Nội dung

The Credit Crunch of 2007-2008: A Discussion of the Background, Market Reactions, and Policy Responses Paul Mizen This paper discusses the events surrounding the 2007-08 credit crunch. It highlights the period of exceptional macrostability, the global savings glut, and financial innovation in mortgage-backed securities as the precursors to the crisis. The credit crunch itself occurred when house prices fell and subprime mortgage defaults increased. These events caused investors to reappraise the risks of high-yielding securities, bank failures, and sharp increases in the spreads on funds in interbank markets. The paper evaluates the actions of the authorities that provided liquidity to the markets and failing banks and indicates areas where improvements could be made. Similarly, it examines the regulation and supervision during this time and argues the need for changes to avoid future crises. (JEL E44, G21, G24, G28) Federal Reserve Bank of St. Louis Review, September/October 2008, 90(5), pp. 531-67. T he concept of a “credit crunch” has a long history reaching as far back as the Great Depression of the 1930s.1 Ben Bernanke and Cara Lown’s (1991) classic article on the credit crunch in the Brookings Papers documents the decline in the supply of credit for the 1990-91 recession, controlling for the stage of the business cycle, but also considers five previous recessions going back to the 1960s. The combined effect of the shortage of financial capital and declining quality of borrowers’ financial health caused banks to cut the loan supply in the 1990s. Clair and Tucker (1993) document 1 The term is now officially part of the language as one of several new words added to the Concise Oxford English Dictionary in June 2008; also included for the first time is the term “sub-prime.” that the phrase “credit crunch” has been used in the past to explain curtailment of the credit supply in response to both (i) a decline in the value of bank capital and (ii) conditions imposed by regulators, bank supervisors, or banks themselves that require banks to hold more capital than they previously would have held. A milder version of a full-blown credit crunch is sometimes referred to as a “credit squeeze,” and arguably this is what we observed in 2007 and early 2008; the term credit crunch was already in use well before any serious decline in credit supply was recorded, however. At that time the effects were restricted to shortage of liquidity in money markets and effective closure of certain capital markets that affected credit availability between banks. There was even speculation Paul Mizen is a professor of monetary economics and director of the Centre for Finance and Credit Markets at the University of Nottingham and a visiting scholar in the Research Division of the Federal Reserve Bank of St. Louis. This article was originally presented as an invited lecture to the Groupement de Recherche Européen Monnaie Banque Finance XXVth Symposium on Banking and Monetary Economics hosted by the Université du Luxembourg, June 18-20, 2008. The author thanks the organizers—particularly, Eric Girardin, Jen-Bernard Chatelain, and Andrew Mullineux—and Dick Anderson, Mike Artis, Alec Chrystal, Bill Emmons, Bill Gavin, Charles Goodhart, Clemens Kool, Dan Thornton, David Wheelock, and Geoffrey Wood for helpful comments. The author thanks Faith Weller for excellent research assistance. © 2008, The Federal Reserve Bank of St. Louis. The views expressed in this article are those of the author(s) and do not necessarily reflect the views of the Federal Reserve System, the Board of Governors, or the regional Federal Reserve Banks. Articles may be reprinted, reproduced, published, distributed, displayed, and transmitted in their entirety if copyright notice, author name(s), and full citation are included. Abstracts, synopses, and other derivative works may be made only with prior written permission of the Federal Reserve Bank of St. Louis. F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W S E P T E M B E R / O C TO B E R 2008 531 Mizen whether these conditions would spill over into the real sector, but there is little doubt now that there will be a decline in the terms and availability of credit for consumers and entrepreneurs. Disorder in financial markets occurred as banks sought to determine the true value of assets that were no longer being traded in sufficient volumes to establish a true price; and uncertainty prevailed among institutions aware of the need for liquidity but unwilling to offer it except under terms well above the risk-free rate. These conditions have now given way to the start of a credit crunch, and the restrictions on the credit supply will have negative real effects. Well-informed observers, such as Martin Wolf, associate editor and chief economics commentator of the Financial Times, are convinced that the credit crunch of 2007-08 will have a significance similar to that of earlier turning points in the world economy, such as the emerging markets crises in 1997-98 and the dotcom boom-and-bust in 2000 (Wolf, 2007). Like previous crises, the credit crunch has global implications because international investors are involved. The assetbacked securities composed of risky mortgages were packaged and sold to banks, investors, and pension funds worldwide—as were equities in emerging markets and dotcom companies before them. The 2007-08 credit crunch has been far more complex than earlier crunches because financial innovation has allowed new ways of packaging and reselling assets. It is intertwined with the growth of the subprime mortgage market in the United States—which offered nonstandard mortgages to individuals with nonstandard income or credit profiles—but it is really a crisis that occurred because of the mispricing of the risk of these products. New assets were developed based on subprime and other mortgages, which were then sold to investors in the form of repackaged debt securities of increasing sophistication. These received high ratings and were considered safe; they also provided good returns compared with more conventional asset classes. However, they were not as safe as the ratings suggested, because their value was closely tied to movements in house prices. While house prices were rising, these 532 S E P T E M B E R / O C TO B E R 2008 assets offered relatively high returns compared with other assets with similar risk ratings; but, when house prices began to fall, foreclosures on mortgages increased. To make matters worse investors had concentrated risks by leveraging their holdings of mortgages in securitized assets, so their losses were multiplied. Investors realized that they had not fully understood the scale of the likely losses on these assets, which sent shock waves through financial markets, and financial institutions struggled to determine the degree of their exposure to potential losses. Banks failed and the financial system was strained for an extended period. The banking system as a whole was strong enough to take these entities onto its balance sheet in 2007-08, but the effect on the demand for liquidity had a serious impact on the operation of the money markets. The episode tested authorities such as central banks, which were responsible for providing liquidity to the markets, and regulators and supervisors of the financial systems, who monitor the activities of financial institutions. Only now are lessons being learned that will alter future operations of the financial system to eliminate weaknesses in the process of regulation and supervision of financial institutions and the response of central banks to crisis conditions. These lessons include the need to create incentives that ensure the characteristics of assets “originated and distributed” are fully understood and communicated to end-investors. These changes will involve minimum information standards and improvements to both the modeling of risks and the ratings process. Central banks will review their treatment of liquidity crises by evaluating the effectiveness of their procedures to inject liquidity into the markets at times of crisis and their response to funding crises in individual banks. Regulators will need to consider the capital requirements for banks and offbalance sheet entities that are sponsored or owned by banks, evaluate the scope of regulation necessary for ratings agencies, and review the usefulness of stress testing and “fair value” accounting methods. This article consists of two parts: an outline of events and an evaluation. The first part discusses the background to the events of the past F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W Mizen year to discover how and why credit markets have expanded in recent years due to an environment of remarkably stable macroeconomic conditions, the global savings glut, and the development of new financial products. These conditions were conducive to the expansion of credit without due regard to the risks. It then describes the market responses to the deteriorating conditions and the response of the authorities to the crisis. The second part discusses how the structure and incentives of the new financial assets created conditions likely to trigger a crisis. It also evaluates the actions of the authorities and the regulators with some recommendations for reform. EVENTS Background: The Origins of the Crisis The beginnings of what is now referred to as the 2007-08 credit crunch appeared in early 2007 to be localized problems among lower-quality U.S. mortgage lenders. An increase in subprime mortgage defaults in February 2007 had caused some excitement in the markets, but this had settled by March. However, in April New Century Financial, a subprime specialist, had filed for Chapter 11 bankruptcy and laid off half its employees; and in early May 2007, the Swissowned investment bank UBS had closed the Dillon Reed hedge fund after incurring $125 million in subprime mortgage–related losses.2 This also might have seemed an isolated incident, but that month Moody’s announced it was reviewing the ratings of 62 asset groups (known as tranches) based on 21 U.S. subprime mortgage securitizations. This pattern of downgrades and losses was to repeat itself many times over the next few months. In June 2007 Bear Stearns supported two failing hedge funds, and in June and July 2007 three ratings agencies—Fitch Ratings, Standard & Poor’s, and Moody’s—all downgraded subprimerelated mortgage products from their “safe” AAA 2 As we will explain in more detail, defaults on subprime mortgages increased, causing losses; but, because investors “scaled up” the risks by leveraging their positions with borrowed funds, which were themselves funded with short-term loans, these small losses were magnified into larger ones. F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W status. Shortly thereafter Countrywide, a U.S. mortgage bank, experienced large losses, and in August two European banks, IKB (German) and BNP Paribas (French), closed hedge funds in troubled circumstances. These events were to develop into the full-scale credit crunch of 200708. Before discussing the details, we need to ask why the credit crunch happened and why now? Two important developments in the late 1990s and early twenty-first century provided a supportive environment for credit expansion. First, extraordinarily tranquil macroeconomic conditions (known as the “Great Moderation”) coupled with a flow of global savings from emerging and oil-exporting countries resulted in lower longterm interest rates and reduced macroeconomic volatility. Second, an expansion of securitization in subprime mortgage– backed assets produced sophisticated financial assets with relatively high yields and good credit ratings. The Great Moderation and the Global Savings Glut. The “Great Moderation” in the United States (and the “Great Stability” in the United Kingdom) saw a remarkable period of low inflation and low nominal short-term interest rates and steady growth. Many economists consider this the reason for credit expansion. For example, Dell’Ariccia, Igan, and Leavan (2008) suggest that lending was excessive—what they call “credit booms”—in the past five years. Beori and Guiso (2008) argue that the seeds of the credit booms were sown by Alan Greenspan when he cut short-term interest rates in response to the 9/11 attacks and the dotcom bubble, which is a plausible hypothesis, but this is unlikely to be the main reason for the expansion of credit. Shortterm rates elsewhere, notably the euro area and the United Kingdom, were not as low as they were in the United States, but credit grew there, too. When U.S. short-term interest rates steadily rose from 2004 to 2006, credit continued to grow. It is certainly true that the low real short-term interest rates, rising house prices, and stable economic conditions of the Great Moderation created strong incentives for credit growth on the demand and supply side. However, another important driving force of the growth in lending was found in the global savings glut flowing from China, Japan, Germany, and the oil exporters S E P T E M B E R / O C TO B E R 2008 533 Mizen Figure 1 Saving Ratios Percent of Disposable Income 20 18 16 14 12 10 8 6 4 2 Canada United Kingdom United States Euro Area 0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 SOURCE: OECD Economic Outlook and ECB Monthly Bulletin. that kept long-term interest rates down, as thenGovernor Bernanke noted in 2005 in a speech entitled, “The Global Saving Glut and the U.S. Current Account Deficit.” After the Asian crisis of 1997, many affected countries made determined efforts to accumulate official reserves denominated in currencies unlikely to be affected by speculative behavior, which could be used to defend the currency regime should events repeat themselves. (With larger reserves, of course, those events were unlikely to be repeated.) Strong demand for U.S. Treasuries and bonds raised their prices and lowered the long-term interest rate. Large savings flows from emerging markets funded the growing deficits in the industrialized countries for a time, and significant imbalances emerged between countries with large current account surpluses and deficits. These could not be sustained indefinitely; but, while they lasted and long-term interest rates were low, they encouraged the growth of credit. 534 S E P T E M B E R / O C TO B E R 2008 Figures 1 and 2 show that saving ratios declined and borrowing relative to income increased for industrialized countries from 1993 to 2006. The U.S. saving ratio fell from 6 percent of disposable income to below 1 percent in little over a decade, and at the same time the total debt– to–disposable income ratio rose from 75 percent to 120 percent, according to figures produced by the Organisation for Economic Co-operation and Development (OECD). The United Kingdom and Canada show similar patterns in saving and debtto-income ratios, as does the euro area—but the saving ratio is higher and the debt-to-income ratio is lower than in other countries. Similar experiences were observed in other countries. Revolving debt in the form of credit card borrowing increased significantly, and as prices in housing markets across the globe increased faster than income, lenders offered mortgages at ever higher multiples (in relation to income), raising the level of secured debt to income. Credit and housing bubbles reinforced F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W Mizen Figure 2 Debt to Income Ratios Liabilities in Percent of Disposable Income 180 160 Canada United Kingdom United States Euro Area 140 120 100 80 60 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 SOURCE: OECD Economic Outlook and ECB/Haver Analytics. each other. Borrowers continued to seek funds to gain a foothold on the housing ladder, reassured by the fact that the values of the properties they were buying were rising and were expected to continue to rise. Lenders assumed that house prices would continue to rise in the face of strong demand. In some cases, lenders offered in excess of 100 percent of the value of the property. Conditions in housing markets were favorable to increased lending with what appeared to be limited risk; lenders were prepared to extend the scope of lending to include lower-quality mortgages, known as subprime mortgages. Growth in the Subprime Mortgage Market. In the United States mortgages comprise four categories, defined as follows: (i) prime conforming mortgages are made to good-quality borrowers and meet requirements that enable originators to sell them to government-sponsored enterprises (GSEs, such as Fannie Mae and Freddie Mac); F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W (ii) jumbo mortgages exceed the limits set by Fannie Mae and Freddie Mac (the 2008 limit set by Congress is a maximum of $729,750 in the continental United States, but a loan cannot be more than 125 percent of the county average house value; the limit is higher in Alaska, Hawaii, and the U.S. Virgin Islands), but are otherwise standard; (iii) Alt-A mortgages do not conform to the Fannie Mae and Freddie Mac definitions, perhaps because a mortgagee has a higher loan-to-income ratio, higher loan-to-value ratio, or some other characteristic that increases the risk of default; and (iv) subprime mortgages lie below Alt-A mortgages and typically, but not always, represent mortgages to individuals with poor credit histories. Subprime mortgages are nevertheless difficult to define (see Sengupta and Emmons, 2007). One approach is to consider the originators of mortS E P T E M B E R / O C TO B E R 2008 535 Mizen gages: The U.S. Department of Housing and Urban Development (HUD) uses Home Mortgage Disclosure Act (HMDA) data to identify subprime specialists with fewer originations, a higher proportion of loans that are refinanced, and, because subprime mortgages are nonconforming, those that sell a smaller share of their mortgages to the GSEs. A second approach is to identify the mortgages by borrower characteristics: The Board of Governors of the Federal Reserve System, the Office of the Comptroller of the Currency, the Federal Deposit Insurance Corporation, and the Office of Thrift Supervision list a previous record of delinquency, foreclosure, or bankruptcy; a credit score of 580 or below on the Fair, Isaac and Company (FICO) scale; and a debt serviceto-income ratio of 50 percent or greater as subprime borrowers. Subprime products also exist in other countries where they may be marketed as interest-only, 100 percent loan-to-value, or selfcertification mortgages, but they are not as prevalent as in the United States. The main differences between a prime mortgage and a subprime mortgage from the borrower’s perspective are higher up-front fees (such as application and appraisal fees), higher insurance costs, fines for late payment or delinquency, and higher interest rates. Therefore, the penalty for borrowing in the subprime market, when the prime market is inaccessible, is a higher cost in the form of loan arrangement fees and charges for failing to meet payment terms. The main difference from the lender’s perspective is the higher probability of termination through prepayment (often due to refinancing) or default. The lender sets an interest rate dependent on a loan grade assigned in light of the borrower’s previous payment history, bankruptcies, debt-to-income ratio, and a limited loanto-value ratio, although this can be breached by piggyback lending. The lender offers a subprime borrower a mortgage with an interest premium over prime mortgage rates to cover the higher risk of default given these characteristics. Many other terms are attached to subprime mortgages, which sometimes benefit the borrower by granting allowances (e.g., to vary the payments through time), but the terms often also protect the lender (e.g., prepayment conditions that make it easier 536 S E P T E M B E R / O C TO B E R 2008 for the lender to resell the mortgage loan as a securitized product). The market for subprime mortgages grew very fast. Jaffee (2008) documents two periods of exceptional subprime mortgage growth. The first expansion occurred during the late 1990s, when the volume of subprime lending rose to $150 billion, totalling some 13 percent of total annual mortgage originations. This expansion came to a halt with the dotcom crisis of 2001. A second expansion phase was from 2002 until 2006 (Figure 3), when the subprime component of mortgage originations rose from $160 billion in 2001 to $600 billion by 2006 (see Calomiris, 2008), representing more than 20 percent of total annual mortgage originations. Chomsisengphet and Pennington-Cross (2006) argue that these expansions occurred because changes in the law allowed mortgage lending at high interest rates and fees, and tax advantages were available for secured borrowing versus unsecured borrowing.3 Another strong influence was the desire of mortgage originators to maintain the volume of new mortgages for securitization by expanding lending activity into previously untapped markets. Subprime loans were heavily concentrated in urban areas of certain U.S. cities —Detroit, Miami, Riverside, Orlando, Las Vegas, and Phoenix—where homeownership had not previously been common— as well as economically depressed areas of Ohio, Michigan, and Indiana, where prime borrowers that faced financial difficulties switched from prime to subprime mortgages. Securitization and “Originate and Distribute” Banking. Securitization was popularized in the United States when the Government National Mortgage Association (Ginnie Mae) securitized mortgages composed of Federal Housing Administration and Veterans Administration (FHA/VA) mortgages backed by the “full faith 3 Chomsisengphet and Pennington-Cross (2006) indicate that the Depository Institutions Deregulation and Monetary Control Act (1980) allowed borrowers to obtain loans from states other than the state in which they lived, effectively rendering interest rate caps at the state level ineffective. The Alternative Mortgage Transaction Parity Act (1982) allowed variable-rate mortgages, and the Tax Reform Act (1986) ended tax deduction for interest on forms of borrowing other than mortgages. These changes occurred well before the growth in subprime mortgage originations, but they put in place conditions that would allow for that growth. F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W Mizen Figure 3 Subprime Mortgage Originations, Annual Volume Billions of U.S. $ 700 600 500 400 300 200 100 0 2001 2002 2003 2004 2005 SOURCE: Data are from Inside Mortgage Finance, as published in the 2006 Mortgage Market Statistical Annual, Vol. 1. and credit” of the U.S. government for resale in a secondary market in 1968.4 In 1981, the Federal National Mortgage Association (Fannie Mae) began issuing mortgage-backed securities (MBSs), and soon after new “private-label” securitized products emerged for prime loans without the backing of the government.5 The European asset securitization market emerged later, in the 1990s, and picked up considerably in 2004. The originations occurred mainly in the Netherlands, Spain, and Italy (much less so in Germany, France and Portugal), but they were widely sold: More than half were sold outside the euro area, with onethird sold to U.K. institutions in 2005-06. 4 Ginnie Mae is a government-owned corporation within the Department of Housing and Urban Development (HUD) that was originally established in 1934 to offer “affordable” housing loans. In 1968 it was allowed by Congress to issue MBSs to finance its home loans. 5 Private-label MBSs dated back to the 1980s, but the process of repackaging and selling on auto loan receivables and credit card receivables goes back much farther—to the 1970s. F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W Securitization was undertaken by commercial and investment banks through special purpose vehicles (SPVs), which are financial entities created for a specific purpose—usually to engage in investment activities using assets conferred on them by banks, but at arm’s length and, importantly, not under the direct control of the banks. The advantage of their off-balance sheet status allows them to make use of assets for investment purposes without incurring risks of bankruptcy to the parent organization (see Gorton and Souleles, 2005). SPVs were established to create new asset-backed securities from complex mixtures of residential MBSs, credit card, and other debt receivables that they sold to investors elsewhere. By separating asset-backed securities into tranches (senior, mezzanine, and equity levels), the SPVs offering asset-backed securities could sell the products with different risk ratings for each level. In the event of default by a proportion of the borrowers, the equity tranche would be the first to incur losses, followed by mezzanine S E P T E M B E R / O C TO B E R 2008 537 Mizen and finally by senior tranches. Senior tranches were rated AAA—equivalent to government debt. In addition, they were protected by third-party insurance from monoline insurers that undertook to protect holders from losses, which improved their ratings. A market for collateralized debt obligations (CDOs) composed of asset-backed securities emerged; these instruments also had claims of different seniority offering varying payments. Banks held asset-backed securities in “warehouses” before reconstituting them as CDOs, so although they were intermediating credit to endinvestors, they held some risky assets on their balance sheets in the interim. Some tranches of CDOs were then pooled and resold as CDOs of CDOs (the so-called CDOs-squared); CDOs-squared were even repackaged into CDOs-cubed. These were effectively funds-of-funds based on the original mortgage loans, pooled into asset-backed securities, the lower tranches of which were then pooled again into CDOs, and so forth. As the OECD explains, the process involved several steps whereby “[the] underlying credit risk is first unbundled and then repackaged, tiered, securitised, and distributed to end investors. Various entities participate in this process at various stages in the chain running from origination to final distribution. They include primary lenders, mortgage brokers, bond insurers, and credit rating agencies” (OECD, 2008). Some purchasers were structured investment vehicles (SIVs)—off balance-sheet entities created by banks to hold these assets that could evade capital control requirements that applied to banks under Basel I capital adequacy rules. Others were bought by conduits—organizations similar to SIVs but backed by banks and owned by them. The scale of these purchases was large; de la Dehasa (2008) suggests that the volumes for conduits was around $600 billion for U.S. banks and $500 billion for European banks. The global market in asset-backed securities was estimated by the Bank of England at $10.7 trillion at the end of 2006. Ironically, many of the purchasers were offbalance-sheet institutions owned by the very banks that had originally sold the securitized products. This was not recognized at the time but 538 S E P T E M B E R / O C TO B E R 2008 would later come home to roost as losses on these assets required the banks to bring off-balancesheet vehicles back onto the balance sheet. A well-publicized aspect of the development of the mortgage securitization process was the development of residential MBSs composed of many different types of mortgages, including subprime mortgages. Unlike the earlier securitized offerings of the government-sponsored agency Ginnie Mae, which were subject to zero-default risk, these private-label MBSs were subject to significant default risk. Securitization of subprime mortgages started in the mid-1990s, by which time markets had become accustomed to the properties of securitized prime mortgage products that had emerged in the 1980s, but unlike government or prime private-label securities, the underlying assets in the subprime category were quite diverse. The complexity of new products issued by the private sector was much greater, introducing more variable cash flow, greater default risk for the mortgages themselves, and considerable heterogeneity in the tranches. In an earlier issue of this Review, Chomsisengphet and PenningtonCross (2006) show that the subprime mortgages had a wide range of loan and default risk characteristics. There were loans with options to defer payments, loans that converted from fixed to flexible (adjustable-rate) interest rates after a given period, low-documentation mortgages—all of which were supposedly designed to help buyers enter the housing market when (i) their credit or income histories were poor or (ii) they had expectations of a highly variable or rising income stream over time. Not all the mortgages offered as subprime were of low credit quality, but among the pool were many low-quality loans to borrowers who relied on rising house prices to allow refinancing of the loan to ensure that they could afford to maintain payments. The link between default risk and the movement of house prices was not fully appreciated by investors who provided a ready market for such securitized mortgages in the search for higher yields in the low-interest-rate environment. These included banks, insurance companies, asset managers, and hedge funds. Developments in the securitized subprime mort- F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W Mizen gage market were the trigger for the credit crunch. For this reason, the crisis is often referred to as a “subprime crisis.” In fact, as we shall see, any number of high-yield asset markets could have triggered the crisis. Subprime as a Trigger for the Credit Crunch Conditions in the housing and credit markets helped fuel the developing “crisis.” Credit scores of subprime borrowers through the decade 19952005 were rising; loan amounts on average were greater, with the largest increases to those borrowers with higher credit scores; and loan-to-value ratios were also rising (see Chomsisengphet and Pennington-Cross, 2006). The use of brokers and agents on commission driven by “quantity not quality” added to the problem, but provided the mortgagees did not default in large numbers (triggering clauses in contracts that might require the originator to take back the debts), there was money to be made. Mortgages were offered at low “teaser” rates that presented borrowers affordable, but not sustainable, interest rates, which were designed to increase. Jaffee (2008) suggested that the sheer range of the embedded options in the mortgage products made the decision about the best package for the borrower a complex one. Not all conditions were in the borrower’s best interests; for example, prepayment conditions that limit the faster payment of the loan and interest other than according to the agreed schedule often were even less favorable than the terms offered to prime borrowers. These conditions were designed to deter a borrower from refinancing the loan with another mortgage provider, and they also made it easier for the lender to sell the loan in a securitized form. In addition, brokers were not motivated as much by their future reputations as by the fee income generated by arranging a loan; in some instances, brokers fraudulently reported information to ensure the arrangement occurred. Policymakers, regulators, markets, and the public began to realize that subprime mortgages were very high-risk instruments when default rates mounted in 2006. It soon became apparent that the risks were not necessarily reduced by F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W pooling the products into securitized assets because the defaults were positively correlated. This position worsened because subprime mortgage investors concentrated the risks by leveraging their positions with borrowed funds, which themselves were funded with short-term loans. Leverage of 20:1 transforms a 5 percent realized loss into a 100 percent loss of initial capital; thus, an investor holding a highly leveraged asset could lose all its capital even when default rates were low.6 U.S. residential subprime mortgage delinquency rates have been consistently higher than rates on prime mortgages for many years. Chomsisengphet and Pennington-Cross (2006) record figures from the Mortgage Bankers Association with delinquencies 5½ times higher than for prime rates and foreclosures 10 times higher in the previous peak in 2001-02 during the U.S. recession. More recently, delinquency rates have risen to about 18 percent of all subprime mortgages (Figure 4). Figure 4 shows the effects of the housing downturn from 2005—when borrowers seeking to refinance to avoid the higher rates found they were unable to do so.7 As a consequence, subprime mortgages accounted for a substantial proportion of foreclosures in the United States from 2006 (more than 50 percent in recent years) and are concentrated among certain mortgage originators. A worrying characteristic of loans in this sector is the number of borrowers who defaulted within the first three to five months after receiving a home loan and the high correlation between the defaults on individual mortgage loans. Why did subprime mortgages, which comprise a small proportion of total U.S. mortgages, transmit the credit crunch globally? The growth in the scale of subprime lending in the United States was compounded by the relative ease with which these loans could be originated and the returns that could be generated by securitizing 6 This is why Fannie Mae and Freddie Mac faced difficulties in July 2008, because small mortgage defaults amounted to large losses when they were highly leveraged. 7 In the United States the process of obtaining a new mortgage to pay off an existing mortgage is known as “refinancing,” whereas in Europe this is often referred to as “remortgaging.” S E P T E M B E R / O C TO B E R 2008 539 Mizen Figure 4 U.S. Residential Mortgage Delinquency Rates Percent 20 18 16 All Mortgages Subprime Prime 14 12 10 8 6 4 2 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 SOURCE: Mortgage Bankers Association/Haver Analytics. the loans with (apparently) very little risk to the originating institutions. Some originators used technological improvements such as automatic underwriting and outsourcing of credit scoring to meet the requirements of downstream purchasers of the mortgage debt, but there is anecdotal evidence that the originators cared little about the quality of the loans provided they met the minimum requirements for mortgages to be repackaged and sold. The demand was strong for high-yielding assets, as the Governor of the Bank of England explained in 2007 (King, 2007): [I]nterest rates…were considerably below the levels to which most investors had become accustomed in their working lives. Dissatisfaction with these rates gave birth to the “search for yield.” This desire for higher yields could not be met by traditional investment opportunities. So it led to a demand for innovative, and inevitably riskier, financial instruments and for greater leverage. And the financial sector responded to the challenge by providing ever 540 S E P T E M B E R / O C TO B E R 2008 more sophisticated ways of increasing yields by taking more risk. Much of this demand was satisfied by residential MBSs and CDOs, which were sold globally, but as a consequence the inherent risks in the subprime sector spread to international investors with no experience or knowledge of U.S. real estate practices. When the lenders foreclosed, the claims on the underlying assets were not clearly defined—ex ante it had not been deemed important. Unlike in most European countries where there is a property register that can be used to identify—and repossess—the assets to sell them to recoup a fraction of the losses, the United States has no property register that allows the lender to repossess the property. As a consequence, once the loans had been pooled, repackaged, and sold without much effort to define ownership of the underlying asset, it was difficult to determine who owned the property. Moreover, differences in the various state laws meant that the rules permitting F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W