we exclude examinations for these banks as well as examinations that are m ore than one year old,106the percentage of failed banks that had CAMEL 1 and 2 ratings two years before failure drops to 16 percent of the total number of failures (see table 1. 11).107In other words, the proportion of failed banks that were not identified as requiring increasedscrutiny two years before their failure was 16 percent. 108 105In the case of First RepublicBank Corporation, the FDIC’ s demand that af filiate banks honor their pledge to back the agency’ s assistance to the lead bank caused the af filiates to fail. In the case of First City Bancorporation, the FDIC provided assistance to the holding company and required that it be downstreamed to the af filiates. One may ar gue that examiners should consider what the condition of the lead bank implies for the condition of af filiated banks in the holding company . However , examiners could not have known two years in advance the nature of the resolution arrangements that would be adopted in these two cases and in post-FIRREA cross-guarantee cases and their ef fects on other banks in the company . 106Exclusion of banks with ratings that were m ore than one year old two years before failure means, in ef fect, that the data refer to examinations conducted between two and three years before failure. 107Banks with CAMEL 1 and 2 ratings are treated here as a separate category from banks with worse ratings. CAMEL 1- and 2-rated banks are defined as “basically sound in every respect” or “fundamentally sound, but may reflect m odest weak - nesses correctable in the normal course of business.” Banks with a CAMEL 3 rating “give cause for supervisory concern and require m ore than normal supervision,” while CAMEL 4 and 5 ratings are reserved for progressively weaker banks. 108For banks with assets of m ore than $250 million, the proportion was 15 percent. This suggests that the ef fectiveness of CAMEL ratings in anticipating failures was about the same for lar ge and small banks.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 61Over the course of the 1980–94 period, the record of CAMEL ratings in anticipating failures improved as the frequency of examinations increased and problems were appar - ently better identified. From the period 1980–86 to the period 1987–94, the proportion offailed banks that had CAMEL 1 and 2 ratings two years before failure declined from 28 to 12 percent. Similarly , the proportion of failed banks that had CAMEL 4 and 5 ratings two years before failure rose from 25 to 46 percent. 109 Limitations of examination ratings. Although CAMEL ratings were reasonably suc - cessful in identifying banks that required greater supervisory attention, they also had limi - tations. First, they did not necessarily capture the severity of the situation of the banks thatsubsequently failed. Second, they are based on the internal operations of the bank and there - fore do not take into account local economic developments that may pose future problemsand are not yet reflected in the bank’ s condition. Third, as noted above, they are generally a measure of the condition of the bank at the time it is examined. They do not systematically track risk factors that may produce future losses. 110Fourth, frequent use of on-site exami - nations imposes a burden on depository institutions. Examinations may seem particularlyburdensome during good economic times, when the condition of m ost banks is healthy and examination ratings change relatively little. An average of 85 percent of all banks examined each year during the 1980–94 period experienced either no change or an improvement inratings; only 15 percent, on average, experienced ratings downgrades. However , examina - tion ratings changed considerably m ore often in particular regions of the country and dur - ing periods of regional recessions. Most banks that are designated as troubled banks (rated CAMEL 4 and 5) do not fail. This may be regarded as a deficiency of CAMEL ratings. On the other hand, examination ratings trigger the supervisory responses that may prevent troubled banks from failing ormay reduce failure costs when the banks have to be closed. From this perspective, when su - pervisory ef forts to cure bank problems as revealed by examinations have been successful, the failure forecasts based on these examinations will necessarily prove to have been inac - curate. Either way , the lar ge number of troubled banks that do not fail and the lar ge number of banks whose ratings do not change through repeated examinations are unavoidable con - sequences of frequent use of on-site examinations. However , on-site examinations provide 109Data are the numbers of failed banks that had the indicated CAMEL ratings two years before failure in each year , weighted by the total number of failures in that year . The data are based on 260 banks after exclusion of examinations m ore than one year old, failures due to fraud, cross-guarantees, and the subsidiaries of two Texas bank holding companies (table 1.1 1). 110Apossible exception is the management rating, which encompasses technical competence, leadership qualities, adequacy of internal controls, and other factors that may determine the bank’ s ability to weather future adversity . However , exam - iners appear to be reluctant to rate management much below capital, asset quality , and other CAMEL components. In this regard, in only 6 percent of failed banks were the management ratings of two years before failure one full number worsethan the average of other components.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 62 Histor y of the Eighties—Lessons for the F utureinformation to the regulators that is otherwise unavailable,111and they also help ensure the accuracy of financial reports issued by the banks.112As a result, the burdens of frequent ex - aminations must be borne if the condition of insured banks is to be m onitored ef fectively . Recognizing these burdens, the FDIC has sought to reduce the time examiners spend inbanks and is developing a program designed to allow individual loan files to be examinedoff-site. Number , kinds, and effectiveness of enforcement actions. After troubled institutions were identified during the 1980–94 period, they were subject to supervisory and enforce - ment actions that appear to have been ef fective in reducing failures and losses to the insur - ance fund. This conclusion is based on evidence concerning the behavior of banks with respect to asset growth rates, dividend payouts, and equity infusions when the banks hadbeen designated as problem institutions and been made subject to informal and formal en - forcement actions. 113 The FDIC used formal enforcement actions (for example, cease-and-desist orders) sparingly in the 1970s but m ore frequently in the early 1980s, as the number of troubled banks increased. Formal enforcement actions are legally enforceable in court, and noncom - pliance with such actions may lead to heavy fines. Most FDIC formal enforcement actionsin the 1980s were issued against 4-rated banks, which are troubled but salvageable; m ost of the remainder were issued against 5-rated banks, which face a high probability of imminentor near -term failure. About one-half of all banks rated 4 and 5 by the FDIC in the 1980s were the subject of formal enforcement actions; many of the remaining banks received in - formal enforcement actions (for example, mem oranda of understanding). 114Enforcement actions require banks to take corrective actions in various areas: compliance with regula - tions, improvement in operating procedures, the raising of new capital, the cutting of divi - dend payments, replacement of managers, and so forth. That supervisory and enforcement actions were ef fective in reducing failures and losses to the insurance fund is suggested by the following: •Of all banks that were rated 4 and 5 sometime during the 1980–94 period, 73 percent recovered, while 27 percent failed. As noted above, one-half of the FDIC-supervised problem 111The view that examinations yield unique information is lar gely based on the belief that banks specialize in evaluating and monitoring idiosyncratic borrowers who do not have practical access to the capital markets. This view suggests that the best way to secure the private information banks have gathered about borrowers is by examining individual loan files. 112See Drew Dahl, Gerald A. Hanweck, and John O’Keefe, “The Influence of Auditors and Examiners on Accounting Dis - cretion in the Banking Industry ,” and Gilbert, “Implications of Annual Examinations.” 113Data on enforcement actions are available for FDIC- and Federal Reserve–supervised banks only . 114In a sample of 307 FDIC-supervised problem banks there were 209 with formal actions, 83 with informal actions only , and merely 15 with neither formal nor informal actions.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 63Table 1.12 Asset Gr owth Rates, Dividend Payments, and Capital Injections, All Banks with CAMEL Ratings of 4 and 5, 1980–1994 Total Banks Failed Banks Surviving Banks (Failed and Surviving) Years befor e Years Years of Recovery* Years of Failure, Recovery ,* Failur e, Recover y, of Failure or Mer ger or Mer ger orMerger 1980–85 1986–91 1992–94 1980–85 1986–91 1992–94 1980–85 1986–91 1992–94 Asset Gr owth Rate (Per cent) 3 14.60 15.65 18.77 10.39 13.38 4.42 11.91 14.09 5.93 2 10.72 1.71 −3.53 3.67 1.25 −0.61 6.21 1.40 −0.92 1 0.91 −10.17 −13.39 1.96 0.96 −0.64 1.58 −2.51 −1.98 Dividends to Average Assets (Per cent) 3 0.34 0.42 0.09 0.20 0.21 0.13 0.25 0.21 0.13 2 0.32 0.52 0.06 0.16 0.14 0.09 0.22 0.15 0.09 1 0.16 0.39 0.02 0.13 0.13 0.08 0.14 0.11 0.07 Capital Injections to Average Assets (Per cent) 3 0.18 0.42 0.45 0.19 0.46 0.42 0.19 0.45 0.42 2 0.22 0.52 0.54 0.39 0.56 0.42 0.33 0.55 0.43 1 0.65 0.39 0.40 0.44 0.45 0.49 0.51 0.43 0.48 Note: Data are unweighted averages of individual bank percentages. *Recovery is either the date of a bank’ s unassisted mer ger, or if the bank survived as an independent institution, the date it re - ceived a CAMEL rating of 1, 2, or 3.banks were the subject of formal enforcement actions, and many others received informal actions. •For all insured banks rated 4 and 5, in the three years before failure or recovery their as - set growth and dividend payout rates declined (see table 1. 12).115(Recovery was defined as ei - ther a CAMEL rating upgrade to 1, 2, or 3 or mer ger without FDIC financial assistance.) Capital injections generally increased over the three years before recovery for the banks that recovered,and from the third to the second year before failure for the banks that failed. 115For dividends, similar results are produced whether dividends are expressed as a percentage of net income or as a per - centage of assets. Capital injections include stock transactions, capital contributed through mer ger, and capital transactions with parent holding companies.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 64 Histor y of the Eighties—Lessons for the F uture•The data in table 1.12 suggest that in the later years of the banking crisis, supervisory efforts to reduce risk taking and insurance losses became increasingly aggressive. During 1992–94, for both failed banks and survivors, the levels to which asset growth rates and divi - dend payouts dropped in the final year before failure or recovery were considerably lower thanhad been the case during the 1980–85 period. 116 Table 1.12 and the preceding pages summarize an analysis of the behavior of problem banks in relation to the dates of their failure or recovery . Problem-bank behavior was also analyzed in relation to the dates of regulatory intervention. For purposes of this secondanalysis, the dates of regulatory intervention were taken to be the dates of on-site examina - tions that led to either formal enforcement actions or downgrades in CAMEL ratings with - out such actions. 117The purpose was to test m ore directly the ef fects of formal and informal enforcement actions on problem-bank behavior . (As noted before, m ost problem banks that did not receive formal enforcement actions received informal ones.) As shown in figure 1.10, at FDIC- and Federal Reserve–supervised banks with CAMEL ratings of 4, median quarterly asset growth rates declined before the date of regulatory intervention and gener - ally remained negative in the four quarters immediately following the intervention.118This was true both for banks that were downgraded to a CAMEL 4 rating and had no formal en - forcement action taken against them and for 4-rated banks that eventually did receive for - mal actions. Growth rates of banks with formal enforcement actions showed greaterchanges, on average, from before intervention to after intervention than growth rates ofbanks without such actions. 119Similar results were produced by other measures of bank be - 116R. Alton Gilbert found that undercapitalized banks during the 1985–89 period generally did not grow rapidly , pay divi - dends, or make loans to insiders. See his “Supervision of Under -Capitalized Banks: Is There a Case for Change,” Federal Reserve Bank of S t. Louis Review 74, no. 4 (1992): 3–20. 117Enforcement data in this analysis are based on 2,398 formal actions issued by the FDIC and 362 by the Federal Reserve. Comparable data are not available for the OCC. Intervention dates are (1) the date of the examination that resulted in adowngrading of the bank to a CAMEL 3, 4, or 5 rating for the first time without a formal enforcement action or (2) the date of the last examination before the issuance of a formal enforcement action for banks receiving such actions. At the end of the examination the bank would normally be informed of conditions that were likely to result in such downgradesor of the likelihood of formal enforcement actions. Actual issuance of the formal enforcement actions would not take place until six to nine m onths after the examination. For FDIC-supervised banks, the median interval between the date of for - mal enforcement actions and the last examination before such actions was 261 days for 4-rated banks and 176 days for 5-rated banks. 118It is not clear that the remedial actions taken by management before regulatory intervention were purely voluntary and would have been undertaken even if such intervention had not been expected. See also Geor ge E. French, “Early Action for Troubled Banks,” FDIC Banking Review 4, no. 2 (1991): 1–12. 119Similar patterns in growth rates were found for banks with CAMEL 5 ratings. For banks with CAMEL ratings of 3 that were subject to formal enforcement actions, however , growth rates were highly variable, perhaps because for these banks the number of such actions was relatively small.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 651990–19951985–1990 1979–1985Figur e 1.10 Median Asset Gr owth Rates of CAMEL 4-Rated Banks befor e and after Regulatory Intervention (Annualized) Note: Data aremedian asset growth rates ofFDIC- andFederal Reserve supervised banks before andafter regulatory intervention. Forthisanalysis, theintervention dates were dates of:– (1) examinations that resulted in the downgrading of the bank 's CAMEL rating to 4 but did not result in a formal enforcement action, or (2) the last examination before the issuance of a formal enforcement action against a bank with a CAMEL 4 rating. Normally ,abank isinformed atthetime oftheexamination oftheprospect ofaCAMEL rating downgrade oraformal enforcement action. Data were runonaconstant population sample foreach period. Thenumber ofobservations ranged from 200toalmost 500forthedifferent periods forbanks downgraded toCAMEL 4rating thatdidnotreceive formal enforcement actions, andfrom 200to300for4-rated banks thatdidreceive formal enforcement actions.Banks Downgraded to CAMEL 4 Rating That Received No Formal Action 4-Rated Banks That Did Receive Formal Action-4 -3 -2 -1 0 1 2 3 4-8-404812 Quarter Relative to InterventionPercent -4 -3 -2 -1 0 1 2 3 4-8-4048 Quarter Relative to InterventionPercent -4 -3 -2 -1 0 1 2 3 4-12-8-4048 Quarter Relative to InterventionPercentAn Examination of the Banking Crises of the 1980s and Early 1990s Volume I 66 Histor y of the Eighties—Lessons for the F uturehavior (see figure 1. 11). Dividend rate reductions and increases in external capital injec - tions began before regulatory intervention and generally continued in the first year after in - tervention, and banks that became the subject of formal enforcement actions showed thegreatest dividend cuts and capital injections. 120Comparable behavior was also exhibited by loan-loss provisions (not shown in figure 1.1 1). The foregoing analysis indicates that bank managements took remedial actions even before the examinations that triggered reductions in CAMEL ratings or led to formal en - forcement actions. Whether these remedial actions were driven by market forces, by man - agement’ s own objectives, or by expectations of future regulatory action cannot be readily ascertained. In any event, regulatory intervention apparently had the ef fect of reinforcing and accelerating these remedial actions. Changes in the behavior of problem banks weregreater for banks that later received formal enforcement actions as compared with bankssubject only to informal actions. However , it is not clear whether these dif ferences in be - havioral change were due primarily to the m ore demanding nature of formal actions or to the condition and behavior of the banks that received them. Formal actions are frequentlytaken when banks fail to comply with informal ones. Such failure may be due to the exis - tence of m ore severe problems at the banks receiving formal actions or to less willingness on the part of their management to cure them. 121 In general, the reduction in asset growth was an indication that m oral hazard was be - ing contained—that troubled banks were not attempting, or were not being allowed, to“grow out of their problems”; indeed, in many cases their assets were shrinking. In the caseof the surviving banks, reduced dividend payouts and increased capital injections helped re - store equity positions and were instrumental in facilitating recovery . In the case of the fail - ing banks, dividend cuts and new capital had the direct ef fect of reducing failure costs. 122 These favorable results, no matter what the immediate stimulus, were consistent with the regulators’ objectives of preventing the failure of troubled banks and reducing the insurance costs of banks that did fail. The policy of encouraging or forcing problem banks to retrench and shrink has been criticized by some observers for inhibiting the banks’ recovery and, in the context of the 1990s, for contributing to the “credit crunch.” For example, it is sometimes ar gued that re - strictions on asset growth may have deprived problem banks of attractive investment op - 120Data for capital injections are annual in figure 1.1 1 because small banks do not report capital injections quarterly . The analysis was confined to 4-rated banks in order to have lar ge samples of banks with similar conditions. 121As noted in Chapter 12, 71 percent of problem banks that failed had received formal enforcement actions, compared with 41 percent of problem banks that survived. This is consistent with the view that formal actions were taken against the m ost unhealthy banks. 122Although dividend payout ratios declined for troubled banks, a significant number of undercapitalized banks did pay div - idends. See David K. Horne, “Bank Dividend Patterns,” FDIC Banking Review 4, no. 2 (1991): 13–24.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 671990–1995 1990–19951985–1990 1985–19901979–1985 1979–1985 00.10.20.300.050.100.150.200.25Percent of Assets 00.40.81.21.62.0Percent of Assets 0.20.40.60.8 -1 0 1 2 300.20.40.60.81.0 Note: Data are averages of individual bank ratios. See note to figure 1.10.Figur e 1.1 1 Dividend Rates and Capital Infusions of CAMEL 4-Rated Banks befor e and after Regulatory Intervention Capital Infusions -1 0 1 2 300.51.01.52.0Dividends YearRelative to Intervention YearRelative to Intervention Banks Downgraded to CAMEL 4 Rating That Received No Formal Action 4-Rated Banks That Did Receive Formal ActionAn Examination of the Banking Crises of the 1980s and Early 1990s Volume I 68 Histor y of the Eighties—Lessons for the F utureportunities and required them to sell high-quality assets they already owned.123Similarly , it is sometimes ar gued that cuts in dividends may have retarded the growth of external capital infusions. It should be remembered, however , that the range of choices available to regula - tors in dealing with problem banks was limited and permeated by uncertainty . Many prob - lem banks had exhibited a tendency toward excessive risk taking and/or managerial andother weaknesses. Amore relaxed supervisory posture might have resulted in the resump - tion of risk taking and an increase in losses when an institution failed. Continued dividendpayments would also have increased insurance losses if failure occurred. It is not surprisingthat bank regulators generally chose the surer course of reducing risk-taking opportunitiesand insurance losses by seeking the retrenchment and shrinkage of problem banks. Effectiveness of Super visor y Tools: Off -Site Sur veillance Off-site m onitoring based on financial reports submitted by banks evolved during the 1980s in response to earlier developments in computer technology and to fundamentalchanges in the OCC’ s examination policies after two lar ge national banks failed in the 1970s. 124The evolution of of f-site m onitoring appeared to justify reductions in examination staffing and frequency . As the number of failures m ounted during the 1980s, however , it be - came clear that of f-site m onitoring was not a substitute for , but potentially a useful comple - ment to, on-site examinations. Compared with on-site examinations, of f-site m onitoring systems have a number of advantages: they are less intrusive and costly , they can be up - dated frequently when new information is received through quarterly Call Reports, they canprovide the basis for a financial evaluation of the bank between examinations, and they arepotentially able to isolate risk factors that may lead to future problems, whereas examina - tions are essentially a measure of the bank’ s current condition. Furtherm ore, Call Report data on which of f-site m onitoring systems are based are lar gely available to the public and can be used by investors and others as the basis for imposing market discipline on thebanks. By identifying those banks that appear to have deteriorated since their last examina - tions, the systems can help regulators allocate examiner resources. The disadvantages of of f-site m onitoring systems are that they provide no direct eval - uation of management, of individual loan characteristics, of underwriting practices, or ofinternal controls and procedures. Moreover , the accuracy of the financial reports on which they are based, particularly the quality of loan portfolios, is dependent on periodic on-siteexaminations. Off-site surveillance systems, despite their distinct advantages, did not play a very helpful role in the 1980s. On the contrary , belief in their usefulness and their potential 123See comments by Joe Peek in volume 2 of this study . 124This section is based on Chapter 13, “Of f-Site Surveillance Systems.” See also Jesse S tiller, OCC Bank Examination: A Historical Overview , (1995).Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 69helped reinforce the idea that fewer on-site examinations were necessary . In addition, with the lar ge number of failed and troubled banks already straining supervisory resources, tar - geting banks for additional examinations was not a high priority (staf f limitations meant that resources were unavailable to examine any additional banks tar geted by of f-site sys - tems). Of f-site systems appear to have worked best when the number of problem institu - tions and failures was not lar ge and when examination resources were suf ficient for identified banks to be examined. Condition and risk factors. Call Report data can be used to provide an indication of the condition of a bank and the level of risk it has undertaken. In this context, conditionvariables are indicators of the current strength or weakness of a bank. Abank in a weak con - dition would typically have low capital and net-income ratios and high nonperforming-loanratios. Such a bank would face insolvency and failure in the near term. Risk factors, on theother hand, are indicators of a longer -term problem. Abank may be pursuing risky policies but still be in a currently healthy condition, with strong earnings and capital. In time, how - ever, the risky policies could result in loan losses, reduced income, deterioration in capital, and eventual failure. (The distinction between condition and risk in this context is essen - tially the same as the distinction between ex post and ex ante risk measures discussedabove.) The possibilities of isolating condition and risk factors by analyzing banks’ financial data are illustrated in figures 1.12 and 1.13. Figure 1. 12 shows various measures of the cur - rent condition of banks—ratios to assets of equity , of equity plus reserves minus nonper - forming loans (coverage), of net income, and of nonperforming loans—as of l982 for banksthat failed four to five years later (in 1986–87) and for banks that existed throughout the pe - riod and never failed. On the basis solely of these condition variables, there was little as of1982 to distinguish banks that subsequently failed from those that did not. 125Although the condition ratios for the future failures were slightly below those for the future survivors,they were nonetheless at levels that would normally be considered healthy; for example, in1982 the average equity/assets ratio of banks that failed in 1986–87 was over 8 percent. As the banks that failed approached their dates of failure, their condition ratios deterioratedmarkedly compared with those of the nonfailures. Comparisons of long-run risk factors show a considerably dif ferent picture (figure 1.13). In 1982 and throughout the subsequent four to five years, the risk profile of banks 125The data in figures 1.12 and 1.13 include, for 1982, all banks that existed in 1982 and failed in 1986–87 and all banks that existed throughout the 1982–87 period and did not fail after 1987. Certain failures are excluded: those due primarily tofraud, cross-guarantee failures subsequent to FIRREA, and bank af filiates of two Texas bank holding companies with CAMEL ratings of 1 and 2 that were essentially branches of the lead bank and were resolved through transactions whose effects were similar to cross-guarantees.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 70 Histor y of the Eighties—Lessons for the F uture1982 1983 1984 1985 1986 1982 1983 1984 1985 1986 C. Return on Assets* *Net income/assets1982 1983 1984 1985 1986-6-4-202PercentB. Coverage Ratio* *(Equity+r eservesnonperformingloans)/assets –-10-5510PercentBank Condition Ratios for Failed and Nonfailed Banks 1982–1986 Note: “Failed” means banks that existed in 1982 and failed in 1986 or 1987; “nonfailed” means banks that existed during the entire period and never failed.Figur e 1.12 D. Nonperforming Loans* *As a percentage of assets1982 1983 1984 1985 198604812PercentA. Equity Ratio* *Equity/assets0246810Percent 0 Banks That Subsequently Failed Banks That Did Not FailChapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 71A. Loans to Assets 1982 1983 1984 1985 19860204060PercentB.Asset Gr owth RatePercent 1982 1983 1984 1985 1986-10010Bank Risk Ratios for Failed and Nonfailed Banks 1982–1986Figur e 1.13 D.Average Employee Salary 1982 1983 1984 1985 198601020$Tho usandsC. Inter est Income and Fees Ratio* *Total interest and fees on loans and leases/total loans and leases1982 1983 1984 1985 1986048Percent Note: “Failed” means banks that existed in 1982 and failed in 1986 or 1987; “nonfailed” means banks that existed during the entire period and never failed.Banks That Subsequently Failed Banks That Did Not FailAn Examination of the Banking Crises of the 1980s and Early 1990s Volume I 72 Histor y of the Eighties—Lessons for the F uturethat failed in 1986–87 was distinctly higher than that of banks that did not fail. Banks that would fail had substantially higher loans-to-assets ratios than survivors did. They also had substantially higher ratios of interest and fee income on their loan and lease portfolios,which suggests that their loans were riskier . Finally , banks that subsequently failed had higher growth rates in 1982 than the banks that did not fail, but as the banks approachedfailure these growth rates were sharply cut back in a manner consistent with the findingscited above on FDIC enforcement actions. Many prediction m odels constructed for the purpose of predicting bank failures use measures of current condition (or ex post risk) as independent variables. Thus, the accuracy of failure predictions falls of f considerably for predictions of failures m ore than one year ahead. As part of the research for this book, an attempt was made to use ex ante risk mea - surements to identify groups of banks that had a high long-term risk of failure. For this pur - pose, nine risk ratios were tested: loans to assets, deposits over $100,000 to liabilities,ROA, asset growth rate, loan growth rate, operating expenses to total expenses, salary ex - penses per employee, interest yield on loans and leases, interest and fee income on loansand leases. The banks were divided into quintiles according to these ratios. The periods of analysis were four and five years from 1980, 1982, 1984, 1986, and l988. In each period therisk ratio with the strongest statistical relationship to failures turned out to be the ratio ofloans to assets. 126For example, 8.20 percent of the banks that were in the highest loans-to- assets ratio quintile in 1984 failed in 1988–89, compared with 2.89 percent of all banks inthe sample, for an increase of 184 percent in the incidence of failure (see table 1. 13). 127 The same statistical procedure was applied to the “low-risk” group (the lowest four quintiles) as measured by the loans-to-assets ratio, and dif ferent risk factors proved to be the best predictors of failure in four to five years (see table 1. 14). ROA was the best pre - dictor of failure for the “low-risk” group in 1984; of the “low-risk” loans-to-assets group in1984, 3.96 percent of the banks in the highest-risk ROA quintile failed in 1988–89. Anumber of observations are in order . First, the risk factors do not predict which in - dividual banks will fail; rather , they identify a group of banks with the highest incidence of failure. Second, the risk group encompassing the highest loans-to-assets ratio quintile plus 126Logit regressions were performed on each of the risk variables where the dependent variable was whether the bank failed or not. The risk variable with the highest predictive power for failure was determined by a Chi-Square test score for each regression. The coef ficients for each quintile of the variable were then compared, and a Chi-Square test was performed to determine which quintile or group of quintiles was the best predictor of failure. The analysis was then repeated on the high- and low-risk quintiles to determine which was the next-best predictor of failure in both groups. See Chapter 13,“Off-Site Surveillance Systems.” 127In an initial inquiry , the ratio of lar ge deposits to total liabilities was found to be the best predictor of failures in 1984 and 1986. However , this ratio was found to be essentially a proxy for location in Texas, where lar ge-scale use of lar ge deposits occurred in part because of restrictions on branching. Once the lar ge-deposit ratio was excluded, the loans-to-assets ratio was the best predictor of failure in all years.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 73Table 1.13 Probability of Failur e Banks in the Highest Loans-to-Assets Quintile All Sample Banks Highest Loans-to-Assets Quintile Increase in Beginning Probability of Number of Probability Number of Percent of Probability Year of Failure Failures of Failure Failures Total Failures of Failure 1980 1.51% 184 3.62% 88 47.8% 140% 1982 2.45 291 6.75 160 55.0 175 1984 2.89 332 8.20 188 56.6 184 1986 2.25 253 6.46 145 57.3 187 1988 1.24 133 3.36 72 54.1 171 Table 1.14 Probability of Failur e for “Low-Risk” Banks (Banks Not in the Highest Loans-to-Assets Quintile) Failur es in Highest-Risk Gr oup of “Low-Risk” Banks Beginning Highest-Risk Indicator for Probability of Number of Percent of Total Failures in Year “Low-Risk” Group Failure* Failures “Low-Risk” Group† 1980 Loan Growth 2.32% 40 41.7% 1982 Interest Yield 3.76 53 40.4 1984 ROA 3.96 65 45.1 1986 ROA 3.74 62 57.4 1988 ROA 2.12 35 57.4 * This is the probability of failure in the 80 percent of banks that are not in the high-risk loans-to-assets quintile. † Excludes failures in the high-risk loans-to-assets quintile.the high-risk ROA group in the remainder of the banks in 1984 accounted for 76 percent of all failures in the entire sample.128Third, to capture 76 percent of the total number of fail - ures in 1988–89, the two risk groups “flagged” a lar ge proportion (34 percent) of the total 128The 76 percent was derived as follows: 188 failures in the high-risk loans-to-assets ratio quintile plus 65 failures in the high-risk ROA group in the four “low-risk” loans-to-assets quintiles, for a total of 253, or 76 percent of all the 332 fail - ures in the sample. See Chapter 13, table 13-A.3.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 74 Histor y of the Eighties—Lessons for the F uturenumber of banks in the sample.129Fourth, m ost banks in the high-risk groups did not fail. For example, nearly 92 percent of the banks in the high-risk loans-to-assets ratio quintile in1984 did not fail four to five years later , in 1988–89. These findings suggest that failing banks shared a comm on characteristic: they fol - lowed a relatively high-risk strategy , as indicated particularly by the ratio of loans to assets, and could be identified well in advance of their failure dates. The findings also indicate that many other banks had similar risk characteristics but were able to avoid failure. As indi - cated above, the success or failure of banks depends on many factors, so predicting failuresfar in advance on the basis of the institutions’ risk characteristics is dif ficult. Specific off-site sur veillance systems. The original of f-site surveillance systems used in the 1970s were a collection of comm only used financial ratios. The OCC’ s system even - tually evolved into the Uniform Bank Surveillance System, whose best-known product isthe Uniform Bank Performance Report (UBPR). The UBPR is a bank-specific report that allows an analyst to compare the financial characteristics of an individual bank with thecharacteristics of comparable (peer) banks. The Federal Reserve and the FDIC developed similar systems. In 1985 the FDIC developed the CAEL system (Capital, Assets, Earnings, Liquidity—Management is not m odeled). This was designed to replicate the examination rating that an “expert examiner” would give an institution solely on the basis of Call Reportdata. Banks were flagged for attention if their CAMEL rating was 2 or worse and their CAEL rating was m ore than one rating worse than their current CAMEL rating. 130 The CAEL system was adopted in the mid-1980s and has been used to help achieve and maintain ef ficient allocation of supervisory resources, primarily by detecting at an early date banks that appear to have a high probability of receiving a CAMEL downgrade at their next on-site examination. CAEL uses 19 financial ratios by which a bank is matched against its peer group. From 1987 to 1994, the CAEL system was reasonably correct in its predic - tions; approximately one-half of all CAMEL downgrades predicted by the system actually occurred within six m onths. The CAEL system identified approximately 25 percent of total rating downgrades in the relevant group (banks downgraded from CAMEL 2 or worse). By design, CAEL misses a lar ge number of actual downgrades in order to avoid tar geting many banks that are in fact in a stable condition. This appears to be appropriate in a regime of fre - quent on-site examinations, for banks whose conditions have deteriorated since their lastexaminations but that were not identified by the CAEL system will in any event be exam - ined without too much delay . 129The 34 percent figure refers to the highest loans-to-assets quintile plus the highest ROA group of the “low-risk” loans-to- assets quintiles, or 3,935 (2,292 + 1,643) banks. This was 34 percent of the 1 1,479 banks in the sample. See Chapter 13, table 13-A.3. 130The various systems of of f-site surveillance are treated in detail in Chapter 13.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 75In the mid-1980s the FDIC also developed the Growth Monitoring System (GMS). Banks flagged by GMS as rapid-growth institutions are identified for of f-site review and may receive increased supervisory attention. The system is based on the levels and quarterly trends of five summary measures: asset growth rate, growth rate of loans and leases, and ra - tios to assets of equity capital, volatile liabilities, and loans and leases plus securities withmaturities of five years or m ore. The system’ s premise is that rapid growth in total assets or loans represents a risky activity . Through the 1980s banks that generated high growth scores in the m odel had a higher -than-average incidence of failure up to four years later . In the years since the CAEL and GMS systems were developed, there has been a sub - stantial body of economic research related to m odeling bank failures and financial distress. The Federal Reserve has based its of f-site surveillance methods on statistical m odeling techniques, beginning with the Financial Institutions Monitoring System (FIMS), whichwas adopted in 1993 and predicted CAEL-like ratings and bank failures. As of mid-1997 the FDIC was considering substantial m odifications in GMS and adoption of a statistical model for predicting CAMEL rating downgrades for banks and thrifts. Open Questions Many of the weaknesses revealed in bank statutes, regulations, and supervisory prac - tices in the 1980s were subsequently addressed and corrected. However , some issues remain open—two in particular: the potential impact of resolving lar ge-bank failures in accordance with FDICIA, and the adequacy of present systems of identifying and pricing risk. Treatment of L arge Banks: Systemic Risk and Market Discipline FDICIA shifted the balance between stability and market discipline toward market discipline. It accomplished this by requiring that the methods used to resolve bank failuresproduce the least cost to the FDIC and by prohibiting the protection of uninsured depositswhen such action would increase the cost to the insurance fund. Under the pre-FDICIA cost test, either the FDIC could choose to sell the failed bank if the estimated resolution cost wasless than that of a deposit payof f or the FDIC could provide open-bank assistance, regard - less of cost considerations, if the bank’ s services were determined to be “essential” to the community . Failures of big banks were generally resolved in ways that protected all de - posits against loss because of fears of depositor runs on other banks, systemwide crisesthrough correspondent accounts, or disruption of the payments system. FDICIA also limits the ability of the Federal Reserve to provide liquidity to problem banks (defined in terms of capital position) through its discount window . For critically un - dercapitalized banks, repayment must be demanded within no m ore than 5 days, and if that limit is violated, the Federal Reserve is liable to the FDIC for any additional cost. In theAn Examination of the Banking Crises of the 1980s and Early 1990s Volume I 76 Histor y of the Eighties—Lessons for the F uturecase of undercapitalized banks, Federal Reserve advances can remain outstanding for no more than 60 days in any 120-day period.131 FDICIA increases the likelihood that lar ge banks will be resolved with losses to unin - sured depositors and reduces the likelihood that open-bank assistance will be used to dealwith lar ge troubled banks. An exception to the least-cost test is allowed in cases of systemic risk: two-thirds of the FDIC Board and two-thirds of the Federal Reserve Board would haveto recommend that an exception be made, with the final decision in the hands of the secre - tary of the treasury in consultation with the president. Any loss incurred by the FDIC as a result of using the systemic-risk exception would have to be made up by a special assess - ment on all institutions insured by the same fund. These provisions were designed to dis - courage use of the exception and to increase accountability . The 1980–94 experience provides only limited guidance as to how the rules pre - scribed by FDICIA will af fect future lar ge-bank resolutions. On the one hand, there are the well-known troubles of Continental Illinois, which in 1984 sustained enorm ous with - drawals of foreign deposits through high-speed electronic transfers. At the time there was concern that if uninsured deposits were not protected, Continental’ s correspondent banks would sustain serious losses, possibly with “ripple” ef fects on other major banks that were perceived to be vulnerable. Action by the regulators in assisting Continental Illinois fore - stalled the possibility of such ef fects on other major banks. On the other hand, in numerous cases the FDIC resolved banks through methods that left uninsured depositors unprotected yet had no serious repercussions. 132These were gen - erally smaller banks that did not pose problems of systemic risk.133Another instance was the m odified payof f method used to resolve 13 banks in 1983–84, a method that caused uninsured depositors to suf fer losses: at closure uninsured depositors were paid a portion of their m oney based on the value of the bank’ s assets that it was estimated would be recov - ered in liquidation. At the time of these resolutions there were no flights of deposits from other institutions. Similarly , in the period since FDICIA, resolutions with losses to unin - sured depositors have not produced lar ge-scale withdrawals at other institutions. From 1992 131Adecision by the FDIC to act in the Federal Reserve’ s stead by providing open-bank assistance might have rendered this provision less substantial. However , this avenue was essentially closed by the Resolution Trust Corporation Completion Act of 1993, which ef fectively prohibited—unless the systemic-risk exception had been invoked—the use of BIF or SAIF funds to benefit the shareholders of insured depository institutions, a likely outcome of FDIC open-bank assistance. 132From 1986 through 1991, 199 banks (representing 19 percent of all bank failures) were resolved through means that didnot protect uninsured depositors. Average assets of these banks am ounted to $57 million. See FDIC, Failed Bank Cost Analysis , 1986–1995 . 133Apossible exception was Penn Square Bank, which was closed through a deposit payof f in 1982. Because of Penn Square, “Some banks had dif ficulty rolling over lar ge CDs. The business of brokers, who divide up lar ge deposits and participate them to several banks, was significantly boosted. Depositors generally became m ore selective in their choice of banks” (FDIC, The First Fifty Y ears, 98).Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 77to 1995, uninsured depositors were unprotected in 63 percent of all failures, compared with 19 percent in 1986–91. The experience since the adoption of FDICIA is, of course, hardly a rigorous test. In this period bank profits have increased to record levels, failures haveslowed to a trickle, and no major bank has been threatened. Some studies, published m ostly in the post-FDICIA period, present evidence suggest - ing that investors recognize the risk of loss on uninsured deposits and that the market re - sponds appropriately to new information about risk in banking firms. For example, onestudy found that when banks’ subordinated debt claims were downgraded by Moody’ s rat- ing service, the stock prices of banks with lar ger proportions of insured deposits declined less, and downgraded banks then increased their reliance on insured deposits. 134Another study found that stock prices reacted negatively after a downgrade in a bank’ s CAMEL rat- ing, and suggested that such information may be transmitted to the market through thebank’ s Call Reports. 135Astudy of subordinated debt concluded that yields on such instru - ments rationally reflected changes in the government’ s policy toward protecting lar ge bank holding company creditors.136Still another concluded that bond rating agencies convey new information to the market and thereby enhance market discipline, since banks that ex - perience downgrades suf fer negative stock returns.137 Studies have also been done to compare the accuracy of “inside” information devel - oped through on-site examinations with that of “outside” information available to marketparticipants. For example, one study of problem banks concluded that stock returns hadfailed to anticipate downgrades in CAMEL ratings; neither the market nor the banks’ man- agements seemed to have been aware of the banks’ problems before the examinations took place. 138Another study concluded that both regulators and market participants price credit risk, but only regulators price capital strength; the results seem to reflect, on the one hand,the supervisors’ concern with preventing bank failures and protecting the deposit insurance fund and, on the other hand, the market’ s emphasis on risk/return trade-of fs. 139But a third study concluded that CAMEL ratings are primarily proxies for available market informa - 134Matthew T. Billet, Jon A. Garfinkel, and Edward S. O’Neill, “Insured Deposits, Market Discipline, and the Price of Risk in Banking,” unpublished paper (November 28, 1995). 135Allen N. Ber ger and Sally M. Davies, “The Information Content of Bank Examinations,” working paper 94-24, Wharton Financial Institutions Center , 1994. 136Mark J. Flannery and Sorin M. Sorescu, “Evidence of Bank Market Discipline in Subordinated Debenture Yields: 1983–1991,” Journal of Finance 51, no. 4 (September 1996): 1347–77. 137Robert Schweitzer , Samuel H. Szewczyk, and Raj Varma, “Bond Rating Agencies and Their Role in Bank Market Disci - pline,” Journal of Financial Ser vices Resear ch6 (1992): 249–63. 138Katerina Sim ons and S tephen Cross, “Do Capital Markets Predict Problems in Lar ge Commercial Banks?” Federal Re - serve Bank of Boston New England Economic Review (May/June 1991): 51–56. 139John R. Hall, Andrew P . Meyer , and Mark D. Vaughan, “Do Markets and Regulators View Bank Risk Similarly?” Federal Reserve Bank of S t. Louis, supervisory policy analysis working paper no. 1-97, February 1997.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 78 Histor y of the Eighties—Lessons for the F uturetion about the condition of banks; the additional informational content of CAMEL ratings did not appear lar ge.140 These studies often address the issue of whether , in m onitoring lar ge, publicly traded banks, market discipline and supervision are interchangeable. However , their results also have a bearing on the issue of the future treatment of lar ge problem banks. If it appears that the market exercises appropriate discipline and readily obtains relevant information, thenthere are grounds for optimism that, in the future, major surprises at lar ge banks may be avoided because weaknesses will become public knowledge at an early stage, the marketwill have suf ficient information to make realistic assessments of bank risk, and investors will be able to distinguish accurately between viable and nonviable banks. Under these con - ditions, the likelihood that contagious runs will cause systemic problems would be reduced.There would be fewer grounds for optimism if it appeared that the market had inadequateor obsolete information (as compared, for example, with information produced by exami - nations) about a bank’ s condition. With respect to contagious runs, the evidence is not clear; some failures apparently have not af fected other banks, whereas others seemingly have. 141Testing for contagious runs on lar ge banks is obviously problematic: federal deposit insurance and the practice of protecting uninsured depositors of lar ge banks eliminated the possibility of such runs dur - ing the 1980s, and an environment highly favorable to banking has minimized their likeli - hood in the 1990s. Experience from the pre-FDIC era or from countries that have no formaldeposit insurance system is not always consistent or clearly applicable to present-day U.S.conditions. 142 The m ost likely scenario in the event of a future lar ge-bank problem is that the FDIC, the Federal Reserve, and the administration will have to make dif ficult judgment calls on whether use of the systemic-risk exception is justified. Such decisions will probably have 140Thomas F . Car gill, “CAMEL Ratings and the CD Market,” Journal of Financial Ser vices Resear ch3, no. 4 (September 1989): 347–58. 141However , one study concluded that “analysis suggests that bank contagion is lar gely firm-specific and rational, as it ap - pears to be in other industries, and that the costs are not as great as they are perceived to be” (Geor ge G. Kaufman, “Bank Contagion: AReview of the Theory and Evidence,” Journal of Financial Ser vices Resear ch8, no. 2 [April 1994]: 123–50). 142Among the studies of this issue are Charles W. Calomiris and Joseph R. Mason, “Contagion and Bank Failures during the Great Depression: The June 1932 Chicago Banking Panic,” 1 10–22; and Fred R. Kaen and Dag Michalsen, “The Ef fects of the Norwegian Banking Crisis on Norwegian Bank and Nonbank S tocks,” both in Proceedings of the 31st Confer ence on Bank S tructur e and Competition, Federal Reserve Bank of Chicago, May 1995, 123–61; Gerald D. Gay , Stephen G . Timme, and Kenneth Yung, “Bank Failure and Contagion Ef fects: Evidence from Hong Kong,” Journal of Financial Re - search(summer 1991): 153–65; Geor ge G. Kaufman, “Bank Contagion: AReview of the Theory and the Evidence,” Jour- nal of Financial Ser vices Resear ch8, no. 2 (April 1994): 123–50; Charles W. Calomiris and Gary Gorton, “The Origin of Banking Panics: Models, Facts and Bank Regulation,” in Financial Markets and Financial Crises, ed. R. Glenn Hubbard (1991), 109–74; and Wall, “T oo-Big-to-Fail after FDICIA,” 7–9.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 79to be made m ore quickly than were decisions relating to lar ge-bank failures in the 1980s. In any event, the combination of least-cost resolutions, PCA, and limitations on Federal Re - serve advances will no doubt increase market discipline and reduce regulatory discretion.These, of course, are what the supporters of these measures sought to achieve. Additional and unintended ef fects of the new requirements may be that some regulatory decisions will have to be made in haste and that the range of potential solutions to lar ge-bank problems will be narrowed. Adequacy of P resent Systems for Identif ying and P ricing Risk Banking operations became m ore complex during the 1980s and deviated increas - ingly from the traditional loan and deposit-taking m odel (the increase in various types of off-balance-sheet activity is one example). These developments pose new risks and have re - quired adaptations in capital standards and reporting requirements to ensure that majortypes of risk are addressed. 143 Another development that has important implications for assessing risk is the contin - ued geographic diversification of the banking industry through consolidation. As m ore banks spread their activities across state boundaries, they will have increased opportunitiesto diversify their loan portfolios. But as a result of consolidation of multibank holding com - panies into out-of-state branch systems, financial reports under current reporting proce - dures will provide increasingly uncertain indications of the geographic concentrations ofcredit risk. 144For example, if multibank holding companies were to consolidate all their bank and thrift af filiates into a single lead bank, 38 states would show an apparent decline in bank loans outstanding, whereas a few states would show substantial gains.145Currently (as of mid-1997) a number of ef forts are being made to ensure that meaningful data on ge - ographic concentrations of lending risk are available. As these remarks suggest, bank regulators are attempting to adapt systems for identi - fying and pricing risk in order to keep up with developments in the banking industry , and one of the principal tools for restraining risk is capital requirements that also serve to triggerincreasingly severe regulatory action under PCA. As emphasized repeatedly in this chapter , 143The revisions in risk-based capital rules are discussed and evaluated in U.S. General Accounting Of fice, Financial Deriv - atives: Actions T aken or Pr oposed since May 1994 (November 1996). 144“Minimum Data Needs in an Interstate Banking Environment,” FDIC staf f analysis, September 16, 1996. 145To the extent that out-of-state af filiates were consolidated into a local lead bank, a particular state would show an increase in loans. To the extent that locally based af filiates were consolidated into an out-of-state lead bank, a particular state would show a decrease in loans.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 80 Histor y of the Eighties—Lessons for the F uturehowever , bank capital positions are poor predictors of failure several years before the fact. If regulatory action were based solely on capital positions, in many cases such action mightcome too late to do much good. Yet a policy of basing costs or penalties on m ore-forward- looking measures would have its own problems. Although ex ante measures of risk—such as the ratio of loans to assets—correctly flagged a lar ge majority of the institutions that failed several years later , they also flagged a much lar ger number of banks that did not fail. The latter group of banks was presumably being compensated—by earning higher returns,at least for a time—for the greater risk it was assuming. Imposing restrictions on this groupof banks might unnecessarily restrain potentially profitable activities. Basing penalties on exante measures of long-term risk might also expose the regulators to char ges of credit alloca - tion, since they might be restraining banks’ efforts to meet rising credit demands in particu - lar regions or sectors of the economy . And basing regulatory restraints on unreliable ex ante risk measures might increase the prospect of a regulator -induced “credit crunch.” All these difficulties may make regulators loath to base supervisory restraints on, or levy penalties on the basis of, ex ante risk measures, a situation raising the possibility that some future episodeof high-risk activity will go unrestrained until the risky behavior is translated into actuallosses and erosion of capital positions. In other words, identifying and restraining risky bankbehavior on a timely basis will continue to be a dif ficult task for bank regulators. Some observers would address the issue by placing greater reliance on bank owners and the marketplace, and less on regulators, to m onitor and restrain risky behavior . Thus, raising regulatory capital requirements considerably above present standards would in - crease stockholders’ stake in the banks, increase their incentive to enforce conservative policies, and provide greater protection for the deposit insurance fund, taxpayers, and theeconomy against the risk of bank failures. However , if capital requirements are set too high, entry into the industry will be discouraged, competition within the industry will be weak - ened, and credit flows through bank and thrift intermediation will be reduced. Atrade-of f exists between the objective of restraining risk through regulatory capital requirements andthe consequences of reduced competition am ong, and credit flows through, depository in - stitutions. Market value accounting has been proposed as a means of substituting the judgment of the marketplace for that of regulators in assessing bank risk. This assumes that market participants are better able (or willing) to evaluate the risk characteristics of depository in - stitutions on the basis of publicly available data than regulators who have access to internalinformation gained through examination of loan files. As has been frequently pointed out, there are serious problems in assigning market values to bank loans that have no secondarymarkets and have little or no inherent marketability because of the dif ficulty of assessing in - formation developed by the banks on the characteristics and behavior of their borrowers.Aside from implementation problems, market value accounting may reduce longer -termChapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 81bank lending, restrict credit flows during periods of falling asset prices, and inject greater instability in the banking system as a result of fluctuations in net worth positions of depos - itory institutions. In short, whether risky behavior is m onitored by regulators, bank owners, or the market, the objective of greater ex ante restraints on risky behavior may conflict withother public policy objectives. 146 Concluding Comment An eminent philosopher once of fered this discouraging view of the lessons policy - makers learn from history: “[P]eople and governments never have learned anything fromhistory , or acted on principles deduced from it.” 147The present study is based on the view that history can be used constructively by policy makers. The lessons to be learned from this history concern the ef fectiveness of the federal bank regulatory and deposit insurance sys - tems during a period of extraordinary stress. How well did they perform in the 1980s, andhow can a study of their performance benefit future policymakers? Despite bank and thrift failures in numbers not seen since the Great Depression, the government’ s promise to protect insured depositors was fully honored: no depositor lost a penny on federally insured deposits, there was no significant disruption of the financial in - termediation process, and a high degree of financial market stability was maintained. These results did not come cheap, but the financial cost for the banking industry was borne by thebanks themselves and by their customers rather than by taxpayers, who ended up bearingmost of the much greater cost of the S&L debacle. There were also other , less-quantifiable costs, particularly those associated with the m oral-hazard risk taking inherent in deposit in - surance. Achief example was the misallocation of resources when banks and thrifts poured funds into high-risk commercial real estate lending, although other factors besides m oral hazard contributed to this outcome, including poorly conceived deregulation and disruptivetax-law changes. In view of these overall results, several lessons can be drawn about theperformance of bank regulators in the 1980s. 1. Pr oblems in the operations of depositor y institutions must be identified at an early stage if serious deterioration in the institutions’ condition is to be pr evented, and early identification r equir es continuous and sometimes burdensome monitoring of the institutions’ activities. Partly to support the objective of reducing the federal work force and partly because of presumed ef ficacy of of f-site m onitoring, the number of bank exam - 146See Allen N. Ber ger, Kathleen Kuester King, and James M. O’Brien, “The Limitations of Market Value Accounting and a More Realistic Alternative,” Journal of Banking and Finance 15 (1991): 753–83; and Allen N. Ber ger, Richard J. Herring, and Gior gio P . Szego, “The Role of Capital in Financial Institutions ,” Journal of Banking and Finance 19 (1995): 393- 430. 147Geor g Wilhelm Friedrich Hegel, Philosophy of Histor y(1832), quoted in John Bartlett, Familiar Quotations , 14th edition.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 82 Histor y of the Eighties—Lessons for the F utureiners and the frequency of on-site examinations were reduced in the first half of the 1980s, at the very time when the number of troubled banks and bank failures began to rise rapidly . As a result, emer ging problems were not always identified on a timely basis, some failures occurred that might have been averted, and losses to the insurance fund were probably in - creased. Examination forces were rebuilt and the frequency of examinations was increasedin the second half of the 1980s, even before legislation requiring such action was passed byCongress in 1991. Up-to-date, on-site examination results appear to yield information onbanks not available through other means, and they help maintain the integrity of Call Reportand other publicly available bank data. In the 1980s, they provided reasonably accurate ad - vance warning of future banking problems, and their accuracy increased during the period. 2. Adequate funding of the deposit insurance agency is essential to ef fective r eg- ulator y contr ol of risk taking by insur ed institutions. The FSLIC suf fered from a num - ber of defects, but am ong the m ost serious was the lack of funding (and the reluctance of the S&L industry and Congress to provide it). As a result, the FSLIC was unable to close lar ge numbers of insolvent S&Ls, which were allowed to continue operating in the hope thathigher -risk investments would pay of f. FDIC resources, although strained during the late 1980s, were suf ficient to close failed banks. Bank regulators generally forced or encour - aged problem banks to cut asset growth, reduce dividend payments, and attract externalcapital. Problem banks were generally not permitted to “throw the long bomb,” and m ost of them survived as independent institutions or were mer ged without FDIC financial assis - tance. With some significant exceptions, m ost problem banks that failed were closed within the time frame later prescribed by the PCA provisions of FDICIA for critically undercapi - talized banks. Forbearance programs mandated or inspired by Congress were administeredin a generally ef fective manner by the bank regulators, in contrast to the unfavorable S&L experience with forbearance. Although other factors obviously af fected the quality of regu - lation, the availability of the financial resources needed to close insolvent institutions wascentral to the bank regulators’ ability to control bank risk and m oral-hazard problems and reduce losses to the insurance fund when failure occurred. 3. The tr eatment of large-bank failur es had undesirable side ef fects, but it is un - clear whether alternative r esolution methods would have been successful in the envi - ronment of the 1980s. Protecting uninsured depositors of lar ge failed banks weakened market discipline and exposed regulators to char ges of treating small banks unfairly . Im- posing losses on uninsured depositors and liquidating a few lar ge banks might have had salutary ef fects on market discipline, and some observers suggest that the regulators should have been m ore willing to take the risk involved in such actions. However , no such experi -Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 83ment was undertaken, and therefore the experience of the 1980s provides little guidance on whether these actions would have led to runs on other lar ge banks and to m ore-general fi - nancial market instability . In other respects, it is clear that the treatment of lar ge banks could have been im - proved. While still profitable and solvent, some lar ge banks that eventually failed were en - gaging in risky behavior that was not suf ficiently restrained by bank regulators. In addition, a few lar ge banks continued to operate with little equity for extended periods before being closed; these banks generated avoidable losses that increased total resolution costs. In theseinstances, m ore-ef fective regulatory action was feasible and could have reduced losses to the insurance fund. 4. Statutor y rules limiting r egulator y discr etion may help pr event a r epetition of the r egulator y lapses that occurr ed in the 1980s, but it r emains to be seen whether such rules will be maintained in a futur e period of widespr ead banking distr ess. Lim- its on the discretionary authority of bank regulators were adopted as part of FDICIA after the banking crisis had lar gely passed, and they have raised few problems in the benign banking climate that has since prevailed. However , the tension between rules and discretion in bank regulation may reappear in some future period of widespread banking problems. Inthe early 1980s, Congress responded to the concerns of the banking and thrift industries andlimited the ability of regulators to close weakened institutions. In that instance, Congressmandated forbearance for thrifts and some banks, delayed recapitalization of the FSLIC’ s insurance fund, and then declined to provide the am ount requested by the Reagan adminis - tration. 148Given this experience, it is dif ficult to predict the ef fect of current statutory rules in some future banking crisis, or the willingness of legislators to retain them. In such a cri - sis, numerous banks might be suf fering substantial operating losses and capital reductions resulting from external shocks and other unforeseen developments. Will it then be politi - cally feasible, for example, to liquidate a significant number of lar ge banks in accordance with least-cost resolution requirements or to close many small and lar ge banks because they fail a statutory solvency test? If so, will such actions be compatible with the objective ofmaintaining financial market stability? Experience in the 1980s provides little basis for con - fident answers to these questions. 5. Bank r egulation can limit the scope and cost of bank failur es but is unlikely to prevent failur es that have systemic causes. The rise in the number of bank failures in the 148National Commission on Financial Institution Reform, Recovery and Enforcement, Origins and Causes, 73.An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 84 Histor y of the Eighties—Lessons for the F uture1980s had many causes that were beyond the regulators’ power to influence or of fset. These included broad economic and financial market changes, ill-considered government policyactions, and structural weaknesses that inhibited geographic diversification and made manybanks vulnerable to regional and sectoral recessions. Earlier implementation of uniformcapital standards and other improvements in regulation might have reduced the number offailures in the 1980s, but it could not have prevented a great many of them. Legislation per - mitting geographic consolidation was a major step toward correcting existing structuralweaknesses in the banking system. However , if significant new structural weaknesses or se - rious economic problems are allowed to develop in the future, bank regulation alone willnot be able to prevent a major increase in the number of bank failures. 6. The ability of r egulators to curb excessive risk taking on the par t of curr ently healthy banks was (and continues to be) limited by the pr oblem of identifying risky ac - tivities befor e they pr oduce serious losses and by competing public policy objectives. As noted, bank regulators were reasonably successful in curbing risk taking on the part ofofficially designated problem banks whose condition had already deteriorated. However , in dealing with ostensibly healthy banks, regulators had dif ficulty restricting risky behavior before the fact, while the banks were still solvent and the risky behavior was widely prac - ticed and currently profitable. It was (and remains) hard to distinguish such behavior fromacceptable risk/return trade-of fs, innovation, and other appropriate activity , or to m odify the behavior of banks while they were (and are) still apparently healthy . Current risk-based cap - ital requirements are forward-looking in the sense that they apply dif ferent weights to dif - ferent asset categories, but the categories are so broad that they permit major increases inhigh-risk loans without requiring m ore capital. On the other hand, current risk-based pre - mium schedules penalize banks after the fact, when losses have already weakened theircondition. In addition to problems of identification, conflicting public policy objectives arealso a limiting factor; this was evident during the “credit crunch” of the early 1990s, whenbank regulators were criticized by legislators and administration of ficials for retarding eco - nomic recovery through their excessive zeal in applying the very supervisory restraints theyhad previously been ur ged to implement. An alternative approach, proposed mainly by academic writers, would be to rely m ore heavily on bank owners and investors, rather than on regulators, to restrain risky behavioron the part of profitable banks; this would be done by raising overall capital standards toconsiderably higher levels than at present in order to increase shareholders’ stake or by adopting market value accounting. Aside from problems of implementation, the potential efficacy of this alternative is also limited by conflicts with other public policy objectives, such as maintaining financial stability and meeting private sector credit demands.Chapter 1 Summar y and Implications Histor y of the Eighties—Lessons for the F uture 857. Dif ferences in perspective among federal bank r egulators may have delayed recognition of the natur e of the pr oblems of the 1980s. Differences am ong the regulators are to be expected, given their various primary responsibilities, and the resulting checks andbalances are frequently cited as one of the main advantages of the present regulatory struc - ture. However , conflicts am ong regulators on the issue of brokered funds persisted until 1985, and on new bank charters until 1989. Arguably , it should have been clear before then that bank failures were the m ost pressing problem, outweighing such considerations as en - couraging innovations in deposit gathering and easing the entry of new institutions intobanking markets. While the present system of divided regulatory responsibilities is believed to have important advantages, in the early 1980s it may have delayed recognition of the se - riousness of a new crisis. * * * Finally , it is appropriate to emphasize that the lessons of the 1980s need to be applied to future problems judiciously . As noted by one of the participants in the FDIC’ s sympo - sium at which an earlier version of this chapter was presented, the problems of the past maybear little or no resemblance to those of the future. 149Therefore, it is important to keep in mind those lessons of the 1980s that appear to be relevant while remaining alert to emer g- ing problems that have few or no precedents in the past. 149See comments by Carter H. Golembe in volume 2 of this study .An Examination of the Banking Crises of the 1980s and Early 1990s Volume I 86 Histor y of the Eighties—Lessons for the F utureFinancial Crises: Explanations, Types, and Implications Stijn Claessens and M. Ayhan Kose WP/13/28 © 2013 International Monetary Fund WP/ IMF Working Paper Research Department Financial Crises: Explanations, Types, and Implications Prepared by Stijn Claessens and M. Ayhan Kose1 January 2013 Abstract This paper reviews the literature on financia l crises focusing on three specific aspects. First, what are the main fact ors explaining financial crises ? Since many theories on the sources of financial crises highlight the importance of sharp fluctuations in asset and credit markets, the paper briefly reviews theoretica l and empirical studies on developments in these markets around financial crises. Second, what are the major types of financial crises? The paper focuses on the main theoretical and empirical explanations of four types of financial crises—currency crises, sudden st ops, debt crises, and banking crises—and presents a survey of the literature that attemp ts to identify these episodes. Third, what are the real and financial sector implications of crises? The paper briefly reviews the short- and medium-run implications of crises for the real economy and financial sector. It concludes with a summary of the main lessons from the literature and future research directions. JEL Classification Numbers: E32, F44, G01, E5, E6, H12 Keywords: Sudden stops, debt crises, banking crises, currency cris es, defaults, policy implications, financial restructuring, asse t booms, credit booms, crises prediction. Author’s E-Mail Address: SC [email protected], [email protected] 1 This paper is written for a forthcoming book, Financial Crises: Causes, Consequences, and Policy Responses, edited by Stijn Claessens, M. Ayhan Kose, Luc Laeven, and Fabián Vale ncia, to be published by the International Monetary Fund. We thank Ezgi Ozturk for outstanding research assistance. This Working Paper should not be reporte d as representing the views of the IMF. The views expressed in this Working Paper are those of the author(s) and do not necessarily represent those of the IMF or IMF policy. Working Papers descri be research in progress by the author(s) and are published to elicit comments and to further debate. 2 Contents Page I. Introduction ............................................................................................................... .............3 II. Explaining Financial Crises ............................................................................................... ...4 A. Asset Price Booms and Busts....................................................................................5 B. Credit Booms and Busts ............................................................................................8 C. Impact of Asset Price and Credit Busts ...................................................................10 III. Types of Financial Crises................................................................................................. ..11 A. Currency Crises .......................................................................................................12 B. Sudden Stops ...........................................................................................................14 C. Foreign and Domestic Debt Crises ..........................................................................15 D. Banking Crises ........................................................................................................18 IV. Identification, Dating and Frequency of Crises .................................................................22 A. Identification and Dating ........................................................................................23 B. Frequency and Distribution .....................................................................................26 V. Real and Financial Implications of Crises ..........................................................................27 A. Real Effects of Crises ..............................................................................................28 B. Financial Effects of Crises ......................................................................................30 VI. Predicting Financial Crises ............................................................................................... .31 VII. Conclusions .............................................................................................................. ........34 References .................................................................................................................... ............41 Figure 1. Evolution of House Price During Financial Crises .............................................................56 2a. Credit and Asset Price Booms ............................................................................................57 2b. Credit Crunches and Asset Price Busts ..............................................................................58 3.Coincidence of Financia l Booms and Crises ........................................................................59 4. Coincidence of Financial Crises ..........................................................................................60 5. Average Number of Financ ial Crises over Decades ............................................................61 6. Coincidences of Recessions and Crises ...............................................................................62 7. Real Implications of Financia l Crises, Crunches, and Busts ...............................................63 8. Financial Implications of Crises, Crunches, and Busts........................................................64 9. Creditless Recoveries ...................................................................................................... .....65 3 I. I NTRODUCTIO N The 2007-09 global financial crisis has been a pa inful reminder of the multifaceted nature of crises. They hit small an d large countries as well as poor an d rich ones. As fittingly described by Reinhart and Rogoff (2009a), “financial crises are an equal opportunity menace.” They can have domestic or external or igins, and stem from private or public sectors. They come in different shapes and sizes, evolve over time in to different forms, and can rapidly spread across borders. They often require immediate and comprehensive policy responses, call for major changes in financial sector and fiscal policies, and can necess itate global coordination of policies. The widespread impact of the latest global financial crisis underlines the importance of having a solid understanding of crises. As th e latest episode has vividly showed, the implications of financial turmoil can be substantial and greatly affect the conduct of economic and financial policies. A thorough anal ysis of the consequences of and best responses to crises has become an integral pa rt of current policy de bates as the lingering effects of the latest crisis are still being felt around the world. This paper provides a selected survey of the literature on financial crises. 2 Crises are, at a certain level, extreme manifestations of the in teractions between the financial sector and the real economy. As such, understanding financial crises requires an understanding of macro- financial linkages, a truly complex challenge in itself. The objective of this paper is more modest: it presents a focused survey consideri ng three specific questions . First, what are the main factors explaining financia l crises? Second, what are the ma jor types of financial crises? Third, what are the real and financial sector im plications of crises? The paper also briefly reviews the literature on the prediction of crises and the e volution of early warning models. Section II reviews the main factors explaining financial crises . A financial crisis is often an amalgam of events, including substantial change s in credit volume and asset prices, severe disruptions in financial intermediation, notably the supply of external financing, large scale balance sheet problems, and the need for large scal e government support. While these events can be driven by a variety of factors, financ ial crises often are preced ed by asset and credit booms that then turn into busts. As such, ma ny theories focusing on th e sources of financial crises have recognized the importance of sharp movements in asset and credit markets. In light of this, this section briefly reviews th eoretical and empirical studies analyzing the developments in credit and asse t markets around financial crises. Section III classifies the types of financial crises identified in many studies. It is useful to classify crises in four groups: currency crises ; sudden stop (or capital account or balance of 2 For further reading on financial crises, the starting point is the authoritative study by Reinhart and Rogoff (2009). Classical references are Minsky (1 975) and Kindleberger (1976). See IMF (1998), Eichengreen (2002), Tirole (2002), Allen and Gale (2007), Allen, Babus, Carletti (2009), Allen (2009), and Gorton (2012) for reviews on causes and consequences of financial crises. 4 payments) crises; debt crises; and banking crises . The section summarizes the findings of the literature on analytical causes and empirical determinants of each type of crisis. The identification of crises is discussed in S ection IV. Theories, that are designed to explain crises, are used to guide the literature on the identification of crises. However, it has been difficult to transform the predictions of the theori es into practice. While it is easy to design quantitative methods to identify currency (a nd inflation) crises and sudden stops, the identification of debt and banking crises is typically based on qualitative and judgmental analyses. Irrespective of the classification one us es, different types of crises are likely to overlap. Many banking crises, for example, are al so associated with sudden stop episodes and currency crises. The coincidence of multiple types of crises leads to further challenges of identification. The literature therefore employs a wide range of methods to identify and classify crises. The section considers various identification approaches and reviews the frequency of crises over time and across different groups of countries. Section V analyzes the implications of fina ncial crises. The macroeconomic and financial implications of crises are typically severe and share many commonalities across various types. Large output losses are common to many crises, and other macroeconomic variables typically register significant dec lines. Financial variables, such as asset prices and credit, usually follow qualitatively similar patterns across crises, albeit with variations in terms of duration and severity of declin es. The section examines the short- and medium-run effects of crises and presents a set of stylized facts w ith respect to their macr oeconomic and financial implications. Section VI summarizes the main methods used for predicting crises. It has been a challenge to predict the timing of crises. Financial market s with high leverage can easily be subject to crises of confidence, making fickleness the ma in reason why the exact timing of crises is very difficult to predict. Mo reover, the nature of crises changes over time as economic and financial structures evolve. Not surprisingly, early warni ng tools can quickly become obsolete or inadequate. This section presents a summary of the evolution of different types of prediction models and considers the cu rrent state of early warning models. The last section concludes with a summary a nd suggestions for future research. It first summarizes the major lessons from this literature review. It then considers the most relevant issues for research in light of these lessons. On e is that future research should be geared to eliminate the “this-time-i s-different” syndrome. However, this is a very broad task requiring to address two major questions: How to preven t financial crises? And, how to mitigate their costs when they take place? In addition, there ha ve to be more intensive efforts to collect necessary data and to develop new methodologi es in order to guide both empirical and theoretical studies. II. E XPLAI NING FINANCIAL CRISES While financial crises have common elements , they do come in many forms. A financial crisis is often associated with one or more of the following phenomena: substantial changes in credit volume and asset prices; severe di sruptions in financial intermediation and the 5 supply of external financing to various acto rs in the economy; larg e scale balance sheet problems (of firms, households, financial interm ediaries and sovereigns); and large scale government support (in the form of liquidity support and recapita lization). As such, financial crises are typically multidimensional events and can be hard to characterize using a single indicator. The literature has clarified some of the factors driving crises, but it remains a challenge to definitively identify their deeper causes. Many theories have been developed over the years regarding the underlying causes of crises. While fundamental factors—macroeconomic imbalances, internal or external shocks—are often observed, many questions remain on the exact causes of crises. Financial crises sometimes appear to be driven by “irrati onal” factors. These include sudden runs on banks, contagion and spillovers among financial markets, limits to arbitrage during times of stress, emerge nce of asset busts, credit crunches, and fire- sales, and other aspects related to financial turm oil. Indeed, the idea of “animal spirits” (as a source of financial market movements) has long occupied a significant space in the literature attempting to explain crises (Keyne s, 1930; Minsky, 1975; Kindleberger, 1978). 3 Financial crises are often preceded by asset and credit booms that eventually turn into busts. Many theories focusing on the sour ces of crises have recognized the importance of booms in asset and credit markets. However, explaini ng why asset price bubbles or credit booms are allowed to continue and eventua lly become unsustainable and turn into busts or crunches has been challenging. This naturally requires answering w hy neither financial market participants nor policy makers foresee the risks and attemp t to slow down the expansion of credit and increase in asset prices. The dynamics of macroeconomic and financial va riables around crises ha ve been extensively studied. Empirical studies have documented the various phases of financial crises, from initial, small-scale financial disruptions to large-scale national, re gional, or even global crises. They have also describe d how, in the aftermath of fina ncial crises, asset prices and credit growth can remain depressed for a l ong time and how crises can have long-lasting consequences for the real economy. Given thei r central roles, we next briefly discuss developments in asset and credit markets around financial crises. A. Asset Price Booms and Busts Sharp increases in asset prices, sometimes called bubbles, and often followed by crashes have been around for centuries. Asset prices sometimes seem to deviate from what fundamentals would suggest and exhibit patterns different than predictions of standard models with perfect financial markets. A bubble, an extreme form of such deviation, can be defined as “ the part of a grossly upward asset price mo vement that is unexplainable based on fundamentals” (Garber, 2000). Patterns of exuberant in creases in asset prices, often followed by crashes, figure prominently in many accounts of financial instabil ity, both for advanced and emerging market countries alike, going back millenniums (see Evanoff, Kaufman, Malliaris (2012) and Scherbina (2013) for detailed review s of asset price bubbles). 3 Related are such concepts as “reflexivity” (Sor os, 1987), “irrational exube rance” (Greenspan, 1996), and “collective cognition” (De La Torre and Ize, 2011). 6 Some asset price bubbles and crashes are well known. Such historical cases include the Dutch Tulip Mania from 1634 to 1637, the Fr ench Mississippi Bubble in 1719-20, and the South Sea Bubble in the United Kingdom in 1720 (Garber, 2000; Kindleberger, 1986). During some of these periods, certain asset prices increased very rapidl y in a short period of time, followed by sharp corrections. These cases are extreme, but not unique. In the recent financial crisis, for example, hous e prices in a number of countri es have followed this inverse U-shape pattern (Figure 1). What explains asset price bubbles? Formal models attempting to explain asset pric e bubbles have been developed for some time. Some of these models consider how individu al rational behavior can lead to collective mispricing, which in turn can result in bubble s. Others rely on microeconomic distortions that can lead to mispricing. Some others assu me “irrationality” on th e part of investors. Although there are parallels, expl aining asset price bu sts (such as fire-s ales) often requires accounting for different factor s than explaining bubbles. Some models employing rationa l investors can explain bubbles without distortions. These consider asset price bubbles as agents’ “justif ied” expectations about future returns. For example, in Blanchard and Watson (1982), under rational expectations, the asset price does not need to equal its fundamental value, lead ing to “rational” bubbles . Thus, observed prices, while exhibiting extremely large fluctuations, are not necessarily excessive or irrational. These models have been applied relatively suc cessfully to explain th e internet “bubble” of the late 1990s. Pastor and Veronesi (2006) s how how a standard model can reproduce the valuation and volatility of internet stocks in the late 1990s, thus arguing that there is no reason to refer to a “dotcom bubble.” Bran ch and Evans (2008), employing a theory of learning where investors use most recent (ins tead of past) data, find that shocks to fundamentals may increase return expectations. This may cause stock pr ices to rise above levels consistent with fundamentals. As pric es increase, investors’ perceived riskiness declines until the bubble bursts. 4 More generally, theories s uggest that bubbles can appear without distortions, uncertaint y, speculation, or bounded rational ity (see Garber (2000) and Scherbina (2013) for reviews of models of bubbles). But both micro distortions and macro factors can lead to bubbles as we ll. Bubbles may relate to agency issues (Allen and Ga le, 2007). For example, due to ri sk shifting – as when agents borrow to invest (e.g., margin lending for stoc ks, mortgages for housing), but can default if rates of return are not sufficiently high – prices can esca late rapidly. Fund managers who are rewarded on the upside more than on the downs ide (somewhat analogous to limited liability of financial institutions), bi as their portfolios towards risky assets, which may trigger a 4 Wen and Wang (2012) argue that systemic risk, commonly perceived changes in the bubble's probability of bursting, can produce asset price movements many times more volatile than the economy's fundamentals and generate boom-bust cycles in the context of a DSGE model. 7 bubble (Rajan, 2005).5 Other microeconomic factors (e.g. , interest rate deductibility for household mortgages and corporat e debt) can add to this, possi bly leading to bubbles (see BIS, 2002 for a general review, and IMF (2009) for a review of debt and other biases in tax policy with respect to the recent financial crisis). Investors’ behavior can also drive asset prices away from fundamentals, at least temporarily. Frictions in financial markets (notably those associated with information asymmetries) and institutional factors can affect asset prices. Theory suggests, fo r example, that differences of information and opinions among investors (relat ed to disagreements about valuation of assets), short sales constraints, and other limits to arbitrag e are possible reasons for asset prices to deviate from fundamentals. 6 Mechanisms, such as herding among financial market players, informational cascades, and market se ntiment, can affect asset prices. Virtuous feedback loops – rising asset prices, incr easing net worth positions, allowing financial intermediaries to leverage up, and buy more of the same assets – play a significant role in driving the evolution of bubbles. The phenomenon of contagion – spillovers beyond what “fundamentals” suggest – may have similar root s. Brunnermeier (2001) reviews these models and show how they can help understand bubbles, crashes, and other market inefficiencies and frictions. Empirical work confirms some of th ese channels, but formal econometric tests are most often not powerful enough to separate bubbl es from rational increases in prices, let alone to detect the cause s of bubbles (Gürkaynak, 2008). 7 Bubbles may also be the results of the same factors that ar e argued to lead to asset price anomalies. Many “deviations” of asset prices from the predictions of efficient markets models, on a small scale with no systemic im plications, have been documented (Schwert, 2003 and Lo and MacKinlay, 2001, and earlier Fama 1998 review). 8 While some of these deviations have diminished over time, possibly as investors have implemented strategies to exploit them, others, even though documented ex tensively, persist to today. Furthermore, deviations have been found in similar ways across various markets, time periods, and institutional contexts. As such, anomalies cannot easily be attributed to specific, institution- related distortions. Rather, they appear to reflec t factors intrinsic to financial markets. Studies under the rubric of “behavioral finance” have tried to explain these patterns, with some 5 In Rajan’s (2005) “alpha-seeking” argument, firms, asset managers, and traders take more risk to improve returns, with private rewards in the s hort-run. See Gorton and He (2000) and Dell’Ariccia and Marquez (2000) for theories linking credit booms to the quality of lending standards and competition. 6 Models include Miller (1977), Harrison and Kreps (1978), Chen, Hong and Stein (2002), Scheinkman and Xiong (2003), and Ho ng, Scheinkman and Xiong (2007). 7 Empirical studies include Abreu and Brunnermeier (2003), Diether, Malloy and Scherbina (2002), Lamont and Thaler (2003), Ofek and Richards on (2003), and Shleifer and Vishny (1997). 8 For example, stocks of small firms get higher ra tes of return than other stocks do, even after adjusting for risk, liquidity and other factors. Spr eads on lower-rated corporate bonds appear to have a relatively larger compensation for default risk th an higher-rated bonds do. Mutual funds whose assets cannot be liquidated when investors sell the funds (so called closed-end funds) can trade at prices different those implied by the intrinsic value of their assets. 8 success (Shleifer 2000, and Barberis and Thaler 2003 review).9 Of course, “evidence of irrationally” may reflect a mis-specified model, i.e., irrational behavior is not easily falsifiable. Busts following bubbles can be triggered by sma ll shocks. Asset prices may experience small declines, whether due to cha nges in fundamental values or sentiment. Changes in international financial and economic conditions , for example, may drive prices down. The channels by which such small declines in asset prices can trigger a cr isis are well understood by now. Given information asymmetries, for ex ample, a small shock can lead to market freezes. Adverse feedback loops may then arise, where asset prices exhibit rapid declines and downward spirals. Notably, a drop in prices can trigger a fire sale, as financial institutions experiencing a decline in asse t values struggle to attract sh ort-term financing. Such “sudden stops” can lead to a cascade of forced sales and liquidations of assets, and further declines in prices, with consequences for the real economy. Flight to quality can further intensify fina ncial turmoil. Relationships among financial intermediaries are multiple and complex. Information asymmetries are prevalent among intermediaries and in financial markets. These problems can easily lead to financial turmoil. They can be aggravated by preferences of in vestors to hold debt claims (Gorton, 2008). Specifically, debt claims are “low information-intensive” in normal states of the world – as the risk of default is remote, they require lit tle analysis of the underl ying asset value. They become “high information-intensive,” however , in times of financial turmoil as risks increase, requiring investors to assess default risks, a complex task involving a multitude of information problems. This puts a premium on sa fety and can create perverse spirals. As investors flight to quality assets, e.g., gove rnment bonds, they avoid some, lower quality types of debt claims, leading to sharper drops in their prices (Gor ton and Ordonez, 2012). B. Credit Booms and Busts A rapid increase in credit is another common thread running th rough the narratives of events prior to financial crises. Leverage buildups and greater risk-taki ng through rapid credit expansion, in concert with increases in asset pr ices, often precede crises (albeit typically only recognized with the benefits of hindsight). Both distant past and more recent crises episodes typically witnessed a period of significant growth in credit (and external financing), followed by busts in credit markets along with sharp corr ections in asset prices. In many respects, the descriptions of the Australian boom and bust of the 1880-90s, for example, fit the more recent episodes of financial instability. Likewise , the patterns before the East Asian financial crisis in the late-1990s resemb led those of the earlier ones in Nordic countries as banking 9 For example, firms tend to issue new stocks when prices (and firm profitability) are high and markets’ reaction to initial public offerings can be “hot” or “cold.” Both contradict the assumption that firms seek external financing only when they need to (due to lack of internal funds while having good growth opportunities). Many in dividual investors also appear to diversify their assets insufficiently (or naively) and rebalance their portf olio too infrequently. At the same time, some investors respond too much to price movements, and sell winners too early and hold on to losers too long. These patterns have been “explained” by various behavioral factors. 9 systems collapsed following peri ods of rapid credit growth re lated to investment in real estate. The experience of the United States in the late 1920s and early 1930s exhibits some features similar to the run-up to the recent global financial crisis with, beside rapid growth in asset prices and land speculation, a sharp increase in (household) leverage. The literature has also documented common patterns in various ot her macroeconomic and financial variables around these episodes. What explains credit booms? Credit booms can be triggered by a wide range of factors, including shocks and structural changes in markets.10 Shocks that can lead to credit booms include changes in productivity, economic policies, and capital flows. Some cr edit booms tend to be associated with positive productivity shocks. These generally start during or after period s of buoyant economic growth. Dell’Ariccia and others (2013) find th at lagged GDP growth is positively associated with the probability of a cred it boom: in the three-year period preceding a boom, the average real GDP growth rate reaches 5.1 percent, compared to 3.4 percent during a tranquil three- year period. Sharp increases in international financial fl ows can amplify credit booms. Most national financial markets are affected by global c onditions, even more so today, making asset bubbles easily spill across borders . Fluctuations in capital fl ows can amplify movements in local financial markets when in flows lead to a signi ficant increase in the funds available to banks, relaxing credit constraints for corporat ions and households (Claessens et al. 2010). Rapid expansion of credit and sharp growth in house and other asset prices were indeed associated with large capital inflows in many countries before the re cent financial crisis. Accommodative monetary policies, especially wh en in place for extended periods, have been linked to credit booms and excessive risk taking. The channel is as fo llows. Interest rates affect asset prices and borrower’s net worth, in turn affecting lending conditions. Analytical models, including on the relationship between agency problems and interest rates (e.g., Stiglitz and Weiss, 1981), suggest more risk-taki ng when interest rates d ecline and a flight to quality when interest rates rise, with conseq uent effects on the availability of external financing. Empirical evidence (e.g., for Spain, Maddaloni and Peydró, 2010; Ongena et al. 2009), supports such a channel as credit standard s tend to loosen when policy rates decline. The relatively low interest rates in the U.S. during 2001-04 are often mentioned as a main factor behind the rapid increases in house prices and household le verage (Lansing, 2008; Hirata et. al, 2012). 11 10 For reviews of factors associated with the onset of credit booms, see further Mendoza and Terrones (2008 and 2012), Magud, Reinhart, and Vesperoni (2012), and Dell’Ariccia and others (2013). 11 However, whether and how monetary policy affects risk taking, and thereby asset prices and leverage, remains a subject of further research (see De Nicolo and others (2010) for recent analysis and review). The extent of bank capitalization a ppears to be an important factor as it affects incentives: when facing a lower interest rate, a well-capitalized bank decreases its monitoring and takes more risk, while a highly levered, low capitalized bank does the opposite (see further Dell’Ariccia, Laeven and Marquez (2010)). 10 Structural factors include financial liberalizat ion and innovation. Fina ncial liberalization, especially when poorly design ed or sequenced, and financia l innovation can trigger credit booms and lead to excessive increases in leve rage of borrowers and lenders by facilitating more risk taking. Indeed, financial liberaliza tion has been found to often precede crises in empirical studies (Kaminsky and Reinhart, 1999; Detragiache and Demirguc-Kunt, 2006). Dell’Ariccia and others (2013) re port that roughly a third of booms they identify follow or coincide with financial liberalization episodes. The mechanisms involved include institutional weaknesses as well as the perverse effects of competition. One channel seems to be that re gulation, supervision, and market discipline is slow to catch up with greater competition a nd innovation (possibly se t in motion by shocks or liberalization). Vulnerabilities in credit markets can naturally ar ise. Another mechanism commonly linking booms to crises is a decline in lending standards. Greater competition in financial services, while generally enhancing efficiency and stability in the long run, can contribute to financial fragility over shorter periods. For the latest crisis in the United States, this was evident in higher delinquency rates in those metropolitan area s with higher growth in loan origination prior to the onset of the cr isis, with the deteriorat ion in lending standards appearing in part related to increases in co mpetition (Dell’Ariccia, Igan and Laeven, 2012). C. Impact of Asset Price and Credit Busts Sharp movements in asset and credit markets duri ng financial crises are quite different from those normally observed. Asset prices and cred it booms and busts differ from the movements observed over the course of a normal business cy cle. Booms in credit and asset markets are shorter, stronger, and faster th an other upturns. For example, th ese episodes often take place over relatively shorter time periods than other episodes and are associated with much faster increases in the financial variables (Figure 2A). The slope of a typical boom is two to three times larger than that of re gular episodes. And crunches and busts are longer, deeper and more violent than other downtur ns. Credit crunches and asset price busts have much larger declines than other declines (Figure 2B). Sp ecifically, credit crunche s and house price busts lead to respectively ro ughly 10 and 15 times larger drops th an other downturns, while equity busts more than 2.5 times as large. These epis odes also last longer, some two times, than other downturns, with house price busts the longes t of all, about 18 quart ers, whereas a credit crunch and equity busts last about 10-12 quarters. Moreover, disruptions are more violent, as evidenced by higher slope coefficients, with bust s in equity prices three times more violent than those in credit and house prices (Claessens, Kose and Terrones, 2010a). There are typically adverse real effects of asset price busts and credit crunches on the real economy. 12 Asset price busts can affect bank le nding and other financial institutions’ 12 Some used to be sanguine on the costs of busts in credit and asset markets. Until the most recent crisis, for example, some appeared to be sanguine on the economic cost of bubbles. For example, Roger W. Ferguson, then Vice Chairman of the Fe deral Reserve Board, argued in January 2005 that “recessions that follow swings in asset prices are not necessarily longer, deeper, and associated with a greater fall in output and investment than other recessions…” There are also theories in which even (continued…) 11 investment decisions and in turn the real economy through two channels. First, when borrowing/lending is collatera lized and the market price of colla teral falls, the ability of firms to rely on assets as collateral for new loans and financial institutions’ ability to extend new credit become impaired, which in turn advers ely affect investment. Second, the prospect of large price dislocations arising from fire sales and related financial turmoil distorts decisions of financial institutions to le nd or invest, prompting them inter alia to hoard cash. Through these channels, fire sales can trigger a credit crunch and cause a severe contraction in real activity. Those asset price booms supported through le veraged financing an d involving financial intermediaries appear to enta il larger risks for the economy. Evidence from past episodes suggests that whether excessive movements in a sset prices lead to se vere misallocations of resources depends in large part on the nature of boom and how it is financed. Booms largely involving equity market activities appear to ha ve lower risks of adverse consequences. The burst of the internet bubble of the late 1990s, which largely involved only equity markets, has not been very costly for the real economy. Wh en banks are involved in financing asset price booms, however, as in real estate mortgage a nd corporate sector fina ncing, risks of adverse consequences of a following asse t bust are typically much higher. The main reason is that these booms involve leverage and banks, implying th at the flow of credit to the economy gets interrupted when a bust occurs. The burst of the latest bubble, as it was financed by ba nks (and the shadow banking system) and involving housing, has been very costly. For the most recent episode, Dell’Ariccia et al (2011) report that, in a 40-country sample, almost all the countries with “twin booms” in real estate and credit markets (21 out of 23) ended up suffering from either a crisis or a severe drop in GDP growth rate relative to the count ry’s performance in the 2003–07 period (Figure 3). Eleven of these countries actually suffered both financial sector da mage and a sharp drop in economic activity. In contrast, of the seven countries that experien ced a real estate boom, but not a credit boom, only two went through a sy stemic crisis and, on average, had relatively mild recessions. We present a broader discussion of the real and financial implications of financial crises and disr uptions in Section V. III. T YPES OF FINANCIAL CRISES While financial crises can take various shapes and forms, in terms of classification, broadly two types can be distinguished. Reinhart and Rogoff (2009a) dist inguish two types of crises: those classified using strict ly quantitative definitions; a nd those dependent largely on qualitative and judgmental analys is. The first group mainly includes currency and sudden fully irrational asset bubbles are not necessarily harmfu l or could even be beneficial (Kocherlakota, 2009). Bubbles can allow for a store of value (“co llateral”) and thereby enhance overall financial intermediation through facilitating exchanges, a nd thereby improve overall economic performance. As such, the presence of bubbles per se , whether rational or irrational, need not necessarily be a cause for concern. 12 stop crises and the second group contains debt and banking crises. Regardless, definitions are strongly influenced by the theori es trying to explain crises. While financial crises can take various shapes an d forms, the literature has been able to arrive at concrete definitions of many types of crises. For example, a currency crisis involves a speculative attack on the curren cy resulting in a devaluati on (or sharp depreciation), or forcing the authorities to defend the currency by expending large amount of international reserves, or sharply raising interest rates, or imposing capital controls. A sudden stop (or a capital account or balance of payments cris is) can be defined as a large (and often unexpected) fall in international capital inflows or a sharp reversal in aggregate capital flows to a country, likely taking place in conjunction with a sharp rise in its credit spreads. Since these are measurable variables, they lend themselves to the use of quantitative methodologies. Other crises are associated with adverse de bt dynamics or banking system turmoil. A foreign debt crisis takes place when a country cannot (or doe s not want to) service its foreign debt. It can take the form of a sovereign or private (or both) debt crisis. A domestic public debt crisis takes place when a country does not honor its domestic fiscal obliga tions in real terms, either by defaulting explicitly, or by inflating or otherwise debasing its currency, or by employing some (other) forms of financial repression. In a systemic banking crisis, actual or potential bank runs and failures can induce banks to susp end the convertibility of their liabilities or compel the government to intervene to preven t this by extending liquidity and capital assistance on a large scale. Since these are not so easily measurable variables, they lend themselves more to the use of qualitative methodologies. Other classifications are possible, but regardle ss the types of crises likely overlap. A number of banking crises, for example, are associated with sudden stop episodes and currency crises. We examine analytical causes and empirical dete rminants of each type of crisis in this section and consider the identi fication, dating and frequency of crises in the next section. A. Currency Crises Theories on currency crises, often more precisely articulated than for other types of crises, have evolved over time in part as the nature of such crises has cha nged. In particular, the literature has evolved from a focus on the fundamental causes of currency crises, to emphasizing the scope for multiple equilibria, and to stressing the role of financial variables, especially changes in balance sheets, in tr iggering currency crises (and other types of financial turmoil). Three generati ons of models are typically us ed to explain currency crises that took place during the past four decades. The first generation of models, largely motivated by the collapse in the price of gold, an important nominal anchor before the floati ng of exchange rates in the 1970s, was often applied to currency devaluations in Latin Amer ica and other developing countries (Claessens, 1991). 13 These models are from seminal papers by Krugman (1979) and Flood and Garber 13 Earlier versions of the canonical crisis model we re Salant and Henderson ( 1978) and Salant (1983). 13 (1984), and hence called “KFG” models. They s how that a sudden speculative attack on a fixed or pegged currency can result from rati onal behavior by investor s who correctly foresee that a government has been runn ing excessive deficits financed with central bank credit. Investors continue to hold the currency as long as they expect the exchange rate regime remain intact, but they start dumping it when they anticipate that the peg is about to end. This run leads the central bank to quick ly lose its liquid assets or hard foreign currency supporting the exchange rate. The currency then collapses. The second generation of models stresses the importance of multiple equilibria. These models show that doubts about whether a government is willing to maintain its exchange rate peg could lead to multiple equilibria and cu rrency crises (Obstfeld and Rogoff, 1986). In these models, self-fulfilling proph ecies are possible, in which th e reason investors attack the currency is simply that they expect other inve stors to attack the currency. As discussed in Flood and Marion (1997), policies prior to the attack in the first generation models can translate into a crisis, whereas changes in polic ies in response to a po ssible attack (even if these policies are comp atible with macroeconomic fundament als) can lead to an attack and be the trigger of a crisis. Th e second generation models are in part motivated by episodes like the European Exchange Rate Mechanism crisis, where countries like the UK came under pressure in 1992 and ended up devaluing, even though other outcomes (that were consistent with macroeconomic fundamentals) were possi ble too (see Eichengreen, Rose and Wyplosz (1996), Frankel and Rose (1996)). The third generation of crisis models explores how rapi d deteriorations of balance sheets associated with fluctuations in asset prices, including exchange rates, can lead to currency crises. These models are largely motivated by th e Asian crises of the late 1990s. In the case of Asian countries, macroeconomic imbalances we re small before the crisis – fiscal positions were often in surplus and current account de ficits appeared to be manageable, but vulnerabilities associated with financial and co rporate sectors were large. Models show how balance sheets mismatches in these sectors can give rise to currency crises. For example, Chang and Velasco (2000) show how, if lo cal banks have large debts outstanding denominated in foreign currency, this may lead to a banking cum currency crisis. 14 This generation of models also considers the roles played by banks a nd the self-fulfilling nature of crises. McKinnon and Pill (1996), Kr ugman (1998), and Corsetti, Pesenti, and Roubini (1998) suggest that over-borrowing by banks can arise due to government subsidies (to the extent that governments would bail out failing banks). In turn, vulnerabilities stemming from over-borrowing can trigger curre ncy crises. Burnside, Eichenbaum, and Rebelo (2001 and 2004) argue that crises can be self-fulfilling because of fiscal concerns and volatile real exchange rate movements (whe n the banking system has such a government guarantee, a good and/or a bad equilibrium can result). Radelet and Sachs (1998) argue more 14 Hallwood and MacDonald (2000) provide a deta iled summary of the first and second generation models and consider their extensi ons to different contexts. Krugman (1999), in an attempt to explain the Asian financial crisis, also provides a similar mechanism operating through firms' balance sheets, and investment is a function of net worth. 14 generally that self-fulfilling pa nics hitting financial intermed iaries can force liquidation of assets, which then confirms the pani c and leads to a currency crisis. Empirical research has not been able to di fferentiate which generation of these models provides the best characterization of currency crises. Early work had good success with the KFG model. Blanco and Garber (1986), for example, applied the KFG model to the Mexican devaluations in 1976 and 1981-82 a nd showed crisis pr obabilities to build up to peaks just before the devaluations (Cumby and van Wij nbergen (1989) and Klei n and Marion (1994)). However, while the KFG model worked well in cases where macroeconomic fundamentals grow explosively, it was not successful when fundamentals are merely highly volatile and money-demand unstable. Later empirical work moved away from explicit tests of structural models. Some studies used censored dependent variable models, e.g., Logit models, to estimate crisis probabilities based on a wide range of lagged variables (Eiche ngreen, Rose and Wyploz (1996), Frankel and Rose (1996), Kumar et al (2003)). Others, su ch as Kaminsky, Lizondo, and Reinhart (1998) and Kaminsky and Reinhart (1999), employed signa ling models to evaluate the usefulness of several variables in signaling an impending crisis . While this literature has found that certain indicators tend to be associated with cr ises, the outcomes have been nevertheless disappointing, with the timing of crises very hard to predict (see Kaminsky, Lizondo and Reinhart (1998) for an early review, Kami nsky (2003) for an update, and Frankel and Saravelos (2012) for an extensiv e recent survey up to the 2000s). We will re visit the issue of crisis prediction later. B. Sudden Stops Models with sudden stops make a closer asso ciation with disruptio ns in the supply of external financing. These models resemble the la test generation of currency crises models in that they also focus on balance sheet mismatch es – notably currency, but also maturity – in financial and corporate sector s (Calvo et al., 2006). They te nd to give gr eater weight, however, to the role of international fact ors (as captured, for ex ample, by changes in international interest rates or spreads on ris ky assets) in causing “sudd en stops” in capital flows. These models can account for the current account reversals and the real exchange rate depreciation typically observed during crises in emerging markets. The models explain less well the typical sharp drops in output and total factor pr oductivity (TFP). In order to match data better, more recent sudden stop models introduce various frictions. While counterintuitive, in mo st models, a sudden stop cum currency crisis generates an increase in output, rather than a drop. This happens through an abrupt increase in net exports resulting from the currency depreciation. This has led to various arguments explaining why sudden stops in capital flows ar e associated with large output losses, as is often the case. Models typically include Fisherian channels and financial accelerator mechanisms, or frictions in labor markets, to generate an output drop during a sudden stop, without losing the ability to account for the m ovements of other variables. Following closely the domestic literature, models with financial frictions help to account better for the dynamics of output and productivit y in sudden stops. With frictions, e.g., when 15 firms must borrow in advance to pay for inputs (e .g., wages, foreign inputs), a fall in credit – the sudden stop combined with rising extern al financing premium – reduces aggregate demand and causes a fall in output (Calvo and Reinhart, 2000). Or because of collateral constraints in lending, a sudden stop can lead to a debt-deflation spiral of declines in credit, prices and quantity of collateral assets, resu lting in a fall in output. Like the domestic financial accelerator mechanism, financia l distress and bankruptcies cause negative externalities, as banks become more cautious and reduce new lending, in turn inducing a further fall in credit, and thereby contributing to a recession (Calvo, 2000). These types of amplification mechanisms can make small shocks cause sudden stops. Relatively small shocks – to imported input prices, the world interest rate, or productivity – can trigger collateral constraint s on debt and working capital, especially when borrowing levels are high relative to asset values. Fish er's style debt-deflation mechanisms can then cause sudden stops through a spira ling decline in asset prices a nd holdings of collateral assets (Fisher, 1933). This chain of events immediately affects output and demand. Mendoza (2009) shows how a business cycle model with collateral constraints can be consistent with the key features of sudden stops. Korinek (2010) provides a model analyzing the adverse implications of large movements in capital flows on real activity. Sudden stops often take place in countries with relatively small tradable sectors and large foreign exchange liabilities. Sudden stops have affected countries with widely disparate per capita GDPs, levels of financ ial development, and exchange rate regimes, as well as countries with different levels of reserve coverage. There are though two elements most episodes share, as Calvo, Izquierdo and Mejía (2008) document: a small supply of tradable goods relative to domestic ab sorption – a proxy for potential ch anges in the real exchange rate – and a domestic banking system with la rge foreign–exchange denominated liabilities, raising the probability of a “perverse” cycle. Empirical studies find that many sudden stops ha ve been associated with global shocks. For a number of emerging markets, e.g., those in Latin America and Asia in the 1990s and in Central and Eastern Europe in the 2000s, following a period of large capital inflows, a sharp retrenchment or reversal of capital flows occurred, triggered by global shocks (such as increases in interest rates or changes in commodity prices). Sudden stops are more likely with large cross-border financial linkages. M ilesi-Ferretti and Tille (2011) document that rapid changes in capital flows we re important triggers of local crises during the recent crisis. Other papers, e.g., Rose and Spiegel (2011), however, find little role for international factors, including capital flows, in th e spread of the recent crisis. C. Foreign and Domestic Debt Crises Theories on foreign debt crises and default are closely linked to t hose explaining sovereign lending. Absent “gun-boat” diplomacy, lenders ca nnot seize collateral from another country, or at least from a sovereign, when it refuse s to honor its debt obl igations. Without an enforcement mechanism, i.e., the analogue to domestic bankruptcy, economic reasons, instead of legal arguments, are n eeded to explain why internati onal (sovereign) lending exists at all. 16 Models developed rely, as a gross simplifica tion, on either intertempor al or intratemporal sanctions. Intertemporal sanctions arise because of a threat of cutoff fr om future lending if a country defaults (Eaton and Ge rsovitz, 1981). With no access (for ever or for some time), the country can no longer smooth idiosyncratic inco me shocks using international financial markets. This cost can induce the country to continue its debt payments today, even though there are no immediate, direct costs to defaul t. Intratemporal sanctions can arise from the inability to earn foreign exchange today because trading partners impose sanctions or otherwise shut the country out of international markets, again forever or for some time (Bulow and Rogoff, 1989a). Both types of co sts can support a certain volume of sovereign lending (see Eaton and Fernandez, (1995) and Panizza, Sturzenegger and Zettelmeyer (2009) for reviews). These models imply that inab ility or unwillingness to pay, i.e., default, can result from different factors. The incentives governments face in repaying debt differ from those for corporations and households in a domestic cont ext. They also vary across models. In the intertemporal model, a country defaults when the opportunity cost of not being able to borrow ever again is low, one such case presum ably being when the terms of trade is good and is expected to remain so (Kletzer and Wright, 2000). In the intratemporal sanction model, in contrast, the costs of a cutoff from tr ade may be the least when the terms of trade is bad. Indeed, Aguiar and Gopinath (2006) demonstr ate how in a model with persistent shocks, countries default in bad times to smooth consum ption. The models thus also have different implications with respect to a country’s borrowing capacity. Such models are unable, however, to fully account why sovereigns default and why creditors lend as much as they do. Many models actually predict that default does not happen in equilibrium as creditors and de btors avoid the dead-weight cost s of default a nd renegotiate debt payments. While some models have been calibrated to match actual experiences of default, models often still underpredict the li kelihood of actual defaults. Notably, countries do not always default when times are bad, as most models predict: Tomz and Wright (2007) report that in only 62 percent of defaults cases output wa s below trend. Models also underestimate the willingness of inve stors to lend to countries in spite of large default risk. Moreover, changes in the institutional environmen t, such as those implemented after the debt crises of the 1980s, do not appear to have modified the relation between economic and political variables and the probabi lity of a debt default. Togeth er, this suggests that models still fail to capture all asp ects necessary to explain defau lts (Panizza, Sturzenegger and Zettelmeyer, 2009). Although domestic debt crises have been prev alent throughout history, these episodes had received only limited attention in the literature until recen tly. Economic theory assigns a trivial role to domestic debt crises since models often assume that governments always honor their domestic debt obligations—the typical assumption is of the “risk-free” government assets. Models also often assume Ricardian equivalence, making government debt less relevant. However, recent reviews of histor y (Reinhart and Rogoff, 2009a) shows that few countries were able to escape default on do mestic debt, with often adverse economic consequences. 17 This often happens through bouts of high infla tion because of the abuse of governments’ monopoly on currency issuance. One such episode was when the U.S. experienced a rate of inflation close to 200 percent in the late 1770 s. The periods of hyperinflation in some European countries following the World War II were also in th is category. Debt defaults in the form of inflation are often followed by cu rrency crashes. In the past, countries would often “debase” their currency by reducing the meta l content of coins or switching to another metal. This reduced the real value of government debt and thus provide d fiscal relief. There have also been other forms of debt “def ault,” including through financial repression (Reinhart, Kirkegaard, and Sbrancia, 2011). After inflation or debasing cr ises, it takes a long time to convince the public to start using the currency with confidence again. This in turn significantly increase s the fiscal costs of inflation stabi lization, leading to large negative real effects of high inflation and a ssociated currency crashes. Debt intolerance tends to be associated w ith the “extreme duress” many emerging economies experience at levels of external debt that would often be easily managed by advanced countries. Empirical studies on debt intolerance and serial default suggests that, while safe debt thresholds hinge on country specific factors, such as a country’s record of default and inflation, when the external debt level of an emerging economy is above 30-35 percent of GNP, the likelihood of an external debt crisis rises substantially (Reinhart and Rogoff, 2009b). More importantly, when an emerging mark et country becomes a serial defaulter of its external debt, this increases its debt into lerance and, in turn, makes it very difficult to graduate to the club of countries that have continuous access to gl obal capital markets. Many challenges remain regarding modeling the countries ’ ability to sustain various types of domestic and external debt. An important challe nge is that the form of financing countries use is endogenous. Jeanne (2003) argues that short-term (fore ign exchange) debt can be a useful commitment device for countries to employ good macroeconomic policies. Diamond and Rajan (2001) posit that banks in developing countries have little choice but to borrow short-term to finance illiquid projects given the low-quality institutional environment they operate in. Eichengreen and Hausmann (1 999) propose the “original sin” argument explaining how countries with unfavorable conditi ons have no choice but to rely mostly on short-term, foreign currency denominated de bt as their main source of capital. More generally, although short-term debt can increase vulnerabilities, especially when the domestic financial system is underdeveloped, poorly supe rvised, and subject to governance problems, it also may be the only source of (external) fi nancing for a capital-poor country with limited access to equity or FDI inflows. This makes th e countries’ choice of ac cumulating short-term debt and becoming more vulnerable to crises simultaneous outcomes. More generally, the deeper causes driving debt crises are hard to separate from the proximate causes. Many of the vulnerabilitie s raising the risk of a debt cr isis can result from factors related to financial integration, political econo my and institutional environments. Opening up to capital flows can make countries with profligate governments and weakly supervised financial sectors more vulnerable to shoc ks. McKinnon and Pill (1996, 1998) describe how moral hazard and inadequate supervision combined with unrestricted capital flows can lead to crises as banks incur currency risks. Debt crises are also likely to involve sudden stops, currency or banking crises (or various combina tions), making it hard to identify the initial 18 cause. Empirical studies on the id entification of causes are thus subject to the usual problems of omitted variables, endogeneity and simulta neity. Although using short-term (foreign currency) debt as a crisis predictor may work, for example, it does not constitute a proof of the root cause of the crisis. The difficulty to identify the deeper causes is more generally reflected in the fact that debt crises have also been around throughout history. D. Banking Crises Banking crises are quite common, but perhaps the least understood type of crises. Banks are inherently fragile, making them subject to runs by depositors. Moreover, problems of individual banks can quickly spread to the w hole banking system. While public safety nets – including deposit insurance – can limit this risk , public support comes with distortions that can actually increas e the likelihood of a crisis. Institutional we aknesses can also elevate the risk of a crisis. For example, banks heavily depend on the information, legal and judicial environments to make prudent investment d ecisions and collect on their loans. With institutional weaknesses, risks can be highe r. While banking crises have occurred over centuries and exhibited some common patterns, their timing remain s empirically hard to pin down. Bank Runs and Banking Crises Financial institutions are inherently fragile en tities, giving rise to many possible coordination problems. Because of their roles in maturity transformation and liquidity creation, financial institutions operate with highly leveraged ba lance sheets. Hence, banking, and other similar forms of financial intermediation, can be precarious undertaki ngs. Fragility makes coordination, or lack thereof, a major challe nge in financial market s. Coordination problems arise when investors and/or inst itutions take actions – like wit hdrawing liquidity or capital – merely out of fear that others also take similar actions. Given this fragility, a crisis can easily take place, where large amounts of liquidity or capital are withdraw n because of a self- fulfilling belief – it happens because investors f ear it will happen. Small shocks, whether real or financial, can translate into turmoil in markets and even a financial crisis. A simple example of a coordination problem is a bank run. It is a trui sm that banks borrow short and lend long. This matur ity transformation reflects pr eferences of consumers and borrowers. However, it makes banks vulnerable to sudden demands for liquidity, i.e., “runs” (the seminal reference here is Diamond a nd Dybvig, 1983). A run occurs when a large number of customers withdraw their deposits because they believe the bank is, or might become, insolvent. As a bank run proceeds, it generates its own momentum, leading to a self- fulfilling prophecy (or perverse feedback loop): as more people withdraw their deposits, the likelihood of default increases, and this encourag es further withdrawals. This can destabilize the bank to the point where it faces bankruptcy as it cannot liquidate assets fast enough to cover its short-term liabilities. These fragilities have long been recognized, and markets, institutions, and policy makers have developed many “coping” mechanisms (s ee further Dewatripoi nt and Tirole, 1994). Market discipline encourages institutions to limit vulnerabilities. At the firm level, intermediaries have adopted risk manageme nt strategies to re duce their fragility. 19 Furthermore, micro-prudential regulation, with s upervision to enforce rules, is designed to reduce risky behavior of individual financial institutions and can help engineer stability. Deposit insurance can eliminate concerns of small depositors and can help reduce coordination problems. Lender of last resort f acilities (i.e., central banks) can provide short- run liquidity to banks during pe riods of elevated financial stress. Policy interventions by public sector, such as public guarantees, ca pital support and purchas es of non-performing assets, can mitigate systemic risk when financial turmoil hits. Although regulation and safety net measures can help, when poorly designed or implemented they can increase the likelihood of a banking crisis . Regulations aim to reduce fragilities (for example, limits on balance sheet mismatches stemming from interest rate, exchange rate, maturity mismatches, or certain activities of financial institu tions). Regulation (and supervision), however, often finds itself play ing catch up with innova tion. And it can be poorly designed or implemented. Support from th e public sector can also have distortionary effects (see further Barth, Caprio and Levine, 2006). Moral hazard due to a state guarantee (e.g., explicit or implicit depos it insurance) may, for example, lead banks to assume too much leverage. Institutions that know they are too big to fail or unwi nd, can take excessive risks, thereby creating systemic vulnerabilities. 15 More generally, fragilities in the banking system can arise because of policies at both micro and macro levels (Laeven, 2011). History of banks runs Runs have occurred in many countries throughout history. In the U.S., bank runs were common during the banking panics of the 1800 s and in the early 1900s (during the Great Depression). Only with the in troduction of deposit insurance in 1933, did most runs stop in the U.S. (Calomiris and Gorton, 1998). Wide-spr ead runs also happened frequently in emerging markets and developing countries in recent decades, such as in Indonesia during the 1997 Asian financial crisis. Runs occurred mo re rarely in other advanced countries, and even less so in recent decades, in part due to the wide spread availability of deposit insurance. 16 Yet, Northern Rock, a bank speciali zing in housing finance in the U.K., constitutes a very recent example of a bank r un in an advanced country (Shin, 2011). Rapid withdrawals of wholesale market funding also took place during the re cent financial crisis, when several investment and some commercial banks faced large liquidity demands from investors. Widespread runs can also take place in non-ban k financial markets. For example, in the U.S. during the fall of 2008, some mutu al funds “broke the buck”, i.e ., their net asset value fell below par. This triggered sharp outflows from individual investors and many other mutual 15 Ranciere and Tornell (2011) m odel how financial innovations can allow institutions to maximize a systemic bailout guarantee, and report evidence suppor ting this mechanism in the context of the 2007 US financial crisis. 16 Deposit insurance, first introduced in the U.S. in 1933, was adopted following the World War II by many advanced countries, and has since employed by developing countries (Demirguc-Kunt, Kane and Laeven, 2008). While deposit insurance can redu ce the risk of bank runs, it can have severe negative side effects, including increased moral hazard, leading to more risk taking. 20 funds (Wermers, 2012). This “run”, in turn, led the government to provide a guarantee against further declines. These guarantees constitu te a continued source of fiscal risk as the government might be forced to step in to pr event a run again. Other investment vehicles specializing in specific asset classes (such as emerging markets) also experienced sharp outflows as there was a general “flight to safety” (i.e., more demand for advanced countries’ government bonds and T-bills). More generally , the 2007-08 crisis has been interpreted by many as a widespread liquidity run (Gorton, 2009). Deeper causes of banking crises Although funding and liquidity problems can be triggers or proximate causes, a broader perspective shows that banking cr ises often relate to problems in asset markets. Banking crises may appear to originate from the liabil ity side, but they typically reflect solvency issues. Banks often run into problems when many of their loans go sour or when securities quickly lose their value. This happened in cris es as diverse as the Nordic banking crises in the late 1980s, the crisis in Japan in the late 1990s, and the recent crises in Europe. In all of these episodes, there were actually no large- scale deposit runs on banks, but large-scale problems arising from real estate loans made many banks undercapitalized and required support of governments. Problems in asset markets, such as those related to the subprime and other mortgage loans, also played a major role part during the recent crisis. These types of problems in asset markets can go undetected for some time, and a banking crisis often comes into the open through the emergence of fundi ng difficulties among a large fraction of banks. Although the exact cau sal sources are often hard to iden tify, and risks can be difficult to foresee beforehand, looking back banking crises and other financial panics are rarely random events. Banking panics more likely occur near th e peak of the business cycle, with recessions on the horizon, because of concer ns that loans do not get re paid (Gorton 1988; Gorton and Wilton, 2000). Depositors, noticing the risks, demand cash from the banks. As banks cannot (immediately) satisfy all requests, a panic may occur. The large scale bank distress in the 1930s was traced back this way to shocks in the real sector. In many emerging markets, banking crises were triggered by external develo pments, such as sharp movements in capital flows, global interest rates and commodity pric es, which in turn led to an increase in non- performing loans. Panics can too be policy induced. Panics ca n take place when some banks experience difficulties and governments intervene in an ad-hoc manner, without providing clear signals as to the status of other institutions. Th e banking panic in Indonesia in 1997, has been attributed to poorly-managed early interven tions (see Honohan and Laeven, 2007, for this and other case studies). Runs can also be di rectly triggered by govern ment actions: the runs on banks in Argentina in 2001 occurred when the government imposed a limit on withdrawals, making depositors question the soundness of the entire banking system. The recent financial crisis in advanced countries ha s in part been attributed to the lack of consistency across government interventions a nd other policy measures (e.g., Calomiris, 2009). 21 Structural problems can also l ead to banking crises. Studies (e.g., Lindgren, Garcia and Saal, 1996; Barth, Caprio and Levine, 2006, and ma ny others) have identified some common, structural characteristics re lated to banking crises. These include notably: poor market discipline due to moral hazard and excessive deposit insurance; limited disclosure; weak corporate governance framework; and poor supervis ion, in part due to conflict of interests.17 Other structural aspects found to increase the ri sk of a crisis include: large state-ownership and limited competition in the financial system, including restricted entry from abroad; and an undiversified financial system, e.g., a dominance of banks (World Bank, 2001). Because the financial sector receives many fo rms of public support, policy distortions that can lead to crises easily arise. In the context of the recent financial crisis in the US, large government support for housing finance (thro ugh the government sponsored enterprises Fannie Mae and Freddie Mac) has been argued to lead to excessive risk taking. The tendency to pursue accommodative monetary and fiscal pol icies following crises , at least in some advanced countries, can also be interpreted as a form of an ex-post systemic bailout, which in turn distorts ex-ante incentives and can lead to excessive risk taki ng (Farhi and Tirole, 2010). Another often cited problem has been “connected lending” which leads to perverse incentives – as corporations and politicians borrow too much from banks – and can cause a buildup of systemic risk. Some well-studied cases of this phenomenon include Mexico (La Porta et al. 2000; Haber 2005), Russia (Laeven, 2001), and Indonesia (Fisman, 2000). Systemic banking panics still require furthe r study as many puzzles remain, especially regarding how contagion arises . The individual importance of the factors listed above in contributing to crises is not known, in part si nce many of them tend to be observed at the same time. Fragilities remain inherent to th e process of financial intermediation, with the causes for panics often difficult to understa nd. For reasons often unknown, small shocks can result significant problems for the entire fina ncial system. Similarly, shocks may spillover from one market to another a nd/or from one country to othe rs leading to financial crises . The latest financial crisis had many elements in its genesis common to other crises. Much has been written about the causes of the recent crisis (see Calomiris (2009), Gorton (2009), Claessens et al. (2012), and many others). While observers differ on the exact weights given to various factors, the list of factors common to previous crises is ge nerally similar. Four features often mentioned in common are: (1) a sset price increases that turned out to be 17 Failures in regulation and supervision remain the most mentioned cause for crises, despite significant upgrading of regulati ons, supervisory capacity and expertise over decades. For analysis how weaknesses in regulation and supervisi on contributed to the recent crisis, see Čihák, Demirgüç- Kunt, Martínez Pería and Mohseni-Cheraghlou (2012). Analysis suggests though that the design of regulation matters for the risk of financial dist ress. Barth, Caprio and Levine (2006; 2012), for example, suggest not relying solely on regulation an d supervision. Rather, they advocate, inter alia, for an active but carefully balanced mix of market discipline and official regulation and supervision. This should all be supported by in stitutional infrastructure that protects property rights, allows for competition, including engagement with global financ e, and ensures adequate information. The wider threats to financial stability, including those aris ing from political economy and corruption, should be kept at bay. 22 unsustainable; (2) credit booms that led to exce ssive debt burdens; (3) build-up of marginal loans and systemic risk; and (4) the failure of regulation and supervis ion to keep up with financial innovation and get ahead of the crisis when it erupted.18 The global financial crisis was, however, also rooted in so me new factors. Four key new aspects often mentioned are: (1) the widesp read use of complex and opaque financial instruments; (2) the increased interconnectedness among financial markets, nationally and internationally, with the U.S. at the core; (3 ) the high degree of leverage of financial institutions; and (4) the central role of the household sector. These factors, in combination with the ones common to other crises, and fue lled at times by poor government interventions during different stages, led to the worst financia l crisis since the Great Depression. It required massive government outlays and guarantees to re store confidence in financial systems. The consequences of the crisis are still being felt in many advanced countries and the crisis is still ongoing in some European countries. IV. I DENTIFICATIO N, DATI NG A ND FREQUE NCY OF CRISES A large body of work has been devoted to th e identification and dating of crises, but ambiguities remain. Methodologies based on the main theories explaining various types of crises can be used to identify (and accordingly classify) crises. 19 In practice, however, this is not straightforward. While currency (and inflat ion) crises and sudden stops lend themselves to quantitative approaches, the dating of debt and banking crises is typically based on qualitative and judgmental analyses. Irrespectiv e of type, variations in methodologies can lead to differences in the star t and end dates of crises. And, as noted, various types of crises can overlap in a single episode, creating possible ambiguities as to how to classify the episode. This in part because the frequency and types of financial crises have evolved over time. In practice, a wide range of quantit ative and qualitative methods i nvolving judgment are used to identify and classify crises. The data also shows that crises have evolved over time. For example, currency crises were dominant duri ng the 1980s whereas banki ng crises and sudden stops became more prevalent in the 1990s and 2000s. This section begins with a summary of common identification and dating methods (see also IMF WEO 1998; Re inhart and Rogoff, 2009a; and Laeven and Valencia, 2008, 2012). It then provides a summary of the frequency of crises over time, across groups of countri es, and the overlap among types of crises. 18 Specifically, there was an increase in real est ate prices in many markets around the world, paralleled by a run-up in other asset prices, espec ially in equity. Reinha rt and Rogoff (2008) demonstrate that the appreciation of equity and hou se prices in the U.S. before the crisis was even more dramatic than appreciations experienced befo re the “Big Five” post-war debt crises. As the global crisis unfolded, those countries that had expe rienced the greatest increas es in equity and house prices during the boom found themselves most vulnera ble (see Feldstein, 2009, and Teslik, 2009). Unfortunately, the similarity in crises patterns w as, as is often the case, only recognized ex-post. 19 Dating does not of course establish causes, including whether the event was a rational outcome to some other “cause” (e.g., a crash in an asset price may be rational in response to a real shock or not). 23 A. Identification and Dating Currency crises, as they involve large changes in exch ange rates, and (related) inflation crises, are relatively easy to identify. Reinhart and Rogoff (2 009a) distinguish these episodes by assigning threshold values for the relevant va riables. In the case of currency crises, they consider exchange rate depreciations in excess of 15 percent per year as a crisis, while, for inflation, they adopt a thre shold of 20 percent per year.20 A currency crisis is defined in Frankel and Rose (1996) as a depreciation of at least 25 percent cumu lative over a 12-month period, and at least 10 percentage points greate r than in the preceding 12 months. The dates identified are obviously sensitive to such thre sholds used. These thresholds can also be universal, specific to the sample of countries under study, or country-specific (as when the threshold is adjusted for the country’s “normal” exchange ra te variations). A measurement issue naturally arises when th ere was no significant adjustment in currency, even if there were pressures or attacks. Movements in international reserves or adjustment in interest rates can absorb excha nge rate pressures and prevent or moderate the fluctuations in the rate. However, episodes invo lving such pressures and/or at tacks are also important to document and study. To address this, starting with Eichengreen, Rose and Wyplosz (1996), different methodologies have been employed. A co mposite index of speculative pressure is often constructed based on actual exchange ra te changes, and movements in international reserves and interest rates, with weights chos en to equalize the variance of the components, thereby avoiding one component dominating the i ndex. Thresholds are then set to date the currency events, including both larg e exchange rate movements a nd periods of pressure (see Frankel and Saravelos (2012) and Glick and Hutchison (2012) for reviews; Cardarelli, Elekdag and Kose (2010) for applications). Sudden stops and balance-of-payments crises can also be objectively classified. Calvo, Izquierdo and Talvi (2004) define systemic sudden stop events as episodes with output collapses that coincide with large reversals in capital fl ows. Calvo, Izquierdo and Mejía (2008) expand on these criteria in two ways: one, the period c ontains one or more year-on- year fall in capital flows that are at least two standard deviations below its sample mean (this addresses the “unexpected” requi rement of a Sudden Stop); two, it starts (ends) when the annual change in capital flows falls (exceeds) one standard deviation below (above) its mean (Mauro and Becker, 2006). Since methodologies vary, various samples of even ts follow. Calvo et al . (2004) identified 33 Sudden Stop events with large and mild output collapses in a sample of 31 emerging market countries. While studies use di fferent cutoff criteria (Calvo and Reinhart (1999), Calvo, Izquierdo and Loo-Kung (2006), and Milesi-Fer retti and Razin (2000), for example differ), 20 Their comprehensive analysis also includes th e 1258-1799 period during which the principal means of exchange was metallic coins. During this earlie r era, instead of modern inflation and currency crises, there were a number of episodes of cu rrency debasements which were associated with a reduction in the metallic content of coins in circulat ion in excess of 5 percent. They also consider the introduction of a brand new currency replacing a much-d epreciated earlier currency in circulation as another form of currency debasement, which h as still been practiced in the modern era. 24 the datings of events are very similar. Some st udies also require a fall in output, but later studies excluded this requirement (since a fall may be endogenous) and replaced it with the requirement of large spikes in the Emerging Markets Bond Index (EMBI) spread, indicating a shift in the supply of foreign capital (see further Izquierdo, 2012). Cardarelli, Kose and Elekdag (2010) consider a large capital inflow episode to end “abruptly” if the ratio of net private capital inflows to GDP in the year after the episode terminates is more than 5 percentage point lower than at the end of the episode – closely following the definition of “sudden stops” in the literature. An episode is also considered to finish abrup tly if its end coincides with a currency crisis. Balance-of-payments crises and other parallel episodes can similarly be identified using capital flows data. Although there are some di fferences in approach es (e.g., how reserves losses are treated) and statistical variations across studies (e.g., whether the same current account deficit threshold is used for all coun tries or whether country-specific variables thresholds are used), but many of them point to similar samples of actual events. Forbes and Warnock (2012) analyze for a large set of countri es gross flows, instead of the more typical net capital flows (or current account). They id entify episodes of extreme capital flow movements using quarterly data, differentiatin g activity by foreigners and domestics. They classify episodes as “surge”, “s top”, “flight,” or “retrenchment, with surges and stops related respectively to periods of large gross capital in- or outflows by foreig ners, and flights and retrenchments respectively related to periods of large capital out- or inflows by domestic residents. External sovereign debt crises are generally easy to identify as well, although there remain differences in classifications across studies. Sove reign defaults are relatively easy to identify since they involve a unique ev ent, the default on payments. Typical dating of such episodes relies on the classification of rating agencies or on information from international financial institutions (see McFadden, Eckaus, Feder, a nd Hajivassiliou (1984); and papers summarized in Sturzenegger and Zettelmeier (2007)). Still, there are choi ces in terms of methodology. For example differences arise from considering the ma gnitude of defaults (whether default has to be widespread or on just one cl ass of claims), default by type of claims (such as bank claims or bond claims, private or public claims), and the length of default (missing a single or several payments). Others look instead at the in creases in spreads in sovereign bonds as an indicator of (the probability of) default (Edwards, 1984). The end of a default is harder to date though. A major issue with dating, including of default and sovereign debt crises, can be identifying their e nd, i.e., when default is over. Some studies date this as when countries regained acc ess in some form to pr ivate financial markets. Others use as a criteria when countries rega in a certain credit rating (IMF, 2005 and 2011). Differences consequently arise as to how l ong it takes for a country to emerge after a sovereign default. Domestic debt crises are more difficult to identify. First, consistent historical data on domestic public debt across countries was missing, at least until recently. Furthermore, following a crisis, unrecorded debt obligations can come to light. However, Abbas et al (2011) and Reinhart and Rogoff (2009a) have since made significant progress in putting 25 together historical series on (domestic) debt . Second, countries can default in many ways: outright direct defaults; period s of hyper- or high inflation; punitive taxation of interest payments; forced interest rate or princi pal adjustments or conversions; gold clause abrogation; debasing of currency; and forms of financial repression. Reinhart and Rogoff (2009a) describe these and make clear that there remains considerable ambiguity in classifications of defaults, especially of “inflation-related default” episodes. Banking crises can be particularly challenging to date as to when they start and especially when they end. Such crises have usually be en dated by researcher s using a qualitative approach on the basis of a combination of events – such as forced closures, mergers, or government takeover of many fina ncial institutions, runs on seve ral banks, or the extension of government assistance to one or more fi nancial institutions. In addition, in-depth assessments of financial conditions have been used as a criterion. Anothe r metric used has been the fiscal costs associated with resolvi ng these episodes. The end of a banking crisis is also difficult to identify, in part sin ce its effects can linger on for some time. There are large overlaps in the dating of banking crises across different studies. Reinhart and Rogoff (2009a) date the beginning of banking crises by two types of events: bank runs that lead to closure of, merging or takeover by the public sector of one or more financial institutions. If there are no runs, they check the closure, merging, takeover, or large-scale public assistance of an important financial in stitution. As they acknowledge, this approach has some obvious drawbacks: it could date cr ises too late (or too early) and gives no information about the end date of these episod es. Still, the classifi cation of Reinhart and Rogoff (2009a) largely overlaps with th at of Laeven and Valencia (2012). Still, there remain differences in the dating of crises which ca n affect analyses. One example of difference is the start of Japan’s banking crisis which is dated by Reinhart and Rogoff (2009a) as of 1992 and as of 1997 by Laeven and Valencia. Another example, with significant implications for analyses, is from Lopez-Salido and Ne lson (2010). Analyzing events surrounding financial market difficulties in the U.S. over the past 60 years, Lopez- Salido and Nelson report thre e distinct crises: 1973–75; 198 2–84; and 1988–91. This differs from Reinhart and Rogoff, who identify only one crisis (1984–91), and Laeven and Valencia (2012) who also have only one crisis, 1988 (and since then 2007), over that period. Importantly, using their new chronology, Lopez-Sa lido and Nelson argue th at crises need not impact the strength of recoveries , in contrast to most claims th at recoveries are systematically slower after financial crises. 21 These differences clearly show the importance of dating. Lastly, asset price and credit booms, busts and crunches, common to many crises, are relatively easy to classify, but again specific approaches vary across studies. Asset prices (notably equity and to a lesser degree house pr ices) and credit volumes are available from standard data sources. Large changes (in nominal or real terms) in th ese variables can thus easily be identified. Still, since approaches and focus vary, so do the cl assifications of booms, busts, and crunches. Claessens, Kose and Terro nes (2012) use the clas sical business cycles 21 Bordo and Haubrich (2012) and Howard, Martin and Wilson (2011) also argue that recoveries following financial crises do not appear to be different than typical recoveries. 26 approach, looking at the level of real asset prices or credit to identify peaks and troughs in these variables. They then focus on the top and bottom quartile of these changes to determine the booms, busts, or crunches. Other methods exis t: large deviations from trend in real credit growth (Mendoza and Terrones, 2008) and from the credit-to-GDP ratio can be used to classify credit booms. And Gourinchas, Valdes , and Landerretche (2001) classify 80 booms based on absolute and relative (to the credit-to-GDP ratio) de viation from trend, but rather than setting the thresholds first, they limit th e number of episodes they want to classify. Regardless, it is important to recognize that di fferent types of crises can overlap and do not necessarily take place as independent events. One type of crisis can lead to another type of crisis. Or two crises can take place simultaneously due to co mmon factors. To classify a crisis as only one type can then be mislead ing when one event is really a derivative of another. Crises in emerging markets, for example, often have been combinations of currency and banking crises, associated with sudden stops in capital flows, and often subsequently turning into sovereign debt crises. Overall, considerable ambiguity remains on the identification and dating of financial crises, which should serve as an important caveat when one reviews the frequency and distribution of cr ises over time as we do in the next section. B. Frequency and Distribution Crises have afflicted both emerging markets a nd advanced countries throughout centuries. In the three decades before 2007, most crises o ccurred in emerging markets. Emerging market crises during those decades include the Latin American crises in the late 1970s-early 1980s, the Mexican crisis in 1995, and the East Asian cr ises in the mid- to late 1990s. “Emerging” markets being more prone to cr ises is not new (Reinhart an d Rogoff, 2013). History shows that many countries which are developed today e xperienced financial cris es when they were going through their own process of emergence, including Australia, Spain, the U.K. and the U.S. in the 1800s. For example, France defaulte d on its external debt eight times over the period 1550-1800. Some advanced countries experi enced crises in recent decades as well, from the Nordic countries in the late 1980s, to the Japan in the 1990s. The most recent crises starting with the U.S. subprime crisis in la te 2007 and then spreadin g to other advanced countries show (once again) that crises can affect all types of countries. Some claim that crises have become more fr equent over time. The three decades after the World War II were relatively crises-free, wher eas the most recent three decades have seen many episodes (Figure 4). Some relate this in crease to more liberalized financial markets, including floating exchange rates, and gr eater financial integration. Indeed, using macroeconomic and financial series for 14 advanced countries for the 1870-2008 period, Jordà, Schularick and Taylor (2012) report no financial cr ises during the Bretton Woods period of highly regulated financial markets an d capital controls. Also, Bordo et al. (2001) argue that the sudden stop problem has become more severe since the abandonment of the Gold Standard in the early 1970s. More recent crises seem to have lasted s horter though, but banking cr ises still last the longest. The median duration of debt defau lt episodes in the post-World War II has been much shorter than for the period 1800-1945, possibly because of improvement in policies in the later period, improved international financ ial markets, or the active involvement of 27 multilateral lending agencies (see further Das a nd others (2012)). Currency and sudden stop crises are relatively short (almost by definition) . With the major caveat that their end is hard to date, banking crises tend to last the longest, consistent with their large real and fiscal impacts. Financial crises clearly often come in bunches. Sovereign defaults tend to come in waves and in specific regions. Jordà, Schularick and Tayl or (2012) report that there were five major periods when a substantial number of now-advan ced countries experien ced a crisis: 1893, the early 1890s, 1907, 1930-31, and 2007-08. Earlier cris es bunched around events such as the Napoleonic Wars. Examples of bunches over the la st three decades include in the 1980s, the Latin America debt crises; in 1992, the European ERM currency crises; in the late 1990s, the East Asian, Russia and Brazil financial crisis ; the multiple episodes observed in 2007-2008, and the ongoing crises in Europe. Periods of wi despread sovereign defaults often coincide with a sharp rise in the number of countries going th rough a banking crisis. These coincidences point towards common factors driv ing these episodes as well as spillovers of financial crises across borders. Some types of crises are more frequent than others. Comparisons can be made for the post Bretton Woods period (while some types of cris es have been documented for longer periods, not all have; and currency crises were non-exis tent during the fixed exchange rate period; together this necessitates the common, but shor ter period). Of the to tal number of crises Laeven and Valencia (2013) repo rt, there are 147 banking crises , 217 currency crises, and 67 sovereign debt crises over the period 1970 to 2011 (note that several countries experienced multiple crises of the same type). However, as noted before, there is some overlap between the various types of crises. Currency crises frequently tend to overlap with banking crises – so called twin crises (Kaminsky and Reinhart, 1999). In addition, sudde n stop crises, not surprisingly, can overlap with currency and balance-of-payments crises, and sometimes sovereign crises (Figure 5). Of the 431 banking (147), currency (217) and soverei gn (67) crises Laeven and Valencia (2013) report, they consider 68 as twin crises, and 8 can be classified as triple crises. The overlaps are thus far from complete. Th ere are also relative differen ces in coincidences of these episodes. A systemic banking crisis, for example, often involves a currency crisis and a sovereign crisis sometimes overlaps with other cr ises, 20 out of 67 sovereign crises are also a banking and 42 also a currency crisis. V. R EAL A ND FINANCIAL IMPLICATIO NS OF CRISES Macroeconomic and financial c onsequences of crises are ty pically severe and share many commonalities across various types. While there are obviously differences between crises, there are many similarities in terms of the patterns macroeconomic variables follow during these episodes. Large output losses are comm on to many crises and other macroeconomic variables (consumption, investment and industr ial production) typically register significant declines. And financial variables like asset pr ices and credit usually follow qualitatively similar patterns across crises, albeit with vari ations in terms of durat ion and severity. This 28 section provides a summary of the litera ture on the macroeconomic and financial implications of crises. A. Real Effects of Crises Financial crises have large economic costs. Cr ises have large effects on economic activity and can trigger recessions (Claessens, Kose, and Terrones, 2009 and 2012). There are indeed many recessions associated with financial crises (Figure 6). And financia l crises often tend to make these recessions worse than a “norma l” business cycle rece ssion (Figure 7). The average duration of a recession associated with a financial crisis is some six quarters, two more than a normal recession. There is also ty pically a larger output decline in recessions associated with crises than in other recessions. And the cumulative loss of a recession associated with a crisis (computed using the lost output relative to the pr e-crisis peak) is also much larger than that of a recession without a crisis. The real impact of a crisis on output can be computed using various approaches. For a large cross-section of countries a nd long time period, Claessens, Ko se and Terrones (2012) use the traditional business cycles methodology to identify recessions. They show that recessions associated with credit crunc hes and housing busts tend to be more costly than those associated with equity price busts. Overall losses can also be estimated by adding up the differences between trend grow th and actual growth for a nu mber of years following the crisis or until the time when a nnual output growth returned to its trend. On this basis, Laeven and Valencia (2012) estimate that the cumulati ve cost of banking crises is on average about 23 percent of GDP during the first four years. 22 Regardless of the methodology, losses do vary across countries. While overall losses tend to be larger in emer ging markets, the large losses in recent crises in advanced countri es (e.g., both Iceland and Ireland’s output losses exceed 100 percent) paint a different picture. The median output loss for advanced countries is now about 33 percent which exceeds th at of emerging markets, 26 percent. Crises are generally associated with significant declines in a wide range of macroeconomic aggregates. Recessions following crises exhibi t much larger declin es in consumption, investment, industrial production, employment, exports and imports, compared to those recessions without crises. For example, th e decline in consumption during recessions associated with financial crises is typically seven to ten times larger than those without such crises in emerging markets. In recessions with out crises, the growth rate of consumption slows down but does not fall below zero. In cont rast, consumption tends to contracts during recessions associated with financial crises, anothe r indication of the signifi cant toll that crises have on overall welfare. There are also large declines in global output during financial cr ises episodes. The significant cost for the world economy associated with the Great Depression ha s been documented in many studies. The global financial crisis wa s associated with the worst recession since 22 These loss numbers rely on an estimated trend gr owth, typically proxied by the trend in GDP growth up to the year preceding the crisis. They can overstate output losses, however, as the economy could have experienced a growth boom before the cris is or been on an unsustainable growth path. 29 WWII, as it saw a 2 percent decline in wo rld per capita GDP in 2009. In addition to 2009, there were two other years after WWII the wo rld economy experienced a global recession and witnessed crises in multiple countries (Kose, Loungani and Terrones, 2013). In 1982, a global recession was associated w ith a host of problems in advan ced countries, as well as the Latin American debt crisis.23 The global recession in 1991 al so coincided with financial crises in many parts of the world, including diff iculties in US credit markets, banking and currency crises in Europe, and the burst of th e asset price bubble in Japan. While the world per capita GDP grows by about 2 pe rcent in a typical y ear, it declined by about 0.8 percent in 1982 and 0.2 percent in 1991. Recent studies also document that recoveries fo llowing crises tend to be weak and slow, with long-lasting effects. Kannan, Scott, and Te rrones (2013) employ cro ss-country data and conclude that recoveries following financial cr ises have been typically slower, associated with weak domestic demand and tight credit c onditions. These findings are consistent with those reported in several other studies (Rei nhart and Rogoff, 2009a; Claessens, Kose, and Terrones, 2012; Papell and Pr udan, 2011; and Jordà, Schula rick and Taylor, 2012). Abiad and others (2013) analyze the medium term imp act of financial crises and conclude that output tends to be depressed s ubstantially following banking crises. Specif ically, seven years after a crisis, the level of output is typically a bout 10 percent lower relati ve to precrisis trend (even though growth tends to even tually return to its precrisi s rate). They report that the depressed path of output is associated w ith long-lasting reductions of roughly equal proportions in the employment ra te, the capital-to-labor ratio, and total factor productivity. From a fiscal perspective, espe cially banking crises can be ve ry costly. Both gross fiscal outlays and net fiscal costs of resolving financial distress and restructuring the financial sector can be very large. For banking crises, Laeven and Valenc ia (2013), estimate that fiscal costs, net of recoveries, associ ated with crisis are on averag e about 6.8 percent of GDP. They can, however, be as high as 57 percent of GDP and in several cases are over 40 percent of GDP (for example Chile and Argentina in th e early 1980s, Indonesia in the later 1990s, and Iceland and Ireland in 2008). Net resolution cost s for banking crises tend to be higher for emerging markets, 10 percent vs . 3.8 percent for a dvanced countries. A lthough gross fiscal outlays can be very large in advanced count ries as well—as in many of the recent and ongoing cases, the final direct fiscal costs have generally been lower in advanced countries, reflecting the better recove ries of fiscal outlays. Debt crises can be costly for the real ec onomy. Borensztein and Panizza (2009), Levy-Yeyati and Panizza (2011), and Furceri and Zdzienicka (2012) all document that debt crises are associated with substantial GDP losses. Furceri and Zdzienicka (2012) report that debt crises are more costly than banking and currency cris es and are typically associated with output declines of 3-5 percen t after one year and 6- 12 percent after 8 years. Gupta, Mishra, and Sahay (2007) find that currency cr ises are often contractionary. 23 Mexico’s default in August 1982 marked the begi nning of the crisis and the region’s decade long stagnation (i.e., the lost decade). A number of Latin American countries, including Argentina, Mexico and Venezuela in 1982, and Brazil and Chile in 1983, experienced debt crises during the period. 30 The combination of financial system restructur ing costs and a slow economy can lead public debt to rise sharply during fi nancial crises. Reinhart and R ogoff (2009a) document that crises episodes are often associated with substantia l declines in tax revenues and significant increases in government spending. For exampl e, government debt on average rises by 86 percent during the three years following a banking crisis. Using a larger sample, Laeven and Valencia (2013) report the median increase in public debt to be about 12 percent for their sample of 147 systemic banking crises. Including i ndirect fiscal costs, such as those resulting from expansionary fiscal policy and reduced fi scal revenues as a consequence of a recession, makes the overall fiscal costs of the recent crises in advanced countries actually greater than those in emerging markets, 21.4 percent vs. 9.1 percent of GDP.24 Although empirical work has not been able to pinpoint the exact reasons, sudden stops are especially costly. Using a panel data se t over 1975–1997 and covering 24 emerging markets, Hutchison (2008) finds that while a currenc y crisis typically re duces output by 2–3%, a sudden stop reduces output by an additional 6–8 percent in the year of th e crisis. The cumulative output loss of a sudden stop is ev en larger, about 13–15 pe rcent over a 3-year period. 25 Edwards (2004) finds sudden stops and cu rrent account reversals to be closely related, with reversals in turn having a negative effect on real growth and more so for emerging markets. Cardarelli, Kose and Elekdag (2010), examin ing 109 episodes of large net private capital inflows to 52 countries ove r 1987–2007, report that the typical post-inflow decline in GDP growth for episode s that end abruptly is about 3 percentage points lower than during the episode, and about 1 percentage point lower than during the two years before the episode. These fluctuations are also accompanie d by a significant deterioration of the current account during the inflow period and a sharp reversal at the end. B. Financial Effects of Crises Crises are associated with large downward correc tions in financial variables. A large research program has analyzed the evolu tion of financial variables around crises. Some of the studies in this literature focus on crises episodes us ing the dates identified in other work, others consider the behavior of the fi nancial variables during periods of disruptions, including credit crunches, house and equity pric e busts. Although results differ ac ross the types of crises, both credit and asset prices tend to decline or grow at much lower rates during crises and disruptions than they do dur ing tranquil periods, confirming the boom-bust cycles in these variables discussed in previous sections. In a large sample of advanced countries (Figure 8), credit declines by about 7 percent, house pric es fall by about 12 percent and equity prices drop by more than 15 percent during cred it crunches, house and equity price busts, respectively (Claessens, Kose and Terrones, 2011) . Asset prices (exchange rates, equity and house prices) and credit ar ound crises exhibit qualitatively sim ilar properties in terms of their 24 Reinhart and Rogoff (2011) provide further statis tical analysis of the linkages between debt and banking crises. 25 Of course, this and other analyses can suffer fro m reverse causality. That is, private agents see events that lead them to predict future drops in a country’s output, and as a result, these agents pull their capital from the country. In this view, anticipat ed output drops drive sudden stops, rather than the reverse. While possible and reasonable, is hard to document or refute quantitatively this view. 31 temporal evolution in advanced and emer ging market countries, but the duration and amplitude of declines tend to be larger for the latter than for the former. The most notable drag on the real economy from a financial crisis is th e lack of credit from banks and other financial institutions. De ll’Ariccia, Detragiache, Rajan (2005) and Klingebiel, Laeven and Kroszner (2007) show how after banking crises, sectors grow slower that naturally need more external financing, likely because banks are impaired in their lending capacity. Recoveries in aggregate out put and its components following recessions associated with credit crunche s tend to take place before the revival of credit growth and turnaround in house prices (Figure 9). These tempor al patterns are similar to those in the case of house price busts, i. e., economic recoveries start befo re house prices bottom out during recessions coinciding with sharp drops in house prices. Both advanced and emerging market count ries have experien ced the phenomenon of "creditless recoveries". Creditless recoveries ar e quite common to financial crises associated with sudden-stops in many emerging market economies (Calvo, Izquierdo and Talvi, 2006). Abiad, Dell’Ariccia, and Li (2013) using a large sample of countries, show that about one out of five recoveries is creditless. Creditless r ecoveries are, as expected, more common after banking crises and credit booms. The average GDP growth during these episodes is about a third lower than during “normal” recoveries. 26 Furthermore, sectors more dependent on external finance grow relatively less and more financially dependent activities (such as investment) are curtailed more (see also Kannan (2009)). Micro evidence for individual countries also shows that financial crises are associated with reductio ns in investment, R&D and employment, and firms passing up on growth opportunities (Campello, Graham, and Harvey, 2010 review evidence for the U.S.). Co llectively, this suggests that the supply of credit following a financial crisis can constrain economic growth. VI. P REDICTI NG FINANCIAL CRISES It has long been a challenge to predict the timi ng of crises. There is obviously a great benefit in knowing whether and if so when a crisis ma y occur: it can help put in place measures aimed at preventing a crisis from occurring in the first place or limiting the damage if it does happen. As such, there is much to be gained from better detecting the likelihood of a crisis. Yet, in spite of much effort, no single set of indicators has proven to explain the various 26 The fact that the economy recovers without cred it growth and increases in asset prices reflects a combination of factors. First, consumption is typi cally the key driver of recoveries. In particular, private consumption is often the most important contributor to output growth during recoveries. Investment (especially non-residential) recovers on ly with a lag, with the contribution of fixed investment growth to recovery often relatively sm all. Second, firms and households may be able to get external financing from sources other than comm ercial banks that are adversely affected by the crisis. These sources are not captured in the aggregat e credit series most studies focus on. Thirdly, there can be a switch from more to less credit-inte nsive sectors in such a way that overall credit does not expand, yet, because of productivity gains, output increases. The aggregate data employed in many studies hide such reallocations of credit acr oss sectors, including between corporations and households that vary in their “credit-intensity.” 32 types of crises or consistently so over time. Periods of turmoil of ten arise in endogenous ways, with possibilities of multiple equilibria and many non-linearities.27 And while it is easier to document vulnerabilitie s, such as increasing asset prices and high leverage, it remains difficult to predict with some accuracy the timing of crises. This section presents a short review of the evolution of the empi rical literature on prediction of crises. 28 Early warning models have evolved over time, with the first generation of models focusing on macroeconomic imbalances. In early crisis prediction models, mostly aimed at banking and currency crises, the focus was largely on macroeconomic and financial imbalances, and often in the context of emerging markets. Ka minsky and Reinhart (1999) show that growth rates in money, credit, and several other vari ables exceeding certain thresholds made a banking crisis more likely. In a comprehensive review, Goldstein, Kaminsky and Reinhart (2000) report that a wide range of monthly indi cators help predict currency crises, including the appreciation of the real exch ange rate (relative to trend) , a banking crisis, a decline in equity prices, a fall in exports, a high ratio of broad money (M2) to international reserves, and a recession. Among annual indicators, the two best were both current-account indicators, namely, a large current-account deficit relativ e to both GDP and investment. For banking crises, the best (in descendi ng order) monthly indicators were : appreciation of the real exchange rate (relative to trend), a decline in equity prices, a rise in the money (M2) multiplier, a decline in real output, a fall in exports, and a rise in the real interest rate. Among eight annual indicators tested, the best were a high ratio of short-term capital flows to GDP and a large current-account de ficit relative to investment. 29 In the next generation of models, still largely geared towards external crises, balance sheet variables became more pronounced. Relevant indi cators found include s ubstantial short-term debt coming due (Berg et al. 2004) . The ratio of broad money to international re serves in the year before the crisis was found to be higher (and GDP growth sl ower) for crises in emerging markets. In these models, fiscal deficit, public debt, inflation, and real broad money growth, however, were often found not to be consisten tly different between crisis and non-crisis countries before major crises. Ne ither did interest ra te spreads or sovereign credit ratings generally rank high in the list of early warning indicators of currency and systemic banking crises. Rather, crises were more likely preceded by rapid real exchange rate appreciation, current account deficits, domestic credit e xpansion, and increases in stock prices. Later models showed that a combination of variables can help identify situations of financial stress and vulnerabilities. Frankel and Sarave los (2012) perform a meta-analysis based on reviews of crises prediction models and seve n papers published since 2002. The growth rate 27 The slow movement of the financial system from stability to crisis is something for which Hyman Minsky is best known, and the phrase "Minsky moment " – the sudden occurrence of an open financial crisis – refers to this aspect of his work, see Minsky (1992). 28 Babecky and others (2012) present a detailed re view of the empirical studies of early warning models. 29 Crespo-Cuaresma and Slacik (2009) report that mo st of the early warning variables for currency crises in the literature are quite fragile whereas the extent of real exchange rate misalignment and financial market indicators appear to be relatively r obust determinants of crisis in certain contexts. 33 of credit, foreign exchange reserves, the real exchange rate, GDP gr owth, and the current account to GDP are the most fre quent significant indicators in the 83 papers reviewed (see also Threhan, 2009; Lane and Milesi-Ferretti, 2011). Crises are typically preceded by somewhat larger current account deficits relative to historical averages , although credit trends more than external imbalances appear to be the best predictor (Schularick and Taylor, 2011; Taylor, 2013; Alessi and Detken, 2011). Global factors can play important roles in driving sovereign, currency, balance-of-payments, and sudden stops crises. A vari ety of global factors is ofte n reported to trigger crises, including deterioration in the terms of trade, and shocks to world interest rates and commodity prices. For example, the sharp rise in US interest rates at the time has been identified as a trigger for the Latin American sovereign debt crises of the 1980s. More generally, crises are often preceded by interest rate hikes in advanced economies and by sudden changes in commodity, esp ecially oil, prices. But low interest rates can matter as well. For example, Jordà, Schularick and Taylor (2011) report that global financial crises often take place in an environment of low intere st rates. Other studies argue that the global imbalances of the 2000s and the recent crisis are intimately connected (Obstfeld and Rogoff, 2009; Obstfeld, 2012). International trade and other real linkages can be channels of transmission, and contagion in financial market s is associated with crises (Forbes, 2012). Studies highlight for example the role of a common lender in part icular in spreading the East Asian financial crisis (Kaminsky and Reinhart, 2001). These global factors can themselves be outcomes, as in the most recent crisis, when in terest rates and commodity prices experienced sharp adjustments following th e onset of the crisis. Overall though, rapid growth in cr edit and asset prices is found to be the most reliably related to increases in financial stress and vulnerab ilities. Borio and Lowe (2002) document that out of asset prices, credit and investment data, a measure based on credit and asset prices is the most useful: almost 80 percent of crises can be predicted on the basis of a credit boom at a one-year horizon, while false positive signals ar e issued only about 18 percent of the time. Building on this, Cardarelli, Elekdag, and Lall (2009) find that banking crises are typically preceded by sharp increases in credit and house prices. Many others have found the coexistence of unusually rapid in creases in credit and asset prices, large booms in residential investment, as well as deteriorating current account balances, to contribute to the likelihood of credit crunch and asset price busts. Recent studies confirm that credit growth is the most important, but still imperfect predictor. Many of the indicators, such as sharp asset pri ce increases, a sustained worsening of the trade balance, and a marked increase in bank leverage, lose predictive significance once one condition for the presence of a credit boom. Still, there are both Type I and Type II errors. As Dell’Ariccia et al (2012) show, not all booms are associated with crises : only about a third of boom cases end up in financial crises. Others do not lead to busts but are followed by extended periods of below-trend economic gr owth. And many booms result in permanent financial deepening and benefit long-term ec onomic growth. While not all booms end up in a crisis, the probability of a crisis increases with a boom. Furt hermore, the larger the size of a boom episode, the more likely it results in a cr isis. Dell’Ariccia and others (2013) find that close to half or more of the booms that either lasted longer than six years (4 out of 9), 34 exceeded 25 percent of average annual growth (8 out of 18), or started at an initial credit-to- GDP ratio higher than 60 percent ( 15 out of 26) ended up in crises. In practical terms, recent early warning models typically use a wide array of quantitative leading indicators of vul nerabilities, with a heavy focus on international aspects. Indicators used capture vulnerabilitie s that stem from or are centered in the external, public, financial, nonfinancial corporate, or hous ehold sectors – and combine th ese with qualitative inputs (IMF-FSB, 2010). Since international financia l markets can play multiple roles in transmitting and causing, or at least triggering, va rious types of crises, as happened recently, several international linkages measures are typi cally used. Notably banking system measures, such as exposures to international funding risks and the ratio of non-core to core liabilities, have been found to help signal vulnerabilities (Shin, 2013). 30 Since international markets can also help with risk-sharing and can reduce vola tility, and the empirical evidence is mixed, the overall relationship of international financial integration and crises is, however, much debated (Kose and ot hers, 2010; Lane, 2012). VII. C ONCLUSIO NS A Summary This paper presents a survey on financial crises to answer three specific questions. First, what are the main factors explaining financial crises? Alt hough the literatu re has clarified some of the main factors driving crises, it remains a ch allenge to definitively identify their causes. Many theories have been developed over the ye ars regarding the underl ying causes of crises. These have recognized the importance of booms in asset and credit market s that turned into busts as the main driving forces of most crises episodes. Given their central roles, the paper briefly summarizes the theoretical and empirical literature analyz ing developments in credit and asset markets around financial crises. Second, what are the major types of crises? While financial crises can take various shapes and forms, the literature has focused on four major types of crises: currency crises; sudden stop (or capital account or balance of payments) cr ises; debt crises; and banking crises. It is possible to classify crises in other ways, but regardless they can often overlap in types. A number of banking crises, for example, are also sudden stop episodes and currency crises. The paper examines the literature on the analytic al causes and empirical determinants of each type of crisis. In addition, it presents a revi ew of studies on various approaches for the identification of crises, their frequency over ti me and across different groups of countries. Third, what are the real and fina ncial sector implications of crises? Large output losses are common to many crises and other macroeconom ic variables (consumption, investment and 30 Shin (2013) compares the predictive power from pr ice-based measures (CDS and other spreads, implied volatility, Value-at-Risk, etc.), the gap of credit-to-GDP ratio from a trend, and monetary aggregates and other bank liability aggregates, and s hows that the last group has the most predictive power. 35 industrial production) typically register significant declines. Financial variables like asset prices and credit usually follow qualitativel y similar patterns across crises, albeit with variations in terms of duration and severity of declines. The paper provides a summary of the literature on the macroeconomic and fi nancial implications of crises. The paper also briefly reviews the literature on the prediction of crises. While there are many benefits in knowing whether and if so when a crisis may occur, it has been a challenge to predict crises. It is easy to document vulnerabi lities, such as increasi ng asset prices and high leverage, but it remains difficult to predict with some accuracy the timing of crises. No single set of indicators has proven to predict the vari ous types of crises. Th e paper reviews how the empirical literature on prediction of crises ha s evolved and analyzes its current state. Is this time really different? One of the main conclusions of the literature on financial crises is that it has been hard to beat the “this-time-is-different” syndrome. This, as aptly de scribed by Reinha rt and Rogoff (2009a), is the belief that “ financial crises are things that happen to other people in other countries at other times; cr ises do not happen to us, here and now. We are doing things better, we are smarter, we ha ve learned from past mistakes.” Although often preceded by similar patterns, policy makers tend to ignore the warnings and argue that: “ the current boom, unlike the many booms that preceded catast rophic collapses in the past (even in our country) is built on sound fundamentals…” Leading up to every crisis, it is often claimed that developments appear to be different from those before the earlier episodes. Before the latest episode, for example, the extensive diversif ication of risks and advanced institutional frameworks were touted as such features argu ed to justify the belief that “this time is different”. As the literature reviewed here makes abundantly clear, there are many similarities in the run-ups to crises. In the latest one, increases in credit and asset prices were common to those observed in the earlier ones. Given these co mmonalities, it should be possible to prevent crises. Yet, that seems to have been an impo ssible task. This suggests that future research should be geared to beat the “this-time-is-di fferent” syndrome. This is a very broad task requiring addressing of two ma jor questions: How to prevent financial crises? How to mitigate their costs when they take place? In a ddition, there have to be more intensive efforts to collect necessary data to guide both empirical and theoretical studies. The rest of this conclusion takes each of these issues in turn and points to future research directions. How to prevent financial crises? In light of the lessons from the latest cris is, many agree that asset price bubbles and credit booms can entail substantial costs, if they deflate rapidly. Specifi cally, many now agree on a number of issues with respect to asset price bubbles and cred it booms. First, rapid increases in asset prices and credit can lead to financial turmoil and crises with significant adverse macroeconomic effects. Second, it is important to monitor vulnerabilities stemming from such sharp increases, and determine if they could be followed by large and rapid declines (crashes, busts or crunches, cap ital outflows). Third, the subse quent busts and crunches are 36 likely to be more harmful if bubbl es arise due to “distortions.” Fourth, even if not due to distortions, evidence of irrationality can be interpreted as a sign of inefficiency and a potential source of welfare loss. As such, bubbles and credit booms can ca ll for interventions. The challenge for policy makers and researchers is twofold: when to intervene and how to intervene. First, they need to determine when (and to what extent) increases in asset prices and credit represent substantial deviations from those that can be explained by fundamentals. Second, if the behavior of cred it and asset markets suggests si gns of risk, they need to determine what would be the optimal policy re sponses to minimize risks and mitigate the adverse effects when risks materialize. There has been an active debate on if, and how, monetary policy should respond to movements in asset prices and credit. The consensus before the crisis was that the formulation of monetary policy only needed to c onsider asset prices to the extent that they were relevant for forecasting economic outlook a nd inflation, but not otherwise (see Mishkin 2008, and Kohn 2008, for reviews; and Campbell 2008 for a collection of papers). However, the crisis has made clear (again) that both fi nancial stability and ec onomic activity might be affected by asset price movements and a view has emerged that monetary policy should take into account to some degree developments in asset prices (Blanchard, Dell’Ariccia and Mauro, 2009; Bernanke, 2009 and 2011; Trichet, 2009). How to operationalize this, remains under discussion though (Eichengreen et. al 2011 ; Mishkin, 2011). While the case for policy intervention is considered stronger when the bank ing system is directly involved in financing the bubble, whereas other asset prices bubbles can more justifiably be left to themselves (Crowe et. al, 2011), the exact adjustment of monetary policy rema ins unclear (Bean, Paustian, Penalver, and Taylor, 2010; King, 2012). There remain important lessons to be le arned about the design of micro-prudential regulations and institutional stru ctures for the prevention of crises. The latest crisis has once again exposed flaws in the micro-prudential regulatory and institutional frameworks. The global nature of the crisis has also shown that financially inte grated markets have benefits, but also present risks, with the international fi nancial architecture still far from institutionally matching the policy demands of the closely-inte grated financial syst ems. Although elements of existing frameworks provide foundations, the crisis has fo rced a rethink of regulatory policies, with many open questions. While rule s calling for well capitalized and liquid banks that are transparent and adhere to sound acc ounting standards are being put in place (e.g., Basel III), clarity on how to deal with large, complex financial institutions that operate across many borders is still needed. In addition, it remains unclear what types of changes to the institutional environments – e.g., changes in the accounting standards for mark-to-market valuation, adaptations of employee compensation ru les, moves of some derivatives trading to formal exchanges, greater use of central count er parties – help best to reduce financial markets’ procyclicality and the buildup of systemic risks. The crisis has also showed that fiscal policies, both micro – such as deductibility of interest pa yments – and macro – as in the amount of resources available to deal with financial crises – can play a role in creating vulnerabilities, but which adaptations are needed is not always clear. 37 While there is also a call for th e use of macro-prudential policie s, the design of such policies and their interactions with other policies, es pecially monetary polic y, remain unclear. By constraining ex-ante financial markets participants’ be havior, macroprudential policies can reduce the impact of externalities and market failu res that lead to systemic vulnerabilities. It that way, they can reduce the risks of financ ial crises and help improve macroeconomic stability (De Nicolò and others (2012)). But th e exact design of such policies is yet to be formulated. Although it is clear that multiple tools are needed, complications are abound. Different financial distortions, for example, can l ead to different types of risks, which in turn imply the use of multiple intermediate targets. Moreover, the relevant distortions can change over time and vary by country circumstances. Ex cessive leverage among corporations may give way, for example, to excessive leverage in the household sector. Factors, such as development of financial sector and exchange rate regime, can greatly affect the types of risks economies face. Much is still unknown on these factors and implications for the formulation of macroprudential policies. As new macroprudential frameworks are being established, policymakers have also been increasingly turning their attention to the complex dynamics between macroprudential and monetary policies. These hinge importantly on the “side effects” that one policy has on the other, but conceptual models and empirical evidence on these issues are still at early st ages (see IMF (2013) for a review). The review here clearly shows that further an alytical research and empirical work on these issues are needed. Macroeconomic models need to better reflect the roles of financial intermediaries. Current models are often limite d in the way that they capture financial frictions. In terms of financial stress, they of ten assume that available instruments can fully offset financial shocks and abstract from ef fects, such as those of monetary policy on financial stability. More realistic modeling of the channels that give rise to financial instability and the actual transmission of policies and instruments is needed. In particular, the supply side of finance is not we ll understood and models with realistic calibrations reflecting periods of financial turmoil are still missing (Brunnermeier and Sanikov, 2012). The roles of liquidity and leverage in such periods have yet to be examined using models better suited to address the relevant policy questions. More insi ghts, including from em pirical studies, are necessary to help calibrate these models and allow the formulation of policy prescriptions that can be adapted to different country circumstances. Only with progress in modeling financial crises, can one hope to not only avoid some of these episodes and be prepared with better policies when they occur, but also to minimize their impacts. From an applied perspectiv e, there remains a need for better early warning models . An issue extensively discussed in policy forums and recei ving substantial attention from international organizations is the need to improve the prediction of the onset of crises (IMF, 2010) . As the review here shows, the predictive power of available models remains limited. Historical record indicates that asset price busts have been especially difficult to predict. Even the best indicator failed to raise an alarm one to th ree years ahead of roughl y one-half of all busts since 1985. This was the case again for the recent crisis. Although a number of recent papers that analyze the ability of various models in pred icting the latest crisis come to negative conclusions as well, others have found some predictive patterns. Regardless, there is scope to improve these models. 38 While known risks are being addressed, new risks can emerge. The limited ability of crises prediction models arises in part because countries do take step s to reduce vulne rabilities. In response to increased financial globalizati on and sudden stop risks, many emerging markets increased their international reserves since the late 1990s, which may have helped some countries avoid the impact of the recent cris is (De Gregorio, 2013; Kose and Prasad, 2010). Similarly, improvements in institutional enviro nments which many countries have put in place over the last decades likely helped reduc e some vulnerabilities. At the same time, however, new risks have emerged. In the latest crisis, the explosion of complex financial instruments and greater balance-sheet opa queness and reliance on wholesale funding in highly integrated global financial market s led to greater risks of a crisis. How to mitigate the costs of financial crises? It has been a challenge to explain the substan tial (real) costs associated with crises. As documented, there are various theories regardi ng the channels by which different types of crises affect the real economy. There also ex ist many descriptions of the empirical patterns around crises episodes. Yet, w hy crises cause large costs remains an enigma. Many of the channels that lead to macro-financial linka ges during normal times also “cause” the adverse effects of crises, but it is also clear that there are other dyn amics at work. Normal lending seems undermined for an extended period as ev idenced by creditless recoveries following crises. Fiscal policy and public debt dynamics can be affected for deca des, in part since governments often end up directly supporting fina ncial systems (by injecting liquidity or recapitalization) or suffer from the expansionary policies to mitigate the costs of crises. In great part, the major challenge is to expl ain the sharp, non-linear be havior of financial markets in response to “small” shocks. While the procyclicality of leverage among financial institutions, as highlighted by its increase during the run up to the 2007-09 crisis followed by the sharp deleveraging in its aftermath, has ex tensively been documen ted (Adrian and Shin, 2012), the exact causes of this behavior have yet to be identified. Why crises involve the degree of liquidity hoard ing leading to aggregate liquidity shortages and disrupt transmission of monetary policy remains a puz zle. Although credit crunches are in part attributable to capital shortages at financial institutions, th ese do not seem to fully explain the phenomena with lenders becoming overly risk-averse follow ing a crisis. This lack of knowledge of the forces shaping the dynamics before and during pe riods of financial stress greatly complicates the design of proper policy responses. It is also important to explore why financial spillovers across entities (institutions, markets, countries, etc.) are much more potent than most fundamentals suggest (in other words, why is there so much contagion?). Financial crises often generate effects across markets and have global repercussions. The latest episode is a cas e in point as its global reach and depth are without precedent in the post–World War II pe riod. This emphasizes th e value of having a better grasp of transmission mechanisms th rough which such episodes spill over to other countries. In addition to trade and cross-border banking linkages, research needs to consider the roles played by new financia l channels, such as commerci al paper conduits and shadow banking, and new trade channels, such as verti cal trade networks, in the transmission of crises across borders. Given thei r adverse impact, the exact nature of these spillovers matters 39 for the appropriate design of both crisis mitigat ion and crisis management responses. In light of their cross-border im plications, pooling (regional or gl obal) resources to provide ample liquidity proactively becomes, for example, more important as it can avoid liquidity runs escalating into self-fulfilling solvency crises and help break chains of contagion. Although many stylized facts are al ready available, work on the implications of interactions among different crises and sovere ign debt defaults is still lim ited. The review documents that various types of crises can overlap in a singl e episode, but research on the implications of such overlapping crises episode s has been lagging. Although defa ult on domestic debt tends to be less frequent than that on external debt , it still takes place quite often, suggesting the usual assumption of risk-free government debt needs to be revisited. Furthermore, there appear to be interplays between domestic and foreign debt defaults. While domestic debt tends to account for a large share of the tota l debt stock in both advanced countries and emerging markets, many emerging countries defa ult on their external debt at seemingly low thresholds of debt levels. This suggests that for a given level of unsustainable debt, the cost of defaulting on external debt appears less than that on domestic debt. More generally, there are likely tradeoffs that depend on country ci rcumstances, maybe because the risk of high inflation varies. With the rising public debt stocks in many advanced countries, more work on this would be very useful. There are still many questions about the best pol icy responses to financial crises. The global crisis and associated recessions have shown the limits of polic y measures in dealing with financial meltdowns. It has led to an extensiv e discussion about the ab ility of macroeconomic and financial sector policies to mitigate th e costs stemming from such episodes. Some research shows that countercyclical policies might mitigate the cost and reduce the duration of recessions (Kannan, Terrones, and Scott, 2013). Others argue that such policies can worsen recession outcomes (Taylor, 2009 and 2011). And some others find limited effects associated with expansionary policies (Claessens, Kose, and Terrones, 2009; Baldacci, Gupta, Mulas-Granados, 2013). Th e discussion on the potency of policies clearly indicates a fertile ground for future research as well. While there are valuable lessons on crisis resolution, countries are still fa r from adopting the “best” practices to respond to fi nancial turmoil. It is clear now that open-bank assistance without proper restructuring and recapitalization is not an efficient way of dealing with an ailing banking system (Laeven and Valenc ia, 2013; Landier and Ueda, 2013). Excessive liquidity support and guarantees of bank liabil ities cannot substitute for proper restructuring and recapitalization either as most banking cris es involve solvency problems and not only liquidity shortfalls. In the case of banking crises, the sooner re structuring is implemented, the better outcomes are. Such a strategy removes re sidual uncertainty that triggers precautionary contractions in consumption and investment, which in turn further exacerbate recessions. Still in spite of this unde rstanding, many countries do no t adopt these policy responses, including in some current cris es (Claessens et al. 2013), sugge sting that there are deeper factors that research has not be able to unco ver or address. Moreov er, issues related to restructuring of both househol d debt and sovereign debt require more sophisticated theoretical and empirical appro aches (Laeven and Laryea, 2013; Das, 2013; Igan and others, 2013). 40 What are the major needs for additional data and methods? As the review here documents, it is necessary to put together new data series and to design new methodologies to get a better understanding of crises epis odes. The review lists several recent studies that put together new data series on financial crises. In spite of these, there is clearly a case for more research to collect addi tional cross-country data on aspects relevant to financial crises. Better data on domestic debt an d house prices are urgently needed to get a richer understanding of domestic debt dynamics and fluctuations in housing markets. There is also a need for better (int ernational) data for both survei llance and early warning exercises (see Heath, 2013 and Cerutti, Claessens and McGu ire, 2013, for data needs). For a deeper understanding of crises and the policy issues surrounding these episodes, a nother need is to design new methods to classify crises in a more robust manner. Moreover, it would be important to examine periods of financial disruptions, which are not necessarily crises. Although good luck or adequate policy measures may have prevented a financial crisis following such disruption episodes, there are lessons to be lear ned since those are the types of periods that can provide ca se studies of counterfactuals to analyze the macroeconomic outcomes and implications of policy responses. 41 References Abbas, S. A., M. Belhocine, A. El Ganai ny, and M. Horton, 2011, “Historical Patterns and Dynamics of Public Debt- Evidence from a New Database,” IMF Economic Review , Vol. 59, No. 4, pp. 717-42. Abiad, A., R. Balakrishnan, P. K. Brooks, D. Leigh, and I. Ty tell, 2013, “What’s the Damage: Medium-term Output Dynamics after Financial Crises,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Abiad, A., G. Dell’Ariccia, B. Li, 2013, “W hat Have We Learned about Creditless Recoveries?,” in S. Claessens, M. A. Ko se, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Abreu, D., and M. K. Brunnermeier, 2003, “Bubbles and Crashes,” Econometrica , Vol. 71, No.1, pp. 173-204. Allen, R. E., 2010, Financial Crises and Recess ion in the Global Economy , Edward Elgar Publishing, 3rd Edition. Allen, F., A. Babus, and E. Carletti, 2009, “Financial Cris es: Theory and Evidence,” Annual Review of Financial Economics , Vol. 1, pp. 97-116. Allen, F., and D. Gale 2007, Understanding Financial Crises (Clarendon Lect. Finance), Oxford: Oxford University Press. Alessi, L., and C. Detken, 2011, “Quasi Real Time Early Warning Indicators for Costly Asset Price Boom/bust Cycles: A Ro le for Global Liquidity,” European Journal of Political Economy , Vol. 27, No. 3, pp. 520–33. Auguiar, M., and G. Gopinath, 2006, “Defaulta ble Debt, Interest Rates and the Current Account,” Journal of International Economics , Vol. 69, pp. 64-83. Baldacci, E., S. Gupta, and C. Mulas-Gran ados, 2013, “How Effective is Fiscal Policy Response in Financial Crises?,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Barberis, N., and R. Thaler, 2003, “A Su rvey of Behavioral Finance,” in Handbook of the Economics of Finance , G.M. Constantinides, M. Harri s and R. Stulz, eds. Chap.18, pp.1051-121. Barth, J., G. Capri o, and R. Levine, 2006, Rethinking Bank Regulation: Till Angels Govern , Cambridge University Press, New York, NY. 42 –––––, 2012, Guardians of Finance: Making Regulators Work for Us , The MIT Press, First Edition. Bean, C, M Paustian, A. Penalver, and T. Taylor, 2010, “Monetary Policy After the Fall,” proceedings of the Federal Reserve Bank of Kansas City Annual Conference, Jackson Hole, Wyoming. Blanchard O., G. Dell'Ariccia, P. Maur o, 2010, “Rethinking Macr oeconomic Policy,” Journal of Money, Credit and Banking , Vol. 42, No. 1, pp.199-215. Blanchard, O. J., and M. W. Watson, 1982, “Bu bbles, Rational Expectations and Speculative Markets,” in Crisis in Economic and Financial Structure: Bubbles, Bursts, and Shocks , P. Wachtel, ed. Lexington Books: Lexington. Blanco, H., and P. M. Garber, 1986, “Recurren t Devaluation and Specu lative Attacks on the Mexican Peso,” Journal of Political Economy , Vol. 94, No 1, pp. 148-66. Bordo, M. D., B. Eichengreen, D. Klingebiel, and M.S. Martinez-Peria, 2001, “Is the Crisis Problem Growing More Severe?” Economic Policy , Vol.16, pp. 51-82. Bordo, M., and J. Haubrich, 2012, “Deep Recessions , Fast Recoveries, and Financial Crises: Evidence from the American Reco rd,” NBER Working Paper 18194. Borensztein, E., and U. Panizza, 2009, “The Costs of Sovereign Default,” IMF Staff Papers , Vol. 56, No. 4, pp. 683-741. Borio, C., and P. Lowe, 2002, “Ass et Prices, Financial and Mone tary Stability: Exploring the Nexus,” BIS Working Paper, No. 114. Branch, W., and G. W. Evans, 2008, “Learning a bout Risk and Return: A Simple Model of Bubbles and Crashes,” University of Ore gon Economics Department Working Papers 2008-1. Brunnermeier, M., 2001, Asset Pricing Under Asymmetric Information: Bubbles, Crashes, Technical Analysis and Herding , Oxford University Press. Bulow, J., and K. Rogoff, 1989, “Soverei gn Debt: Is to Forgive to Forget?” American Economic Review , Vol. 79, No. 1, pp. 43-50. Burnside, C., M. Eichenbaum, and S. Rebe lo, 2001, "Prospective Defi cits and the Asian Currency Crisis," Journal of Political Economy , Vol. 109, No. 6, pp. 1155-197. –––––, 2004, “Government Guarantees and Se lf-fulfilling Speculative Attacks,” Journal of Economic Theory , Vol.119, pp. 31–63. 43 Calomiris, C. W., 2009, “The Subprime Turmoil: What’s Old, What’s New, and What’s Next,” Journal of Structured Finance , Spring 2009, Vol. 15, No. 1, pp. 6-52 . Calomiris, C, and G. Gorton, 1991, “The Orig ins of Banking Panics , Models, Facts, and Bank Regulation,” in Financial Markets an d Financial Crises , R.G. Hubbard, ed., pp.107–73. Calvo, G. A., 2000. “Betting Against the State Socially Costly Financial Engineering,” Journal of International Economics , Vol. 51, No. 1, pp. 5-19. Calvo, G. A., A. Izquierdo, and R. Loo-K ung, 2006, “Relative Price Volatility under Sudden Stops: The Relevance of Balance Sheet Effects,” Journal of International Economics , Vol. 69, 1, pp. 231-54. Calvo, G. A., A. Izquierdo, and L. Mejía, 2008 , “Systemic Sudden Stops: The Relevance of Balance-Sheet Effects and Financial In tegration,” NBER Working Papers 14026, National Bureau of Economic Research, Inc. –––––, 2004, “On the Empirics of Sudden Stops: Th e Relevance of Balance-Sheet Effects,” Proceedings , Federal Reserve Bank of San Francisco. Calvo, G. A., A. Izquierdo, and E. Talvi, 200 6, “Phoenix Miracles in Emerging Markets: Recovering without Credit from Systemic Fi nancial Crises,” NBER Working Papers 12101, National Bureau of Economic Research. Calvo, G., and C. Reinhart,1999, “Capital Flow Re versals, the Exchange Rate Debate, and Dollarization,” Finance and Development , International Monetary Fund, September. –––––, 2000, “When Capital Inflows Come to a Sudden Stop: Consequences and Policy Options,” in P. B. Kenen and A. K. Swoboda, eds., Reforming the International Monetary and Financial System , Washington, DC, International Monetary Fund. Campbell, J. (ed.), 2008, Asset Prices and Monetary Policy , NBER Conference Report. University of Chicago Press Campello, M., J. Graham, and C. R. Harvey, 2010. "The real effects of financial constraints: Evidence from a financial crisis," Journal of Financial Economics , Vol. 97, No. 3, pp. 470-87. Cardarelli, R., M. A. Kose, and S. Ele kdag, 2010, “Capital Inflows: Macroeconomic Implications and Policy Responses,” Economic Systems , Vol. 34, No. 4. Cardarelli, R., S. Elekdag, and Subir La ll, 2009, Financial Stress, Downturns, and Recoveries, IMF Working Pa per No. 09/100, forthcoming Journal of Financial Stability . 44 Cerutti, E., S. Claessens, and P. McGuire, 2013, “Systemic Risks in Global Banking: What Available Data Can Tell Us and What More Data Are Needed?” in Systemic Risk and Macro Modeling , M. K. Brunnermeier and A. Kr ishnamurthy, eds., forthcoming, NBER/University of Chicago Press. Chang, R., and A. Velasco, 2000, “Liquidity Crises in Emerging Markets: Theory and Policy,” Chapter in: NBER Macroeconomics Annual 1999 , Vol. 14, p. 11-78. Chen, J., H. Hong, and J. C. Stein, 2002, “Breadth of Ownership and Stock Returns,” Journal of Financial Economics , Vol. 66, No. 2-3, pp. 171-205. Čihák, M., A. Demirgüç-Kunt, M. Soledad Ma rtínez Pería, and A. Mohseni-Cheraghlou, 2012, “Bank Regulation and Supervision around the World: A Crisis Update,” World Bank Policy and Research Working Paper, No.6286. Claessens, S., “Balance of Payments Cr ises in an Optima l Portfolio Model,” European Economic Review , 1991, 35, pp. 81-101. –––––, 2009, “Competition in the Financial Sector : Overview of Competition Policies,” World Bank Research Observer , Vol. 24, pp. 83-118. Claessens, S., C. Pazarbasioglu, M. Dobler, F. Valencia, O. Nedelescu, and K. Seal, 2013, “Crisis Management and Resolution: Early Le ssons from the Financial Crisis,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Claessens, S., G. Dell’Ariccia, D. Igan, and L. Laeven, 2010, “Cross-Country Experience and Policy Implications from the Global Financial Crisis,” Economic Policy . A European Forum, April 2010, Vol. 62. PP. 269-93. –––––, 2012, “A Cross-Country Persp ective on the Causes of the Gl obal Financial Crisis,” in Handbooks in Financial Globalization: Th e Evidence and Im pact of Financial Globalization , pp. 737-52. Elsevier UK: Academic Press. Claessens, S., M. A. Kose, and M. Terro nes, 2009, “What Happens During Recessions, Crunches and Busts?” Economic Policy , Vol. 60, pp. 653-700. –––––, 2010a, “Financial Cycles: What? How? When?” in NBER International Seminar in Macroeconomics 2010, Richard Clarida and Francesco Giavazzi, eds. Chicago: University of Chicago Press. –––––, 2010b, “The Global Financial Crisis: How Si milar? How Different? How Costly?” Journal of Asian Economics , Vol. 21, No. 3, pp. 247-64. 45 –––––, 2012, “How do Business and Financial Cycles Interact?” Journal of International Economics , Vol. 87, pp. 178-90. Corsetti, G., P. Pesenti, and N. Roubini, 1998, “P aper Tigers? A Model of the Asian Crises,” European Economic Review , Vol. 43, No. 7, pp. 1211-236. Crespo-Cuaresma, J., and T. Slacik, 2009, “On the Determinants of Currency Crisis: The Role of Model Uncertainty,” Journal of Macroeconomics , Vol. 31, pp. 621–32. Crowe, C., G. Dell’Ariccia, D. Igan, and P. Rabanal, 2013, “Policies for Macrofinancial Stability: Managing Real Estate Booms and Bu sts,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Cruces, J. J., and C. Trebesch, 2011, “Sovere ign Defaults: The Price of Haircuts,” (unpublished; Berlin, Germany: Free University of Berlin). Cumby, R. E., and S. van Wijnbergen, 1989, “Fin ancial Policy and Speculative Runs with a Crawling Peg: Argentina 1979-1981,” Journal of International Economics , Vol.27, pp.111-27. Das, U. S., 2013, “Restructuring Sovereign De bt: Lessons from Recent History,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Das, U. S., M. G. Papaioannou, and C. Treb esch, 2012, “Sovereign Debt Restructurings 1950–2010: Literature Survey, Data, and St ylized Facts,” IMF Working Paper No. 12/203 (Washington: Intern ational Monetary Fund). De Gregorio, J., 2013, “Resilience in Latin America: Lessons from Macroeconomic Management and Financial Policies,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. De la Torre, A., and A. Ize, 2011, “Containing Systemic Risk: Paradigm-based Perspectives on Regulatory Reform,” Policy Research Working Paper Series 5523, The World Bank. Dell'Ariccia, G., E. Detragiache, Enrica, a nd R. Rajan, 2008, “The Real Effect of Banking Crises,” Journal of Financial Intermediation , Vol. 17, pp. 89-112. Dell’Ariccia, G., D. Igan, and L. Laeve n, 2012, “Credit Booms and Lending Standards: Evidence from the U.S. Subprime Mortgage Market,” Journal of Mone y, Credit and Banking , Vol. 44, pages 367-84. 46 Dell’Ariccia, G., D. Igan, L. Laeven, a nd H. Tong, 2013, “Policies for Macrofinancial Stability: Dealing with Credit Booms and Bu sts,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Dell’Ariccia, G., L. Laeven, and R. Marquez, 2010, “Monetary Policy, Leverage, and Bank Risk-taking,” IMF Working Papers 10 /276, International Monetary Fund. De Nicolò, G., G. Dell’Ariccia, L. Laeve n, and F. Valencia, 2010, “Monetary Policy and Bank Risk-Taking,” IMF Staff Position Note, SDN/10/09. Demirgüç-Kunt, A., E. J. Kane, and L. L aeven, 2008, “Determinants of Deposit-insurance Adoption and Design,” Journal of Financial Intermediation , Vol. 17, No. 3, pp. 407-38. Dewatripont, M., and J. Tirole, 1994, The Prudential Regulation of Banks , MIT Press. Diamond, D., and R. Rajan, 2001, “Banks, Short-te rm Debt, and Financial Crises: Theory, Policy Implications and Applications,” Carnegie-Rochester Conference Series, Vol. 54, pp. 37-71. Diamond, D., and P. Dybvig, 1983, “Bank Runs , Deposit Insurance, and Liquidity,” Journal of Political Economics , Vol. 91, pp.401–19. Diether, K., C. J. Malloy, and A. Scherbina, 2002, “Differences of Opinion and the Cross- Section of Stock Returns,” Journal of Finance , Vol. 57, No. 5. Eaton, J., and R. Fernandez, 1995, “Sovereign Debt,” in the Handbook of International Economics , G.M. Grossman and K. Rogoff, eds., Vol. 3, Chapter 3, pp. 2031-077. Eaton, J., and M. Gersovitz, 1981, “Debt with Potential Repudiation: Theoretical and Empirical Analysis,” The Review of Economic Studies , Vol. 48, No. 2, pp. 289-309. Edwards, S., 1984, “The Demand for Internati onal Reserves and Monetary Equilibrium: Some Evidence from Developing Countries,” The Review of Economics and Statistics , MIT Press, Vol. 66, No. 3, pp. 495-500. Eichengreen, B., 2002, Financial Crises: And What To Do About Them , Oxford University Press. Eichengreen, B., A. K. Rose, and C. Wyplosz, 1996, “Speculative Attacks on Pegged Exchange Rates: An Empirical Exploration with Special Reference to the European Monetary System,” in M.B. Canzoneri , P.R. Masson, and V.U. Grilli, eds., Transatlantic economic issues (Cambridge University Press, Cambridge, UK). 47 Eichengreen, B., and R. Hausmann, 1999, “Exchange Rates and Financial Fragility,” Proceedings , Federal Reserve Bank of Kansas City, pages 329-68. Evanoff, D. D., G. G. Ka ufman, A. G. Malliaris, 2012, New Perspectives on Asset Price Bubbles , Oxford University Press. Fama, E. F., 1998, “Market Efficiency, Long-t erm Returns, and Behavioral Finance,” Journal of Financial Economics , Vol. 49, pp. 283-306. Farhi, E., and J. Tirole, 2012, “Collective Mora l Hazard, Maturity Mismatch, and Systemic Bailouts,” American Economic Review , Vol. 102, No. 1, pp. 60-93. Feldstein, M. 2009, “Rethinking the Role of Fiscal Policy,” American Economic Review , Vol. 99, No. 2, pp. 556-59. Fisher, I., 1933, “The Debt Deflati on Theory of Great Depressions” Econometrica , Vol 1, pp. 337-57. Flood, R., and N. P. Marion, 1997, “Policy Imp lications of "Second- Generation" Crisis Models,” IMF Staff Papers , Vol.44, No.3, pp.383-90. Flood, R., and P. Garber, 1984, “Collapsing Exchange-Rate Regimes: Some Linear Examples,” Journal of International Economics , Vol.17, pp. 1-13 Forbes, K. J. 2012, “The Big “C”: Identifying and Mitigating Contagion,” Proceedings 2012 Jackson Hole Symposium, hosted by the Federal Reserve Bank of Kansas City. Forbes, K. J., and F. Warnock, 2012, “Capital Flow Waves: Surges, Stops, Flight, and Retrenchment,” Journal of International Economics , Vol. 88, No. 2, pp. 235-51. Frankel, J. A., and A. K. Rose, 1996, “Currenc y crashes in emerging markets: An empirical treatment,” Journal of International Economics , Vol. 41, Nos. 3-4, pp. 351-66. Frankel, J., and G. Saravelos, 2012, “Can Lead ing Indicators Assess Country Vulnerability? Evidence from the 2008–09 Global Financial Crisis,” Journal of International Economics , Vol. 87, pp. 216–31. Furceri, D., and A. Zdzienicka, A., 2012, “How Costly Are Debt Crises?” Journal of International Money and Finance , Vol. 31. No. 4, pp. 726–42. Garber, P. M., 2000, Famous First Bubbles: The Fundamentals of Early Manias , Cambridge, MA: MIT Press. Glick R., and M. Hutchis on, 2012, Currency Crises, in Elsevier Encyclopedia on Financial Globalization . 48 Goldstein, M., C. Reinha rt, and G. Kaminsky, 2000, Assessing financial vulnerability, an Early Warning System for Emerging Markets , Peterson Institute for International Economics Press, No. 100. Gorton, G., 1988, “Banking Panics and Business Cycles,” Oxford Economics Papers , Vol. 40, pp.751–81. –––––, 2008, “The Panic of 2007,” in Maintaining Stability in a Changing Financial System, Proceedings of the 2008 Jackson Hole Conference , Federal Reserve Bank of Kansas City. –––––, 2009, “Slapped in the Face by the Invisibl e Hand: Banking and the Panic of 2007,” Paper prepared for the Federal Reserve Bank of Atlanta’s 2009 Financial Markets Conference: Financial Innova tion and Crisis, May 11-13. –––––, 2012, Misunderstanding Financia l Crises: Why We Don’t See Them Coming , Oxford University Press. Gorton G. and G. Ordonez, 2012, “Collatera l Crises,” NBER Working Papers, No. 17771, National Bureau of Economic Research, Inc. Gourinchas, P.-O., R. Valdés, and O. Lande rretche, 2001, “Lending Booms: Latin America and the World,” Economia , Vol. 1, pp. 47-99. Gupta, P., D. Mishra, and R. Sahay, 2007, “Beh avior of Output During Currency Crises,” Journal of International Economics , Vol. 72, No. 2, pp. 428-50. Gurkaynak, R. S., 2008, “Econometric Tests of Asset Price Bubbles: Taking Stock,” Journal of Economic Surveys , Vol. 22, No. 1, pp. 166-86. Haber, S., 2005, “Banking and Economic Growth in Mexico,” Working Paper, Hoover Institution. Harrison, J. M., and D. M. Kreps, 1978, “Specula tive Investor Behavior in a Stock Market with Heterogeneous Expectations,” The Quarterly Journal of Economics , Vol. 92. No. 2, pp. 323-36. Heath, Robert, 2013, “Why are the G-20 Data Gaps Initiative and the SDDS Plus Relevant for Financial Stability Analysis?” IMF Wo rking Paper, No. 13/6, Washington, D.C. Hirata, H., M. A. Kose, C. Otrok, and M. E. Terrones, 2012, “Global House Price Fluctuations: Synchronization and Determ inants,” in Francesco Giavazzi and Kenneth West, eds. NBER International Seminar on Macroeconomics 2012 , National Bureau of Economic Research. 49 Hong, H., J. Scheinkman, and W. Xiong, 2008, “Advi sors and Asset Prices: A Model of the Origins of Bubbles,” Journal of Financial Economics , Vol. 89, No. 2, pp. 268-87. Honohan, P. and L. Laeven (eds.), 2007, Systemic Financial Crises , Cambridge Books, Cambridge University Press, Howard, G., R. Martin, B. A. Wilson, 2011, “A re Recoveries from Banking and Financial Crises Really So Different?” Board of Governors of the Fede ral Reserve System, International Finance Discussion Pa pers, Working Paper Number 1037. Igan, D., D. Leigh, J. Simon, and P. Topalova , 2013, “Dealing with Household Debt,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. IMF, 2010, “The IMF-FSB Early Warning Exerci se: Design and Meth odological Toolkit,” IMF, available at www.imf.org/ external/np/pp/eng/2010/090110.pdf. IMF, 2013, “The Interaction of Monetary a nd Macroprudential Policies,” Board Paper, forthcoming. Jeanne, O., 2003, “Why do Emerging Economies Borrow in Foreign Currency?” IMF Working Papers, No. 03/177, Inte rnational Monetary Fund. Jorda, O., M. Schularick, and A. M. Tayl or, 2012, “When Credit Bites Back: Leverage, Business Cycles, and Crises,” San Fr ancisco Fed, Working Paper No. 2011-27. Kaminsky, G. 2003, “Varieties of Currency Crises,” NBER Working Papers, No. 10193, National Bureau of Economic Research, Inc. Kaminsky, G., S. Lizondo, and C. Reinhart, 1998, “Leading Indi cators of Currency Crisis,” IMF Staff Papers , Vol. 45, No. 1. Kaminsky, G., and C. Reinhart, 1999, “The Tw in Crises: The Causes of Banking and Balance-of-Payments Problems,” American Economic Review , Vol. 89, pp. 473–500. –––––, 2001, “Bank Lending and Contagion: Evid ence from the Asian Crisis,” NBER Chapters, in: Regional and Global Capital Flows: Macroeconomics Causes and Consequences , Vol. 10, pp. 73-116. Kannan, P., 2009, “On the Welfare Benefits of an Interna tional Currency,” European Economic Review , Vol. 53, No. 5, pp. 588-606. Kannan, P., A. Scott, and M. E. Terrones, 2013, “From Recession to Recovery: How Soon and How Strong,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. 50 Keynes, J. M., 1930, The Great Slump of 1930 . London: The Nation & Athenæum, (First Edition). Kindleberger, C., 1978, Manias, Panics, and Crashes: A History of Financial Crises , New York: Basic Books –––––, 1986, The World in Depression: 1929-1939 , Berkeley: University of California Press. Klein, M. W., and N. P. Marion, 1994, “Explaini ng the Duration of Exchange-Rate Pegs,” NBER Working Papers 4651, National Bu reau of Economic Research, Inc. Kletzer, K. M., and B. D. Wright, 2000, “S overeign Debt as Intertemporal Barter,” American Economic Review , Vol. 90, No. 3, pp. 621-39. Klingebiel, D., R. Kroszner, and L. Laeven, 2007. "Banking Crises, Financial Dependence, and Growth," Journal of Financial Economics , Vol. 84, No. 1, pp. 187-228. Kose, M. A., P. Loungani, M. Terrones, 2013, “Global Recessions and Recoveries,” IMF Working Paper, forthcoming. Kose, M. A., and E. Prasad, 2010, Emerging Markets: Resilience and Growth Amid Global Turmoil , Brookings Institute. Kose, M. A., E. Prasad, K. Rogoff, and S. -J. Wei, 2010, “Financial Globalization and Economic Policies,” in D. Rodrik and M. Rosenzweig, eds., Handbook of Development Economics Vol. 5, pp. 4283–362. Krugman, P., 1979, "A Model of Ba lance-of-Payments Crises," Journal of Money, Credit and Banking , Vol. 11, No. 3, pp. 311-25. Krugman, P., 1999, “Balance Sheets, the Tran sfer Problem, and Financial Crises,” International Tax and Public Finance , Vol. 6, No. 4, pp. 459-72. La Porta, R., F. Lopez-de-Silanes, A. Shle ifer, and R. Vishny, 2000, “Investor Protection and Corporate Governance,” Journal of Financial Economics , Vol. 58, Nos. 1-2, pp. 3-27. Laeven, L., 2001, “Insider Lending and Ba nk Ownership: The Case of Russia,” Journal of Comparative Economics , Vol. 29, No. 2, pp. 207-29. –––––, 2011, "Banking Crises: A Review," Annual Review of Financial Economics , Annual Reviews , Vol. 3, No. 1, pp. 17-40. Laeven, L., and T. Laryea, 2013, “Principles of Household Debt Restructuring,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. 51 Laeven, L., and F. Valencia, 2008, “Syste mic Banking Crises: A New Database,” IMF Working Papers 08/224, Inte rnational Monetary Fund. –––––, 2013a, “Systemic Banking Crises” in S. Clae ssens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. (also issued as “Systemic Banking Crises Database: An Update,” IMF Working Paper, No.12/163.) –––––, 2013b, “Resolution of Banking Crises: The Good, the Bad, and the Ugly,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. (also issued as “Resolution of Banking Crises: The Good, the Bad, and the Ugly,” IMF Working Paper, No. 10/146.) Lamont, O., and A. R. H. Thaler, 2003, “Can the Market Add and S ubtract? Mispricing in Tech Stock Carve-outs,” Journal of Political Economy , Vol. 111, No. 2, pp. 227-68. Landier, A., and K. Ueda, 2013, “The Economics of Bank Restructuring: Understanding the Options,” in S. Claessens, M. A. Kose , L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Lane, P. R., 2012, “The European Sovereign Debt Crisis,” Journal of Economic Perspectives , Vol. 26, No. 3, pp. 49-68. Lane, P. R., and G-M. Milesi-Ferretti, 2011, “T he Cross-Country Incidence of the Global Crisis,” IMF Economic Review , Vol. 59, No. 1, pp. 77-110. Lansing, K. J., 2008, “Speculative Growth and Overreaction to Technology Shocks,” Working Paper Series 2008-08, Federa l Reserve Bank of San Francisco. Levy-Yeyati, E. L. and U, Panizza, 2011, “The Elusive Costs of Sovereign Defaults,” Journal of Development Economics , Vol. 94(1), pp. 95–105. Lindgren C.-J., G. Garcia, and M. Saal (eds.), 1996, Bank Soundness and Macroeconomic Policy , International Moneta ry Fund, Washington. Lo, A. W., and A. C. MacKinlay, 2001, A Non-Random Walk Down Wall Street , Princeton University Press. Lopez-Salido, D., and E. Nelson, 2010, “Postwar Financial Crises a nd Economic Recoveries in the United States,” Working Paper, Federal Reserve Board. Maddaloni, A., and J-L. Peydr ó, 2010, “Bank Risk-taking, S ecuritization, Supervision and Low Interest Rates: Evidence from the Eu ro Area and the U.S. Lending Standards,” Working Paper Series 1248, European Central Bank. 52 Magud, N. E., C. M. Reinhart, and E. R. Ve speroni, 2012, “Capital Inflows, Exchange Rate Flexibility, and Credit Booms,” IMF Working Paper, No. 12/41. McKinnon, R, and H. Pill, 1996, “Credible Liberali zations and International Capital Flows: The Overborrowing Syndrome,” in T. Ito, and A. Krueger, eds., Financial Deregulation and Integration in East Asia , pp.7-42 (University of Chicago Press). –––––, 1998, “International Overborrowing: A Decomposition of Credit and Currency Risks,” World Development , Vol. 26, No. 7, pp. 1267-282. Mendoza, E. G., 2010, “Sudden Stops, Fi nancial Crises, and Leverage,” American Economic Review , Vol. 100, No. 5, pp. 1941-966. Mendoza, E. G., and M. E. Terrones, 2008, “A n Anatomy of Credit Booms: Evidence From Macro Aggregates and Micro Da ta,” NBER Working Papers 14049. –––––, 2012, “An Anatomy of Credit Booms and Their Demise,” Journal of Economia Chilena , Central Bank of Chile, Vol. 15, pp. 4-32. Milesi-Ferretti, G-M., and C. Tille, 2011, “The Great Retrenchment: International Capital Flows During the Global Financial Crisis,” Economic Policy , Vol. 26, No. 66, pp. 285-342. Milesi-Ferretti, G-M., and A. Razin, 2000, “Curre nt Account Reversals and Currency Crises, Empirical Regularities,” NBER Chapters, in Currency Crises , pp. 285-326, National Bureau of Economic Research. Miller, E.M., 1977, “Risk, Uncertainty and Divergence of Opinion,” Journal of Finance , Vol. 32, pp. 1151-61. Minsky, H. P., 1975, John Maynard Keynes , Columbia University Press. –––––, 1977, “A Theory of Systemic Fragility,” in E.J. Altman and A. W. Sametz, eds., Financial Crises: Institutions and Ma rkets in a Fragile Environment , Wiley, pp.138-52. Mishkin, F. S., 2008, “Does Stabilizing Infla tion Contribute To Stabilizing Economic Activity?” NBER Working Papers 13970, Na tional Bureau of Economic Research, Inc. –––––, 2011, “Monetary Policy Strategi es: Lessons from the Crisis ,” NBER Working Papers 16755, National Bureau of Economic Research. 53 Obstfeld, M., 2012, “Financial Flows, Fina ncial Crises, and Global Imbalances,” Journal of International Money and Finance , Vol. 31, No. 3, pp. 469-80. Obstfeld, M., and K. Rogoff, 1986, “Ruli ng Out Divergent Speculative Bubbles,” Journal of Monetary Economics , Vol. 17, No. 3, pp. 349-62. –––––, 2009, “Global Imbalances and the Financia l Crisis: Products of Common Causes,” CEPR Discussion Papers, No. 7606. Ofek, E., and M. Richardson, 2003, “DotCom Mani a: The Rise and Fall of Internet Stock Prices,” Journal of Finance , Vol. 58, pp. 1113-137. Ongena, S., I. Vasso, and J. L. Peydró, 2009, Monetary policy, risk-taking and pricing: Evidence from a quasi-natural experi ment. Discussion Paper 2009-31, Tilburg University, Center for Economic Research. Panizza, U., F. Sturzenegger, and J. Zettelmeyer, 2009, “The Economics and Law of Sovereign Debt and Default ,”Journal of Economic Literature , Vol. 47, No. 3, pp. 651-98. Papell, D. H., and R. Prudan, 2011, “The Statisti cal Behavior of GDP after Financial Crises and Severe Recessions,” Pape r prepared for the Federal Reserve Bank of Boston conference on “Long-Term Effects of the Great Recession,” October 18-19. Pastor, L., and P. Veronesi, 2006, “Was Ther e a Nasdaq Bubble in the Late 1990s?” Journal of Financial Economics , Vol. 81, No. 1, pp. 61-100. Radelet, S., and J. Sachs. 1998, “The East Asian Financial Crisis: Diagnosis, Remedies, Prospects,” Brookings Papers on Economic Activity , Vol. 1, pp. 1-74. Rajan, R. G., 2005, “Has Financial Deve lopment Made the World Riskier?” Proceedings , Federal Reserve Bank of Kansas City, August, pp. 313-69. Rancière, R., and A. Tornell, 2011, “Financial Black-Holes: The Interaction of Financial Regulation and Bailout Guarantees,” CE PR Discussion Papers 8248, C.E.P.R. Discussion Papers. Reinhart, C. M., and K. S. Rogoff, 2008, “Is the 2007 U.S. Subprime Cris is So Different? An International Historical Comparison,” American Economic Review , Vol. 98, No. 2, pp.339–44. –––––, 2009a, This Time is Different: Eight Centuries of Financial Folly , Princeton Press. –––––, 2009b, “The Aftermath of Financial Crises,” American Economic Review , Vol. 99, pp. 466–72. 54 –––––, 2011, “From Financial Crash to Debt Crisis,” American Economic Review . Vol. 101, No. 5, pp. 1676–706. –––––, 2013, “Financial and Sovereign Debt Crises : Some Lessons and Those Forgotten,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Reinhart, C., J. Kirkegaard, and B. Sb rancia, 2011, “Financial Repression Redux,” IMF Finance and Development , Vol. 48, No. 1. Rose, A. K., and M. M. Spiegel., 2011, “Cro ss-Country Causes and Consequences of the 2008 Crisis: Early Warning.” European Economic Review , Vol. 55, No. 3, pp. 309-24. Salant, S., 1983, “The Vulnerability of Price St abilization Schemes to Speculative Attack,” Journal of Political Economy , Vol. 91, pp.1-38. Salant, S., and D. Henderson, 1978, “Market Antic ipations of Governme nt Policies and the Price of Gold,” Journal of Political Economy , Vol. 86, pp. 627-48. Scheinkman, J. A., and W. Xiong, 2003, “Ove rconfidence and Specu lative Bubbles,” Journal of Political Economy , Vol. 111, No. 6, pp. 1183-219. Schularick, M., and A. M. Taylor, 2012, “Cre dit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870-2008,” American Economic Review , Vol. 102, No. 2, pp. 1029-061. Schwert, G. W., 2003, “Anomalies and Market Efficiency,” in Handbook of the Economics of Finance , G.M. Constantinides, M. Harris, and R. M. Stulz (ed.), Vol. 1, chapter 15, pp. 939-74. Shin, H. S., 2009, “Reflections on Northern Ro ck: The Bank Run that Heralded the Global Financial Crisis,” Journal of Economic Perspectives , Vol. 23, No. 1, pp. 101-19. –––––, 2013, “Procyclicality and the Search for Earl y Warning Indicators,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Shleifer, A., 2000, Inefficient Markets: An In troduction to Behavioral Finance , Oxford University Press. Shleifer, A., and R.W. Vishny, 1997, “The Limits of Arbitrage,” Journal of Finance , Vol. 52, No. 1, pp. 35-55. 55 Stiglitz, J., and A. Weiss, 1981. "Credit Rationi ng in Markets with Imperfect Information," American Economic Review , Vol. 71, No. 3, pp. 393-410. Sturzenegger, F., and J. Zettelmeyer, 2007, “D ebt Defaults and Lessons from a Decade of Crises,” Cambridge and London: MIT Press. Taylor, A. M., 2013, “External Imbalances and Fi nancial Crises,” in S. Claessens, M. A. Kose, L. Laeven, and F. Valencia, eds., Financial Crises, Consequences, and Policy Responses , forthcoming. Taylor, J. B., 2009, “The Financ ial Crisis and the Policy Res ponses: An Empirical Analysis of What Went Wrong,” NBER Working Paper, No. 14631. –––––, 2011, “An Empirical Analysis of the Revi val of Fiscal Activism in the 2000s,” Journal of Economic Literature , Vol. 49, No. 3, pp. 686-702. Tirole, J., 2002, Financial Crises, Liquidity, and the International Monetary System , Princeton University Press. Tomz, M., and M. L. J. Wright, 2007, “Do Countries Default in "Bad Times"?” Journal of the European Economic Association , Vol. 5, No. 2-3, pp. 352-60. Wang, P., and Y. Wen, 2010, “Speculative Bubbles and Financial Crises,” American Economic Journal: Macroeconomics , Vol. 4, No. 3, pp. 184-221. Wermers, R., 2012, “Runs on Money Market Mutual Funds,” Working Paper, University of Maryland. 56 Figure 1. Evolution of House Prices During Financial Crises Notes: Real house price index is equal to 100 five years prior to the banking crises. Big 5 refers to the average of indices for the five major banking crises (Spain - 1977, Norway - 1987, Finland -1991, Sweden - 1991, and Japan - 1992). For the current crisis, the beginning date is assumed to be 2007Q3. House price series for the US is the S&P Case-Shiller National Home Price Index.Sources : BIS, OECD, and Haver Analytics.80100120140160180 -20 -16 -12 -8 -4 0 4 8US Current Big 5 France Current UK Current Spain Current Ireland Current New Zealand Current Sweden Current 57 Figure 2.A. Credit and Asset Price Booms A. Duration B. Amplitude C. Slope Notes : Amplitude and slope correspond to sample median and duration corresponds to sample mean. Duration is the time it takes to attain the level at the previous peak after the trough. Amplitude is calculated based on the one year change in each respective variable after the trough. Slope is the amplitude from peak to trough divided by the duration. Booms are the top 25 percent of upturns calculated by the amplitude. *** indicate that the difference between corresponding financial boom and other upturns is statistically significant at 1 percent level. The sample includes data for 23 advanced countries and covers 1960-2011. 051015202530 Credit House Price Equity PriceBooms Other Upturns ****** 0102030405060 Credit House Price Equity PriceBooms Other Upturns *** ****** 02468101214 Credit House Price Equity PriceBooms Other Upturns *** ****** 58 Notes : Amplitude and slope correspond to sample median and duration corresponds to sample mean. Duration is the number of quarters between peak and trough. Amplitude is calculated based on the decline in each respective variable during the downturn. Slope is the amplitude from peak to trough divided by the duration. Crunches and busts are the worst 25 percent of downturns calculated by the amplitude. ***, ** indicate that the difference between the corresponding disruptions and other downturns is statistically significant at 1 and 5 percent level, respectively. The sample includes data for 23 advanced countries and covers 1960-2011. Figure 2.B. Credit Crunches and Asset Price Busts A. Duration C. SlopeB. Amplitude024681012141618 Credit House Price Equity PriceDisruptions Other Downturns*** *** -60-50-40-30-20-100 Credit House Price Equity PriceDisruptions Other Downturns*** *** *** -7-6-5-4-3-2-10 Credit House Price Equity PriceDisruptions Other Downturns***** ****** 59 Figure 3. Coincidence of Financial Booms and Crises Notes: The sample consists of 40 countries. The numbers, except in the last column show the percent of the cases in which a crisis or poor macroeconomic performance happened after a boom was observed (out of the total number of cases where the boom occurred). Source : Dell'Ariccia et. al (2011)(fraction of total, in percent) 020406080100 Credit House Prices Both Neitherfollowed by financial crisis followed by poor performance followed by financial crisis or poor performance 60 Banking Crises (147)Currency Crises (217)Banking Crises (147)Debt Crises (67) 74 53 122 85 16 44 18 4 22 4 4 23 23 170 Banking Crises (147)Currency Crises (217)Currency Crises (217)Debt Crises (67) 47 54 133 151 36 24 17 6 29 13 24 1 160 188 Sources : The dates of banking, currency, and debt crises are from Laeven and Valencia (2008, 2011) and the dates of sudden stops are from Forbes and Warnock (2011). Figure 4. Coincidence of Financial Crises Debt Crises (67) Sudden Stops (219) Sudden Stops (219) Sudden Stops (219) Notes: A financial crisis starting at time T coincides with another financial crisis if the latter starts at any time between T-3 and T+3. A financial crisis starting at time T coincides with two other financial crisis if the latter two start at any time between T-3 and T+3. The sample consists of 181 countries. 61 Figure 5. Average Number of Financial Crises over Decades Notes: This graph shows the average number of financial crises in respective decades. Sources : The dates of banking, currency, and debt crises are from Laeven and Valencia (2008, 2011) and the dates of sudden stops are from Forbes and Warnock (2011). 012345678910 1970s 1980s 1990s 2000sCurrency Crises Banking Crises Debt Crises Sudden Stops 62 Figure 6. Coincidence of Recessions and Crises Notes: A recession is associated with a financial crisis if the financial crisis starts at the same time with the recession or one year before or two years after the peak of the recession. The sample includes data for 23 advanced countries and 38 emerging market countries, and covers 1960-2011. ( number of events) 050100150200250300350 World Advanced Countries Emerging Market CountriesAll Recessions Recessions with Crises 63 Figure 7. Real Implications of Financial Crises, Crunches, and Busts Notes : For "Duration" means are shown, for "Cumulative Loss" and "Amplitude" medians are shown. Amlitude is calculated based on the decline in output from peak to trough of a recession, duration is the number of quarters between peak and trough, and cumulative loss combines information about the duration and amplitude to measure overall cost of a recession and is expressed i n percent. Disruptions (severe disruptions) are the worst 25% (12.5%) of downturns calculated by amplitude. A recession is associated with a (severe) credit crunch or a house price bust if the (severe) credit crunch or the house price bust starts at the same time or one quarter before the peak of the recession. A recession is associated with a financial crisis if the financial c risis starts at the same time of the recession or one year before or two years after the peak of the recession. The severe financial crises are the worst 50% of financial crises as measured by output decline during the recession. The sample includes data for 23 advanced countries and covers 1960-2011.A. Duration C. Cumulative LossB. Amplitude -10-8-6-4-20 Financial Crises Credit Crunches House Price Bustswithout with with severe0246 Financial Crises Credit Crunches House Price Bustswithout with with severe -5-4-3-2-10 Financial Crises Credit Crunches House Price Bustswithout with with severe 64 Figure 8. Financial Implications of Crises, Crunches, and Busts Notes: Each panel shows the median change in respective variable during recessions associated with indicated financial events. Disruptions (severe disruptions) are the worst 25% (12.5%) of downturns calculated by amplitude. A recession is associated with a (severe) credit crunch or a house price bust if the (severe) credit crunch or house price bust starts at the same time or one quarter before the peak of the recession. A recession is associated with a financial crisis if the crisis starts at the same time of the recession or one year before or two years after the output peak preceding the recession. Severe financial crises are the worst 50% of financial crises as measured by output decline during the recession. The sample includes data for 23 advanced countries and covers 1960-2011.B. Equity PricesA. House Prices -15-12-9-6-30 Financial Crises Credit Crunches House Price Bustswithout with with severe -45-30-15015 Financial Crises Credit Crunches House Price Bustswithout with with severe 65 Figure 9. Creditless Recoveries (Percent change from a year earlier; zero denotes peak; x-axis quarter) -10-5051015 -12 -8 -4 0 4 8 12Credit Notes: Each panel shows the median year-over-year growth rate of the respective variable during recessions associated with credit crunches. Zero is the quarter at which a recession with credit crunch begins. The sample includes data for 23 advanced countries and covers 1960-2011.-4-202468 -12 -8 -4 0 4 8 12Output -15-10-5051015 - 1 2 - 8 - 4 048 1 2House Prices