The current financial crisis is not--as many have said--a crisis of capitalism. It is in fact the opposite: a demonstration that well-intentioned government intervention in the private economy can have devastating consequences.
The crisis has its roots in the U.S. government's efforts to increase home ownership, especially among minority, low-income, and other underserved groups, through hidden financial subsidies rather than through direct government expenditures. Instead of a government subsidy, say, for down-payment assistance to low-income families, the government used regulatory and political pressure to force banks and other government-regulated or -controlled private entities to reduce lending standards, so more applicants would have access to mortgage financing.
The two key instances of this policy are the Community Reinvestment Act (CRA), adopted in 1977, and the affordable-housing "mission" adopted by Congress in the 1990s as a responsibility of the government-sponsored enterprises (GSEs), Fannie Mae and Freddie Mac. Amendments to the CRA in the early 1990s pressured banks into making loans they would not otherwise have made. Together, the tighter CRA requirements and the affordable-housing regulations imposed on the GSEs substantially reduced the standards that had to be met to qualify for a mortgage. The number of CRA loans was not large, but they required banks to devise ways of lending to people who would not previously have qualified for a mortgage. Once Fannie and Freddie began accepting loans with low down payments and other liberalized terms, the same unsound standards were extended to borrowers who could have qualified under the traditional underwriting standards. In addition, federal regulations encouraged bank lending for housing in preference to other lending, and tax policy favored borrowing against (and thus reducing) the equity in a house.
These policies were effective in the sense that they achieved some of the intended results. Between 1995--when lending quotas based on the CRA became effective--and 2005, the proportion of American households that owned their own home rose from 64 percent, where it had been for about twenty-five years, to 69 percent (Vlasenko 2008). A measure of the unintended results of federal policy, however, is that home prices doubled between 1995 and 2007; and that the housing bubble was composed--to an unprecedented degree--of subprime and other nonprime and risky loans. Banking-capital regulations and the deductibility of interest on home-equity loans made a crisis inevitable once this housing bubble collapsed.
The Community Reinvestment Act
As originally enacted in 1977, the CRA was a vague mandate for regulators to "consider" whether a federally insured bank was serving the needs of its entire community. The "community" was not defined, and the act stated only that it was intended to "encourage" banks to meet community needs. This encouragement included the denial of applications for mergers and acquisitions to banks judged to be in violation of the Act. The Act also stated, however, that serving community needs had to be done within the context of safe and sound lending practices. Although the Act was adopted to prevent "redlining"--the practice of refusing loans to otherwise-qualified borrowers in low-income areas--it also contained language that included small business, agriculture, and similar groups among the community interests that banks had to serve. With all of its ambiguities, the CRA was invoked relatively infrequently when banks applied for permission to merge or needed other regulatory approvals, until the Clinton administration decided to strengthen the Act.
This decision was probably due to the substantial amount of media and political attention that had been paid to the Boston Federal Reserve Bank's 1992 study (Munnell et al. 1992) of discrimination in home mortgage lending. The study concluded that while there was no overt discrimination in the allocation of mortgage funds, more subtle forms of discrimination led to better treatment of white applicants by loan officers as compared to minority-group applicants. The methodology of the study has since been questioned (e.g., McKinley 1994), but at the time of its publication, it seems to have been highly influential with regulators and members of the incoming Clinton administration.
In 1993, bank regulators initiated a major effort to reform the CRA regulations. Some of the context in which this was occurring can be gleaned from the following statement by Attorney General Janet Reno in January 1994:
We will tackle lending discrimination wherever and in whatever form it appears. No loan is exempt, no bank is immune. For those who thumb their nose at us, I promise vigorous enforcement. (Quoted in McKinley 1994, 30)
New rules were adopted in May 1995 that were phased in fully by July 1997. These rules attempted to establish objective criteria for determining whether a bank was meeting the standards of the CRA, taking much of the discretion out of the hands of bank examiners. "The emphasis on performance-based evaluation," writes A.K.M. Rezaul Hossain (2004, 54), "can be thought of as a shift of emphasis from procedural equity to equity in outcome," according to which it was no longer "sufficient for lenders to prove elaborate community lending efforts directed towards borrowers in the community"; instead, they had to prove "an evenhanded distribution of loans across LMI [low and moderate income] and non-LMI areas and borrowers."
In other words, it was now necessary for banks to show that they had actually made the requisite loans, not just that they were trying to find qualified borrowers. To help achieve this result, one of the standards in the new regulations required the use of "innovative or flexible" lending practices to address credit needs of LMI borrowers and neighborhoods (ibid., 57). This clearly meant the relaxation of lending standards, despite the original language about safe and sound practices.
Before the increases in housing prices that began in 2001, reviews of the CRA were generally unfavorable. The Act increased costs for banks, and there was an inverse relationship between their CRA lending and their regulatory ratings (cf. Benston 1999). One of the very few studies of CRA lending in comparison to normal lending was done by the Federal Reserve Bank of Cleveland, which reported in 2000 that "respondents who did report differences [between regular and CRA housing loans] most often said they had lower prices or higher costs or credit losses for CRA-related home purchase and refinance loans than for others" (Avery et al. 2000).
After 2000, however, and until housing prices stopped rising in late 2006 and early 2007, CRA lending occurred during a period of enormous growth in housing values, which tended to suppress the number of defaults and reduce loss rates. In this light, a recently released Fed Study (Kroszner 2008) arguing that CRA loans in 2005–6 performed comparably to other loans is irrelevant, at best. The real question, moreover, is not the default rates on the mortgages made under the CRA, which could not have been sufficient to cause a worldwide financial crisis. The most important fact about the CRA is the associated effort to reduce underwriting standards so that more low-income people could purchase homes. Once these standards were relaxed--particularly by allowing loan-to-value (LTV) ratios higher than the 80 percent that had previously been the norm--they spread rapidly to the prime market and to subprime markets, where loans were made by lenders other than insured banks.
The effort to reduce mortgage underwriting standards was led by the Department of Housing and Urban Development's enforcement of the National Homeownership Strategy, which was published in 1994 in response to a request by President Clinton. Among other things, it called for "financing strategies, fueled by the creativity and resources of the private and public sectors, to help homeowners that lack cash to buy a home or to make the payments." After this effort began, and the new regulations of 1997 were issued, LTV ratios and other indicators of loosened lending standards rose (cf. Demyanyk and Hemert 2008; England 2002). The era of subprime lending had begun.
There is no universally accepted definition of either "subprime" or "Alt-A" loans, except that neither of them is considered a prime mortgage (fifteen- or thirty-year amortization, fixed interest rate, good credit history). Thus, both represent enhanced risk. The Federal Reserve Bank of New York defines a subprime loan as one made to a borrower with blemished credit, or who provides only limited documentation. The federal bank regulators define a loan to a borrower with less than a 660 FICO score as subprime. Alt-A loans generally have a higher balance than subprime loans, and one or more elements of added risk, such as a high LTV ratio (often as a result of a piggyback second mortgage), interest-only payments, little or no income documentation, or an investor rather than a homeowner as the borrower. The term subprime, accordingly, generally refers to the financial capabilities of the borrower, while Alt-A generally refers to the quality of the loan terms.
The growth in housing demand doubled home prices between 1995 and 2007. According to data published by the Joint Center for Housing Studies at Harvard University, the share of all mortgage originations that were made up of conventional/conforming mortgages (that is, the twenty-percent down, thirty-year fixed-rate mortgage that had always been the mainstay of the U.S. mortgage market) fell from 57.1 percent in 2001 to 33.1 percent in 2006. Correspondingly, subprime loans (defined here as loans made to borrowers with blemished credit) rose from 7.2 percent to 18.8 percent, and Alt-A loans (defined as those made to speculative buyers or to buyers without the usual underwriting standards) rose from 2.5 percent to 13.9 percent.
The Role of Fannie Mae and Freddie Mac
Fannie Mae and Freddie Mac were two key mechanisms for transmitting relaxed lending standards to the prime market. The two big GSEs repurchased mortgages from originators of all types, and they were both key implementers of the new HUD policies.
In 1994, HUD had required that 30 percent of GSE mortgage purchases consist of affordable-housing mortgages. In 1997, apparently doubting that Fannie and Freddie were doing all they could for affordable housing, HUD commissioned the Urban Institute to study the GSEs' underwriting guidelines. The report concluded that
the GSEs' guidelines, designed to identify creditworthy applicants, are more likely to disqualify borrowers with low incomes, limited wealth, and poor credit histories; applicants with these characteristics are disproportionately minorities. Informants said that some local and regional lenders serve a greater number of creditworthy low-to-moderate income and minority borrowers than the GSEs, using loan products with more flexible underwriting guidelines than those allowed by Fannie and Freddie. (Temkin et al. 1999)
Following this report, Fannie and Freddie modified their automated underwriting systems to accept loans with characteristics that they had previously rejected. This opened the door to the acquisition of large numbers of nontraditional and subprime mortgages.
By 1997, Fannie was purchasing 97-percent LTV mortgages (meaning those with a 3-percent down payment), and by 2001, it was buying mortgages with no down payment at all. By 2007, Fannie and Freddie were required to show that 55 percent of their mortgage purchases were LMI. Moreover, 38 percent of all purchases had to be from underserved areas (usually in inner cities), and 25 percent had to be purchases of loans that had been made to low-income and very-low-income borrowers.
The decline in underwriting standards is clear in the financial disclosures of Fannie and Freddie. From 2005 to 2007, Fannie and Freddie bought approximately $1 trillion in subprime and Alt-A loans, amounting to about 40 percent of their mortgage purchases. This was only the end of the process: Freddie acquired 6 percent of its Alt-A loans in 2004; this jumped to 17 percent in 2005, 29 percent in 2006, and 32 percent in 2007. Fannie purchased 73 percent of its Alt-A loans during the latter three years. Similarly, in 2004, Freddie purchased 10 percent of the loans in its portfolio that had FICO scores of less than 620 (well under the 660 FICO score that federal bank regulators consider subprime); it increased these purchases to 14 percent in 2005, 17 percent in 2006, and 30 percent in 2007, while Fannie purchased 57.5 percent of its loans in this category during the same period. All told, Fannie and Freddie probably hold or have guaranteed $1.6 trillion in subprime and Alt-A mortgages, or roughly 40 percent of the total outstanding.
The GSEs' purchases of subprime and Alt-A loans affected the rest of the market for these mortgages in two ways. First, it increased the competition for these loans with private-label issuers. Prior to 2004, the financial advantages of the GSEs, including their access to cheaper financing, enabled them to monopolize the conventional/conforming market. When the GSEs ramped up their purchases of subprime and Alt-A loans, however, they not only took market share from the private-label issuers, but also created greater demand for subprime and Alt-A loans from mortgage originators and brokers. This drove up the value of subprime and Alt-A mortgages, reducing the risk premium that had previously suppressed subprime origination and securitization. As a result, many more marginally qualified or unqualified applicants for mortgages were accepted by private originators, including banks and other lenders (such as Countrywide), and then purchased for securitization by GSEs and private firms such as Bear Stearns. During this period, conventional loans (including jumbo loans) declined from 78.8 percent of all mortgages in 2003 to 50.1 percent at the end of 2006. From 2003 to 2006, subprime and Alt-A loans increased from 10.1 percent to 32.7 percent of all mortgages issued. (The remainder of non-conventional mortgages consisted of home-equity loans and F.H.A./VA loans.)
The GSEs' regulation-induced competition with private-label issuers almost certainly had the same effect on the quality of the mortgages the private-label issuers were securitizing. Since these mortgages aggregate to more than $2 trillion, this accounts for the weakness in bank assets that was the principal underlying cause of the financial crisis. In a very real sense, then, competition from Fannie and Freddie beginning in late 2004 caused both groups to scrape the bottom of the barrel--Fannie and Freddie in order to demonstrate to Congress their ability to increase support for affordable housing, and the private-label issuers trying to maintain their market share against the GSEs' increased demand for subprime and Alt-A products.
Thus, the gradual decline in lending standards that began with the revised CRA regulations in 1993, and that continued with the GSEs' attempts to show Congress that they were meeting their affordable-housing mission, had come to dominate mortgage lending in the United States.
Homeowner Options under U.S. Law
State-based residential finance laws, accommodated by the national mortgage system, give American homeowners two options that contributed substantially to the financial crisis.
First, any homeowner may, without penalty, refinance a mortgage whenever interest rates fall, or whenever home prices rise to a point where there is significant equity in the home.
The right to refinance without penalty is very rare in the commercial world, because it increases the difficulty of matching assets and liabilities and thus places significant risks on financial intermediaries. Because home mortgages can be refinanced at any time, banks and others must engage in sophisticated and expensive hedging transactions to protect themselves against the disappearance of their mortgage assets if interest rates decline. Moreover, the ability of homeowners to refinance their mortgages whenever they want to enables them to extract any equity that has accumulated in the home between the original purchase and any subsequent refinancing.
"Cash-out refinancing" permitted homeowners to obtain in cash, at the time of refinancing, a significant portion of the equity that had accumulated in the home up to that point. That equity, of course, could have been the result of a general increase in home prices rather than a homeowner's gradual amortization of principal under the mortgage loan. Once the housing bubble drove up the value of the equity in homes, cash-out refinancing became widespread: Homeowners treated their homes like savings accounts, drawing out funds through refinancing to buy cars, boats, or second homes, or to pay for other family expenditures. By the end of 2006, 86 percent of all home-mortgage refinancings were cash-out refinancings, and amounted to $327 billion in that year. Unfortunately, this meant that when the bubble burst, there was little equity in many homes, and therefore little reason to continue making payments on the mortgage. With subprime and Alt-A loans, where it is likelier that the homeowner will find it difficult or unprofitable to keep making payments, the default risk has proven to be much greater.
Second, walking away from an "underwater" mortgage is, de facto, legal.
In most states, mortgages are either "without recourse"--meaning that defaulting homeowners are not personally responsible for paying any difference between the value of the home and the principal amount of the mortgage obligation--or else the process for enforcing this obligation is so burdensome and time-consuming that lenders simply cannot afford to pursue it. The homeowner's opportunity to walk away from a home that is no longer more valuable than the mortgage it carries exacerbates the effect of the cash-out refinancing that occurred throughout the bubble period.
With the bursting of a housing bubble, cash-out refinancing and the ability to default on a mortgage without penalty would normally have led to concerns about how many prime loans will default. To this uncertainty was added the greater worry about the much higher default prospects of subprime and Alt-A loans. The resulting fears about the value of securitized mortgage holdings triggered the financial crisis.
Taxation policy also aggravated both the housing bubble and the problem of homeowners extracting equity from their homes.
The deductibility of mortgage interest is, of course, very well known, and it substantially tilts the decision about whether to rent or buy a home in favor of ownership. This involves more people in a given housing bubble, widening the scale of defaults when the bubble bursts. In this case, however, the default risk applies to high-end borrowers, not the poor: Only people who are well-enough off to (1) pay income tax and (2) itemize their deductions get the mortgage-interest deduction.
The deductibility of interest on home-equity loans has a similar effect, but it is less widely appreciated, even though its magnitude is greater. Interest on consumer loans of all other kinds--for cars, credit cards, and so on--is not deductible, but interest on home-equity loans is deductible no matter what the purpose of the loan or the use of the funds. As a result, homeowners are encouraged to take out home-equity loans to pay off their credit-card or auto loans, or to make the purchases that would ordinarily be made with credit cards, auto loans, or ordinary consumer loans. The resulting reduction in home equity enhances the likelihood that defaults and foreclosures will rise precipitously as housing prices fall, or as the economy contracts.
Under a 1988 international protocol known as Basel I (now being replaced by Basel II), the bank regulators in most of the world's developed countries adopted a uniform system of assigning bank assets to different risk categories. The purpose of the system was to permit some flexibility in the allocation of capital, based on the perceived riskiness of various types of assets. Capital is viewed as a shock absorber, and thus more capital should be held against the possibility of losses from riskier assets.
The general rule under Basel I is that banks are required to hold 8 percent risk-weighted capital in order to be adequately capitalized, and 10 percent in order to be well capitalized. Sovereign debt is assigned a risk weight of zero, while commercial loans receive a risk weight of 100 percent, meaning that to be adequately capitalized, a bank must have capital equal to 8 percent of the value of its commercial loans. Because residential mortgages are considered to be half as risky as commercial loans, they are assigned a 50-percent risk weight, so banks are required to hold only 4 percent capital against the value of a residential mortgage. In addition, asset-backed securities rated AAA or issued by a GSE are assigned a 20-percent risk weight, so only 1.6 percent capital (.08 × .20) is necessary for a bank to hold AAA-rated securities, including residential mortgage-backed securities.
The 50-percent risk weight placed on mortgages by the Basel rules provides an incentive for banks to hold mortgages in preference to commercial loans. Even more important, by purchasing a portfolio of GSE or AAA-rated residential mortgage-backed securities (MBSs), or by converting their portfolios of originated mortgages into an MBS portfolio rated AAA, banks can reduce their capital requirement to 1.6 percent, enabling them to lend more (and to make more money on these loans).
The 1.6-percent capital reserve might have been sufficient if the mortgages were of high quality, or if the AAA rating correctly predicted the risk of default, but the gradual decline in underwriting standards meant that even the prime mortgages in a given MBS pool often had high loan-to-value ratios, low FICO scores, or other indicators of low quality. This is all the more true of subprime and Alt-A loans, which were mixed in to MBSs containing prime loans, yet still received an AAA rating. Thus, the combination of the Basel rules with the subprime-lending boom, and its associated reduction in underwriting standards, left an indeterminate proportion of the world's financial institutions with deteriorating assets. It is these continuing and indeterminate losses that caused the financial crisis, and thus the global recession.
* * *
The policies I have reviewed here sparked the subprime-lending boom, helped inflate the housing bubble, and magnified the effects of its bursting. The financial crisis should thus be attributed to public policies--not to any "failure of capitalism."
Peter J. Wallison is the Arthur F. Burns Fellow in Financial Policy Studies at AEI. He thanks Edward Pinto for research in preparing the original version of this paper.
1. See the extensive discussion of the Community Reinvestment Act's development in Hossain 2004.
2. Quoted in Mason 2007. The National Homeownership Strategy was removed from the Department of Housing and Urban Development website in 2007.
3. Economagic.com, "Economic Time Series Page: US: Average Price of Houses Actually Sold," www.economagic.com/em-cgi/data.exe/cenc25/c25q07. The average price of homes sold increased from $153,500 in the fourth quarter of 1995 to $322,100 in the first quarter of 2007.
4. Federal National Mortgage Association, Fannie Mae's 2007 Annual Housing Activities Report, 5. Washington, D.C.: Department of Housing and Urban Development, 17 March 2008, http://184.108.40.206/offices/hsg/gse/reports/2007aharfnmanarrative.pdf
5. Federal National Mortgage Association, "2008 Q2 10-Q Investor Summary," 8 August 2008, 29–30, www.fanniemae.com/media/pdf/newsreleases2008_Q2_10Q_Investor_Summary.pdf; and Federal Home Loan Mortgage Corporation, "Freddie Mac Update," 30 August 2008, www.freddiemac.com/investors/pdffiles/investor-presentation.pdf
6. Joint Center for Housing Studies, The State of the Nation's Housing 2008, 39. Cambridge, Mass.: Harvard University, 2008, www.jchs.harvard.edu/publications/markets/son2008/index.htm
7. Ibid., 37.
Avery, Robert B., Raphael W. Bostic, and Glenn B. Canner. 2000. "The Performance and Profitability of CRA-Related Lending." Economic Commentary, Federal Reserve Bank of Cleveland (November). www.clevelandfed.org/research/Commentary/2000/1100.htm
Benston, George J. 1999. "The Community Reinvestment Act: Looking for Discrimination That Isn't There." Policy Analysis 354. Cato Institute, Washington, D.C., 6 October. www.cato.org/pub_display.php?pub_id=1213
Demyanyk, Yuliya, and Otto Van Hemert. 2008. "Understanding the Subprime Mortgage Crisis." Social Science Research Network, 19 August. http://ssrn.com/abstract=1020396
England, Robert Stowe. 2002. "Giving It 100 Percent." Mortgage Banking, 1 February.
Hossain, Rezaul A. K. M. 2004. "The Past, Present and Future of the Community Reinvestment Act (CRA): A Historical Perspective." Working Paper 2004-30, Department of Economics, University of Connecticut. www.econ.uconn.edu/working/2004-30.pdf
Kroszner, Randall S. 2008. "The Community Reinvestment Act and the Recent Mortgage Crisis." Speech at the Confronting Poverty Policy Forum Board of Governors of the Federal Reserve System, Washington, D.C., 3 December.
Mason, Joseph R. 2007. "A National Homeownership Strategy for the New Millennium." Market Commentary, 26 February.
McKinley, Vern. 1994. "Community Reinvestment Act: Ensuring Credit Adequacy or Enforcing Credit Allocation?" Regulation 17(4): 32. www.cato.org/pubs/regulation/regv17n4/vmck4-94.pdf
Munnell, Alicia H., Lynn E. Browne, James McEneaney, and Geoffrey M. B. Tootell. 1992. "Mortgage Lending in Boston: Interpreting HMDA Data." Working Paper 92-7, Federal Reserve Bank of Boston. www.bos.frb.org/economic/wp/wp1992/wp92_7.pdf
Temkin, Kenneth, George Galster, Roberto Quercia, and Sheila O'Leary. 1999. "A Study of the GSEs' Single-Family Underwriting Guidelines." Executive summary, Urban Institute, Washington, D.C., 9 April. www.urban.org/url.cfm?ID=1000205&renderforprint=1
Vlasenko, Polina. 2008. "Home Ownership in the United States." Great Barrington, Mass.: American Institute for Economic Research.