The Federal Reserve Board eagle logo links to home page

Remarks by Governor Edward M. Gramlich
At the Financial Services Roundtable Annual Housing Policy Meeting, Chicago, Illinois
May 21, 2004

Subprime Mortgage Lending: Benefits, Costs, and Challenges

One of the key financial developments of the 1990s was the emergence and rapid growth of subprime mortgage lending. Because of regulatory changes, the desire for increased profits, significant technological innovations, and liberalization in some government mortgage support programs, lending institutions began extending credit to millions of borrowers who previously would have been denied credit, both for mortgages and for other consumer loans. The increased availability of subprime mortgage credit has created new opportunities for homeownership and has allowed previously credit-constrained homeowners to borrow against the equity in their homes to meet a variety of needs. At the same time, increased subprime lending has been associated with higher levels of delinquency, foreclosure, and, in some cases, abusive lending practices. On a social level, one question is whether the gains afforded by these new market developments outweigh the losses. Another question is whether anything can be done to limit foreclosures. These are my topics today.

Basic Facts of Subprime Mortgage Lending
Subprime lending can be defined simply as lending that involves elevated credit risk. Whereas prime loans are typically made to borrowers who have a strong credit history and can demonstrate a capacity to repay their loans, subprime loans are typically made to borrowers who are perceived as deficient on either or both of these grounds. Obviously, lenders take a borrower's credit history into account when determining whether a loan is subprime; however, they also take into account the mortgage characteristics, such as loan-to-value ratio, or attributes of the property that cause the loan to carry elevated credit risk.

A borrower's credit history is usually summarized by a Fair Isaac and Company (FICO) credit score. Everything else being the same, borrowers with FICO scores below 620 are viewed as higher risk and generally ineligible for prime loans unless they make significant downpayments. But it is noteworthy that about half of subprime mortgage borrowers have FICO scores above this threshold, indicating that a good credit history alone does not guarantee prime status.

Compared with prime loans, subprime loans typically have higher loan-to-value ratios, reflecting the greater difficulty that subprime borrowers have in making downpayments and the propensity of these borrowers to extract equity during refinancing. They are also somewhat smaller in size. Whereas only about 1 percent of prime mortgages are in serious delinquency, the rate for serious delinquency on subprime is more than 7 percent. Not surprisingly, subprime mortgages also carry higher interest rates than those for prime loans. Evidence from surveys of mortgage lenders suggests that a weak credit history alone can add about 350 basis points to the loan rate.

The growth in subprime lending represents a natural evolution of credit markets. Two decades ago subprime borrowers would typically have been denied credit. But the 1980 Depository Institutions Deregulatory and Monetary Control Act eliminated all usury controls on first-lien mortgage rates, permitting lenders to charge higher rates of interest to borrowers who pose elevated credit risk, including those with weaker or less certain credit histories. This change encouraged further development and use of credit scoring and other technologies in the mortgage arena to better gauge risk and enabled lenders to price higher-risk borrowers rather than saying no altogether. Intense financial competition in the prime market, where mortgage lending was becoming a commodity business, encouraged lenders to enter this newer market to see if they could make a profit.

This evolutionary process was pushed along by various federal actions. The Community Reinvestment Act (CRA) of 1977, and later revisions to the regulation, gave banking institutions a strong incentive to make loans to low- and moderate-income borrowers or areas, an unknown but possibly significant portion of which were subprime loans. The Federal Housing Administration, which guarantees mortgage loans of many first-time borrowers, liberalized its rules for guaranteeing mortgages, increasing competition in the market and lowering interest rates faced by some subprime mortgage borrowers. Fannie Mae and Freddie Mac, giant secondary market purchasers, sought to meet their federally mandated affordable housing goals by expanding into the prime and lower-risk segment of the subprime mortgage market. They now provide many direct mortgage lenders with other potential buyers for their subprime mortgages. Fannie and Freddie are both working on techniques to extend automated underwriting to the subprime market, an innovation that should further lower costs in this market.

Some data can illustrate these and other features of the subprime market. Table 1 shows the dramatic growth in subprime lending. Subprime mortgage loan originations rose by the whopping rate of 25 percent per year over the 1994-2003 period, nearly a ten-fold increase in just nine years. Even prime mortgage lending grew by a strong annual rate of 17 percent, reflecting many of the same trends. While the annual share of originations accounted for by subprime lending varies with credit conditions and the business cycle, this share has roughly doubled over the same period.

Table 2 shows one consequence of this striking rise in mortgage credit. Overall homeownership rates have gone from 64 percent to more than 68 percent over this period. Nearly 9 million more households own their home now than just nine years ago. A major portion of this expansion in homeownership seems clearly attributable to the increased access to credit afforded by expansions in prime and subprime mortgage lending. Internationally, our current homeownership rate now puts the United States in the top tier of developed countries on homeownership rates, on a par with the United Kingdom, still slightly behind Spain, Finland, Ireland, and Australia, but well ahead of the homeownership rate in most other developed countries.

The distribution of this added homeownership looks promising as well. More than half of these new homeowning households are minorities. While numbers of white homeowners did advance slightly more than 4 million, blacks gained 1.2 million, Hispanics 1.9 million, and the residual "other" category, including Asians and those reporting other races, 1.6 million. Nearly half of all black and Hispanic households now own their own home. These homeownership rates are still well below those of whites but are catching up. With respect to household income, the data also show homeownership rates increasing on both sides of the income distribution.

A more fine-grained disaggregation of subprime mortgage lending is shown in table 3. These data are derived from information reported by mortgage lenders under the Home Mortgage Disclosure Act (HMDA). Subprime loans in the HMDA data are identified using a list of subprime mortgage lenders developed by the Department of Housing and Urban Development (HUD). The table gives the share of subprime mortgage loans in total mortgage loans originated by these subprime lenders within the relevant income, neighborhood, and racial or ethnic categories for 2002. Nationally 9 percent of mortgages were subprime in that year (table 1). For lower-income home purchase loans, the share of subprime lending was above the overall national rate, as was the share of lower-income home equity lending. Similar results are shown for lower-income neighborhoods, for the Native American, black, and Hispanic racial and ethnic categories, and for neighborhoods with high minority concentrations. These differential subprime shares do not necessarily suggest exploitation of these borrowers because borrowers with the relevant characteristics may also be worse credit risks, but they bear watching.

Finally, on the supply side of the market, table 4 shows the institutions that were identified by HUD as mainly subprime lenders in the HMDA data. Although many other institutions do some subprime lending, a review of subprime lenders identified by this list is informative. In numbers, only five commercial banks were mainly subprime lenders in 2002, though these five banks were large enough that they accounted for 27 percent of the mortgage loans of the subprime lenders. Similarly, affiliates of financial holding companies (such as City Financial) constituted only 19 percent of these lenders but 43 percent of the subprime loans. On the other side, independent mortgage companies were quite numerous but accounted for the relatively small total of 12 percent of subprime mortgage loans.

This lender breakdown is important for a reason that may not be obvious. As will be argued below, the growth of subprime mortgage lending has had its positive aspects in bringing credit to borrowers who previously would have been denied, but it has also entailed risks. Subprime borrowers pay higher rates of interest, go into delinquency more often, and have their properties foreclosed at a higher rate than prime borrowers. Many subprime lenders operate under the highest lending standards, but fraud, abuse, and predatory lending problems have also been a troublesome characteristic of the subprime market.

A good defense against predatory lending, perhaps the best defense society has devised, is a careful compliance examination for banks. All commercial banks, thrifts, and subsidiaries of banks undergo compliance exams on approximately a three-year cycle, with these three types of institutions constituting 45 percent of the identified subprime mortgage loans made in 2002. Affiliates of financial holding companies, covering an additional 43 percent of mortgage loans among these subprime specialists, can be revised for compliance with lending laws, though on a less thorough and less timely basis. Independent mortgage companies, covering 12 percent of these mortgage loans, are not systematically examined at all, though they are subject to the jurisdiction of the Federal Trade Commission, which handles complaints about these entities and may take enforcement measures.

Homeownership and Delinquencies
The obvious advantage of the expansion of subprime mortgage credit is the rise in credit opportunities and homeownership. Because of innovations in the prime and subprime mortgage market, nearly 9 million new homeowners are now able to live in their own homes, improve their neighborhoods, and use their homes to build wealth. Studies of neighborhood effects consistently impute a large importance to the rate of homeownership in encouraging a host of positive spillover benefits. In addition, subprime lending has enabled millions of cash-strapped home owners to liquefy the equity in their homes to help reduce the burden of other financial obligations or to improve their homes.

Given the greater credit risks of subprime lending, the obvious disadvantage involves elevated rates of foreclosure and of the incidence of households seriously delinquent on their mortgages. Households in foreclosure lose all the equity they have built up in their homes, typically the largest component of their wealth. There is also evidence of serious neighborhood blight if foreclosure rates, and abandoned properties, proliferate in a given city area.

To see better how those trends balance, I find it useful to analyze delinquency rates, shown in table 5. The table shows prime and subprime mortgage delinquency rates for mortgages outstanding at the end of 2003. Serious delinquency is defined as mortgages either in the foreclosure process or more than ninety days delinquent--the rates are about 1 percent in the prime market and slightly more than 7 percent in the subprime market. Delinquency rates are also higher in the subprime than prime market at the thirty- to sixty-day mileposts. It is difficult to tell whether delinquency rates are rising over time because the trend effect is confounded with the aging effect--that is, loans of an earlier vintage would have more chance to become delinquent. But from annual breakdowns of these data (not shown), I do not see major trends in delinquency rates.

Given the generally low level of serious delinquencies, a purely numerical analysis seems to suggest that significant net social benefits have resulted from the rise in credit extensions and homeownership. Breakdowns of the mortgage data (again not shown) indicate that the majority of new homeowners were able to obtain prime mortgages, with very low resulting delinquency rates. Even in the subprime market, where delinquencies are more common, more than 90 percent of these borrowers are not seriously delinquent.

But delinquencies could be more of a problem than such calculations suggest. First, as table 5 shows, another 8 percent of subprime borrowers are not in serious delinquency, but they are thirty or sixty days delinquent on their loans. Presumably these subprime borrowers are in danger of reaching serious delinquency, and they may be strapped for cash. Borrowers who are strapped for cash become vulnerable to predatory lenders and to later foreclosure proceedings. Second, the individual pain of a foreclosure, with the borrower's losing all home equity and most of his or her wealth, probably exceeds the individual gain from a nonforeclosure. If so, a mere tally of nondelinquent new homeowners may substantially overstate net social benefits. Third, there could be problems that are revealed only with a more-refined analysis of the data: For example, delinquencies may represent significant problems for certain racial or ethnic groups or for certain neighborhoods.

In response to the incidence of foreclosures and delinquencies in some urban markets and other evidence of predatory lending, housing groups have marshalled new resources to fight mortgage foreclosure. The Chicago affiliates of the Neighborhood Reinvestment Corporation (NRC), for example, have initiated a multipronged fight against foreclosures that features data analysis (to determine the source of the foreclosure among the complex set of brokers, lenders, and secondary purchasers), lending counseling, and alternative sources of credit. Many lower-income homeowners are led to high-cost refinance loans simply because they need cash, and alternative credit on reasonable terms can be highly effective in reducing foreclosures.

Despite the caveats, the net social evaluation of these trends is probably a strong positive. The 9 million new homeowners, more than half of whom are minorities and many of whom have lower incomes, suggest that credit and ownership markets are democratizing. Millions of lower-income and minority households now have a chance to own homes and to build wealth; and the vast majority of these new homeowners do not appear to be having credit problems. The rates of serious delinquencies and near-serious delinquencies do raise important warning flags and should inspire renewed efforts to prevent foreclosures, but they do not seem high enough to challenge the overall positive assessment.

While the basic developments in the subprime mortgage market seem positive, the relatively high delinquency rates in the subprime market do raise issues. Even further social benefits would result if various institutions could agree on and implement changes that would lower foreclosures.

For mortgage lenders the real challenge is to figure out how far to go. Over the 1990s both prime and subprime mortgage loans grew rapidly, and homeownership possibilities were extended to millions of households. Ownership rates have now risen to more than 68 percent, and foreclosures are relatively high in the subprime market, an important source of new mortgage loans. Do these circumstances mean that the expansion phase is over, or should lenders seek new possibilities for extending prime and subprime mortgage credit? If lenders do make new loans, can conditions be designed to prevent new delinquencies and foreclosures?

The secondary market purchasers, Fannie Mae and Freddie Mac, face similar challenges. Both purchasers have conditions on their purchases of mortgage loans requiring disclosure, limitations on certain practices, and restrictions against inappropriate classification of prime borrowers as subprime. These restrictions may have limited some subprime lending, but they also have had the beneficial effect of influencing lending standards within a portion of the subprime market. Fannie Mae and Freddie Mac should be continually testing their restrictions to find a set of rules that adequately protect borrowers without unduly constricting lenders.

Housing support groups, such as the NRC, must have a strategy of consolidating gains. These housing support groups are typically in the business of encouraging homeownership. But pushing homeownership so far that all the gains are offset by higher foreclosures does no good. Groups must develop more-effective lending programs, provide alternative and cheaper sources of credit for cash-strapped homeowners, and develop data systems so that institutions and brokers who are facilitating unnecessary foreclosures can be identified and penalized.

Federal regulators face challenges as well. The Federal Reserve Board has already revised HMDA to ask for rate information on subprime mortgage loans so that subprime mortgage markets can be better analyzed and understood. Using its authority to regulate high-cost loans under the Home Ownership and Equity Protection Act, the Fed has also made several changes to protect consumers with high-cost mortgages. These changes include a provision that effectively classifies loans that carry single-premium credit insurance as high-cost loans and allows them to receive the protections of the act. Further, as umbrella supervisor of financial holding companies, the Fed is also instituting new procedures for monitoring the lending practices of affiliates of these holding companies, which are responsible for an important share of subprime mortgage lending. The Fed and the other bank regulators are currently considering tightening CRA so that lenders guilty of lending violations do not get automatic credit for their loans. But despite these actions by the Fed and other bank regulators, we still have no obvious way to monitor the lending behavior of independent mortgage companies.

There are challenges for everybody. Rising to these challenges will ensure that continued subprime mortgage lending growth will generate even more social benefits than it seems to have already generated.

Table 1
Subprime Mortgage Originations, 1994-2003
Billions of current dollars except as noted

Year Subprime
Subprime as a
percent of total
1994 35.0 773.1 4.5
1995 65.0 635.8 10.2
1996 96.5 785.3 12.3
1997 125.0 859.1 14.5
1998 150.0 1,430.0 10.5
1999 160.0 1,275.0 12.5
2000 138.0 1,048.0 13.2
2001 173.0 2,100.0 8.2
2002 241.0 2,780.0 8.7
2003 332.0 3,760.0 8.8
growth rate (%)
25.0 17.6 --

Source: Mortgage Statistical Annual, March 2004.

Return to text

Table 2
Homeownership Totals and Rates
By Race and Household Income
Millions except as noted

  1994 2003
Characteristic Households Owners Rate (%) Households Owners Rate (%)
Total 98.7 63.1 63.9 105.6 72.0 68.2
  White 76.6 53.6 70.0 76.5 57.7 75.4
  Black 11.6 4.9 42.2 12.6 6.1 48.4
  Hispanic 7.7 3.2 41.6 11.0 5.1 46.4
  Other1 2.9 1.5 51.7 5.5 3.1 56.4
  Greater than median NA NA 78.5 NA NA 83.6
  Less than median NA NA 48.4 NA NA 51.8

1. "Other" includes other races and households indicating more than one race. Rates are
calculated from counts of households and owners. Return to table

Source: U.S. Census Bureau.

Return to text

Table 3
Profile of Subprime Mortgage Lending, 2002

Characteristic Subprime share of number of
home purchase loans
Subprime share of number of
home equity loans
Borrower income1
  Lower 10.9 14.4
  Middle 11.2 10.5
  Higher 9.0 6.7
Neighborhood income2
  Lower 16.4 17.8
  Middle 10.7 9.8
  Higher 7.7 6.1
Race of borrower
  Native American 16.0 13.6
  Asian or Pacific Islander 9.4 5.7
  Black 27.0 20.8
  Hispanic 19.6 14.5
  White 7.4 5.7
Neighborhood racial composition
  Less than 10 percent minority 6.9 6.8
  10 - 49 percent minority 12.0 10.0
  Greater than 50 percent minority 20.8 20.8

1. Lower income means that the borrower's income is less than 80 percent of the median family income of their MSA. Higher income is 120 percent. Return to table

2. Lower income means that the median family income in the local Census tract is less than 80 percent of median of their MSA. Higher income is above 120 percent. Return to table

Source: 2002 HMDA data.

Return to text

Table 4
Subprime Lending by Type of Institution
and Volume of Lending, 2002
Percent except as noted

Type of institution Number of
Share of
Number of
Commercial bank 5 2.7 27.0
Thrift 11 6.0 13.8
Independent mortgage company 113 61.8 11.8
Subsidiary of bank 19 10.4 4.4
Affiliate of financial holding company 35 19.1 43.0
Total 183 100.0 100.0

Source: Lenders classified according to Department of Housing and Urban Development (HUD) list of subprime lenders. Number of loans from HMDA data for 2002.

Return to text

Table 5
Loan Delinquency Rates, 2003

Type of mortgages Overdue Foreclosure
30 days 60 days 90 days
Prime 2.26 0.58 0.64 0.48 1.12
Subprime 6.75 2.12 3.98 3.38 7.36

1. Defined as in foreclosure status or ninety or more days delinquent. Return to table

Source: Loan Performance.

Return to text

Return to topReturn to top

2004 Speeches