by MN Baily · Cited by 223 — Banks created off-balance sheet affiliated enti- ties such as Structured Investment Vehicles (SIVs) to purchase mortgage-related assets that were not subject to
Missing: bahamas | Must include: bahamas
37 KB – 47 Pages
PAGE – 1 ============
FIXING FINANCE SERIES Œ PAPER 3 | NOVEMBER 2008 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.
PAGE – 3 ============
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.
PAGE – 4 ============
TH E O RIGI NS OF THE FINANCIA L CR ISIS NO VEMB ER 2008 5 CON TEN TS 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 14Economic Incentives in the Housing and Mortgage Origination Markets 20 Securitization and the Funding of the Housing Boom 22 More Securitization and More LeverageŠCD Os, SIVs, and Short-Term Borrowing 27 Credit Insurance and Tremendous Growth in Credit Default Swaps 32 The Credit Rating Agencies 34Federal Reserve Policy, Foreign Borrowing and the Search for Yield 36 Regulation and Supervision 40The Failure of Company Risk Management Practices 42The Impact of Mark to Market 43Lessons from Studying the Origins of the Crisis 44 References 46About the Authors 47
PAGE – 5 ============
TH E O RIGI NS OF THE FINANCIA L CR ISIS NO VEMB ER 2008 7The ˜nancial 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 ˜nancial 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 signi˜cant factor in in˚ating 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 in˚ate 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 ˜nancial innovations, suddenly made subprime borrowers Š previously shut out of the mortgage markets Š attractive customers for mortgage lenders. Lend -ers devised innovative Adjustable Rate Mortgages (ARMs) Š with low ﬁteaser rates,ﬂ 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 Š 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 ﬁsecuritizingﬂ mortgages Š or the pooling of mortgages into packages and then sell -SUMMA RYing 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 len ding, Fannie Mae and Freddie Mac, developed this ˜nancing technique in the 1970s, adding their guarantees to these ﬁmortgage-backed securitiesﬂ (MBS) to ensure their marketability. For roughly three decades, Fannie and Freddie con˜ned their guarantees to ﬁprimeﬂ borrowers who took out ﬁconformingﬂ 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 ﬁcredit enhancement,ﬂ but this market stayed relatively small until the late 1990s. In this fashion, Wall Street inve stors effectively ˜nanced homebuyers on Main Street. Banks, thrifts, and a new industry of mortgage brokers originated the loans but did not keep them, which was the ﬁoldﬂ way of ˜nanc -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 ﬁColla teralized Debt Obligationsﬂ (sometimes with other asset-backed securities), and then divid -ing the cash ˚ows into different ﬁtranchesﬂ to ap -peal to diffe rent classes of investors with different tolerances for risk. By ordering the rights to the cash ˚ows, 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 ﬁmonolineﬂ 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
PAGE – 6 ============
TH E O RIGI NS OF THE FINANCIA L CR ISIS8 The Initiative on Business and Public Policy | THE BROOK ING S INSTITUTI ON the near equivalent by selling ﬁcredit default swapsﬂ (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 in˚ate home prices. These new ˜nancial 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, ˜na ncial institutions borrowed more and more money (i.e. increased their leverage) to ˜nance their pu rchases of mortgage-related securi -ties. Banks created off-balance sheet af˜liated 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 ﬁcollater -alized borro wingﬂ 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 the mselves 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 pro˜t- 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.
PAGE – 8 ============
TH E O RIGI NS OF THE FINANCIA L CR ISIS10 The Initiative on Business and Public Policy | THE BROOK ING S 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 ˜nancial 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 ˜nancial crisis that is wreaking havoc in ˜nancial markets in the U.S. and across the world has its origins in an asset price bubble that interacted both with new kinds of ˜nancial 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 ˜rst 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 -INTRO DUCTI ON
PAGE – 9 ============
TH E O RIGI NS OF THE FINANCIA L CR ISIS NO VEMB ER 2008 11 The driving force behind the mortgage and ˜ -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 Of˜ce of Federal Housing Enterprise Oversight (OFHEO) index of house prices hardly ever dropped. Only in 1981-82 did this index fall to any signi˜cant extentŠ5.4 per -centŠand 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 in˚ation. 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 in˚ation, 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 ˜rst 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 Š and witness those around them investing in homes Š a ﬁcontagionﬂ 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. 3The third factor is interest rates. After soaring to double digits and beyond in the in˚ationary 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 in˚ation) 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 Investment
PAGE – 10 ============
TH E O RIGI NS OF THE FINANCIA L CR ISIS12 The Initiative on Business and Public Policy | THE BROOK ING S INSTITUTI ON FIGU RE 1: Real Home Prices and Real Household Income (1976=100); 30-year Conventional Mortgage Rate Source: OH FEO ; Federal Reserve; Bureau of the Census. Home Prices and Income are de˜ated 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-˜nanced 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 ˜nally 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 ˜nally 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. 4The 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 1976198019841988199219962000200480100120140160180200024681012141618Index (1976 = 100)PercentAnnualized 30-year Conventional MortgageFixed Rate (Right Axis)Mean Real Household Income (1976 = 100)Left AxisReal Home Price Index (1976 = 100)Left Axis
PAGE – 11 ============
TH E O RIGI NS OF THE FINANCIA L CR ISIS NO VEMB ER 2008 13 4. 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 ˚at in cities like Charlotte and Pittsburgh. 6. Germany is the exception, where there was a huge building boom following reuni˜cation, 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)).
37 KB – 47 Pages